CN1925625B - Device for recording and reproducing digital video signal and method for recording and reproducing the same - Google Patents

Abstract

A digital video signal recording and reproducing apparatus for recording digital video signals coded by motion compensation prediction and orthogonal transformation on a recording medium, and reproducing the data from the recording medium. In the data structure of a digital video signal, an I picture that can be reproduced individually in an image is divided into n areas in the vertical direction, and data is arranged in area units by giving priority to an area at the center of the screen in one GOP. front end. A reproduced image is output by reproducing the I picture. In the case where the entire I picture cannot be read out within a limited time, the area that cannot be read out is interpolated by the data of the previous screen.

Description

Device for recording and reproducing digital video signal and method for recording and reproducing the same

technical field

The present invention relates to recording and reproducing apparatus for recording and reproducing digital video signals of digital video signals, in particular to recording and reproducing digital video signals encoded according to motion compensation prediction and orthogonal transform on a medium such as an optical disc Recording and playback equipment for digital video signals.

Background technique

Fig. 1 is a circuit block diagram of a conventional optical disk recording and reproducing apparatus shown in Japanese Patent Application Laid-Open No. 4-114369 (1992). Referring to FIG. 1, reference numeral 201 denotes an A/D (Analog/Digital) converter for converting video signals, audio signals, etc. into digital information. Reference numeral 202 denotes a data compression circuit, 203 denotes a frame sector conversion circuit for converting compressed data into sector data equal to an integral multiple of a frame period, and 204 denotes an error correction circuit for adding an error correction signal to the sector data Encoder, 205 represents a modulator for modulating the interference signal between codes in the recording medium into a predetermined modulation code to reduce interference, 206 represents a laser drive circuit that modulates the laser according to the modulation code, and 207 represents the laser output switch. In addition, reference numeral 208 denotes an optical head for emitting laser light, 209 denotes an actuator for scanning the laser beam from the optical head 208 on a track, 210 denotes a reciprocating motor for extending the optical head 208, 211 denotes a disc motor for rotating the optical disc 212, 219 denotes a motor drive circuit, 220 denotes a first control circuit, and 221 denotes a second control circuit. In addition, reference numeral 213 denotes a playback amplifier for amplifying a playback signal from the optical head 208, 214 denotes a demodulator for acquiring data from a recorded modulated signal, 215 denotes an error correction decoder, and 216 denotes a frame sector. 217 denotes a data expansion circuit that expands compressed data, and 218 denotes a D/A (Digital/Analog) converter that converts expanded data into, for example, analog video signals and audio signals.

FIG. 2 is a block diagram showing the internal structure of the data compression circuit 202 in FIG. 1 . In FIG. 2, a digital video signal input from an A/D converter 201 is input to a memory circuit 301. As shown in FIG. The video signal 321 output from the storage circuit 301 is supplied to the first input terminal of the subtractor 302 and the second input terminal of the motion compensation prediction circuit 310 . The output of the subtractor 302 is output to a quantizer 304 via a DCT (Discrete Cosine Transform) circuit 303 . The output of quantizer 304 is sent to an input terminal of transmission buffer 306 via variable length coder 305 . The output of the transmission buffer 306 is supplied to the frame sector conversion circuit 203 . At the same time, the output of the quantizer 304 is sent to an inverse DCT circuit 308 via an inverse quantizer 307 . The output of the inverse DCT circuit 308 is applied to a first input of an adder 309 . The output 322 of the adder 309 is supplied to a first input of the motion compensation prediction circuit 310 . The output 323 of the motion compensation prediction circuit 310 is sent to the second input terminal of the adder 309 and the second input terminal of the subtractor 302.

FIG. 3 is a block diagram showing the internal structure of the motion compensation prediction circuit 310 in FIG. 2 . In FIG. 3, the output 322 of the adder 309 is supplied to the input terminal 401a, and the output 321 of the storage circuit 301 is supplied to the input terminal 401b. The signal 322 input from the input terminal 401a is input into the frame memory 404a or the frame memory 404b via a switch 403 . The reference image output from the frame memory 404a is supplied to a first input terminal of a motion vector detection circuit 405a. The video signal 321 input from the input terminal 401b is input to the second input terminal of the motion vector detection circuit 405a. The output of the motion vector detection circuit 405a is sent to the prediction mode selector 406. At the same time, the reference image output from the frame memory 404b is supplied to the first input terminal of the motion vector detection circuit 405b. The video signal 321 input from the input terminal 401b is supplied to the second input terminal of the motion vector detection circuit 405b. The output of the motion vector detection circuit 405b is supplied to the second input terminal of the predictive mode selector 406 . The video signal 321 input from the input terminal 401b is supplied to a third input terminal of the prediction mode selector 406 . A zero signal is sent to the second input of switch 407 . A second output of the predictive mode selector 406 is connected to a third input of a switch 407 . The output 323 of the switch 407 is output from the output terminal 402 .

FIG. 4 is a block diagram showing the internal structure of the data expansion circuit 217 in FIG. 1. Referring to FIG. In FIG. 4 , the video signal input from the frame sector inverse conversion circuit 216 is input to the reception buffer 501 . The output of the receive buffer 501 is input to a variable length decoder 502 , and the output of the decoder is inversely quantized at an inverse quantizer 503 . Its output is then subjected to an inverse discrete cosine transform at an inverse DCT circuit 504 . The transformed output is fed to a first input of adder 506 . At the same time, the output of the receiving buffer 501 is sent to the predictive data decoding circuit 505 , and the output of the predictive data decoding circuit 505 is sent to the second input terminal of the adder 506 . The output of the adder 506 is sent to the D/A converter 218 via the storage circuit 507 .

The operation of the device is described below. As a high-efficiency coding method in the field of video signal coding, there is a coding algorithm by means of the MPEG (Moving Picture Efficient Grouping) method. This is a hybrid coding method that combines inter-frame prediction coding using motion compensation prediction with intra-frame conversion coding. A commonly used example in this regard is to utilize the data compression circuit 202 having the structure shown in FIG. 2, while adopting the above-mentioned MPEG method.

FIG. 5 shows a simplified data organization structure (layered structure) of the MPEG method. In FIG. 5, reference numeral 621 denotes a sequential layer including a group of pictures (hereinafter referred to as "GOP") including a plurality of frame data entries. 622 represents a GOP layer containing several images (screens), 623 represents a screen divided into several blocks, 624 represents a slice layer with several macro blocks, 625 represents a macro block layer, and 626 represents an 8-pixel A block layer consisting of ×8 pixels.

This macroblock layer 625 is a block composed of at least units of 8 pixels by 8 pixels in, for example, MPEG. This block is a unit for performing DCT. At this time, the entire 6 blocks, including four adjacent Y signal blocks, a Cb block corresponding to the existing Y signal block, and a Cr block, are called macroblocks. A plurality of such macroblocks constitute a slice. Furthermore, several macroblocks constitute a minimum unit of motion compensation prediction, and motion vectors for motion compensation prediction are formed in units of some macroblocks.

Next, the inter prediction encoding process will be explained. Fig. 6 shows an overview of interframe predictive coding. The picture is divided into three types, namely intra-frame coded picture (hereinafter referred to as I picture), unidirectional predictive coded picture (hereinafter referred to as P picture), and bidirectional predictive coded picture (hereinafter referred to as B picture) elephant).

For example, assuming that one of the N images is an I image, and one of the M images is a P image or an I image, then the (N×n+M)th image constitutes An I image, the (N×n+M×m) image (m≠1) constitutes a P image, from the (N×n+M×m+1) image to the ( Each picture of Nxn+Mxm+M-1) pictures constitutes a B picture, where n and m are integers, and 1≤m≤N/M. At this time, each picture from the (Nxn+1)th picture to the (Nxn+N)th picture is collectively referred to as a GOP (group of pictures).

FIG. 6 shows a case where N=15 and M=3. In FIG. 6, the I picture is not subjected to inter-frame prediction but only intra-frame transform-coded. A P-picture undergoes prediction from an I-picture immediately preceding the P-picture or from a P-picture. For example, the sixth picture in Fig. 6 is a P picture. The 6th picture undergoes prediction from the 3rd I-picture. Also, the ninth P-picture in Fig. 6 undergoes prediction from the sixth P-picture. B-pictures undergo predictions from I-pictures or P-pictures immediately preceding and following the B-picture. For example, the 4th and 5th B-pictures in FIG. 6 are predicted from both the 3rd I-picture and the 6th P-picture. Therefore, the 4th and 5th pictures are encoded after the encoding of the 6th picture.

Hereinafter, processing of data including the circuit 202 will be described with reference to FIG. 2 . The storage circuit 301 outputs the input digital video image signal after rearranging the signals in the encoding sequence. In other words, the first B picture is coded after the third I picture in FIG. 6, for example, as described above. Therefore, the order of the images is rearranged. Figure 7 illustrates the operation of this rearrangement. The images input in the order shown in Fig. 7A are then output in the order shown in Fig. 7B.

Furthermore, after the video signal 321 output by the storage circuit 301 is subtracted between the images with the output prediction image 323 from the automatic compensation prediction circuit 310 on the subtractor 302 to reduce redundancy in the time axis direction, A DCT in the direction of the spatial axis is performed. Next, the transform coefficients are quantized, subjected to variable-length coding, and then output through the transmission buffer 306 . At the same time, the quantized transform coefficients are inversely quantized and subjected to inverse DCT. Afterwards, the coefficient is added to the adder 309 and added to the predicted image 323, so that a decoded image 322 is obtained. The decoded picture 322 is input to motion compensated prediction circuit 310 for later encoding of the picture.

Next, the operation of the motion compensation prediction circuit 310 will be described with reference to FIG. 3 . The motion compensation pre-circuit 310 utilizes two reference images stored in the frame memories 404a and 404b to perform motion compensation prediction of the video signal 321 from the storage circuit 301 to output a prediction image 323 .

Initially, when a picture 322 encoded and decoded as described above is either an I picture or a P picture, it is stored in frame memory 404a or frame memory 404b for encoding of subsequent pictures. At this time, the switch 403 is switched, thereby selecting one of the two frame memories 404a and 404b, which is updated temporally prior to the other. However, when the decoded picture 322 is a B picture, neither the frame memories 404a nor 404b are written.

For example, when the 1st and 2nd pictures in Fig. 7 were coded with the switching mode of the switch 403, the 0th P picture and the 3rd I picture were stored in the frame memories 404a and 404b respectively . Also, when the 6th P picture is coded and decoded, the decoded picture of the 6th P picture is rewritten in the frame memory 404a.

Next, when the 4th and 5th B pictures are coded, the 6th P picture and the 3rd I picture are stored in the frame memories 404a and 404b, respectively. Further, when the 9th P picture is coded and decoded, the decoded picture of the 9th P picture is rewritten in the frame memory 404b. Thus, when the seventh B picture and the eighth B picture are encoded, the sixth P picture and the ninth P picture are stored in the frame memories 404a and 404b, respectively.

When the video signal 321 from the storage circuit 301 is input to the motion compensation prediction circuit 310, the motion vector detection circuits 405a and 405b detect a motion vector according to the reference images stored in the frame memories 404a and 404b, and output a motion compensation prediction map elephant. In other words, the video signal 321 is divided into blocks. Then, a block is selected that minimizes the predicted distortion in its reference image for each block. Then, the relative position of the block is output as a motion block, while the block itself is output as a motion-compensated predicted image.

At the same time, the prediction mode selector selects an image with the least prediction distortion or an average image thereof from the two motion compensation predicted images from the automatic vector detection circuits 405a and 405b. The selected image is then output as a predicted image. At this time, if the video signal 321 is not a B picture, the motion compensation predicted picture corresponding to the reference picture input before another picture in time is always selected and output. Further, the prediction mode selector 406 selects coding in a picture in which prediction has not been completed, or selects predictive coding, according to the selected predicted picture, so that the selected coding has better coding efficiency.

At this time, when the video signal 321 is an I picture, the coding in the picture is always selected. When coding in picture is selected, a signal representing coded in picture mode is output as the predictive mode. Meanwhile, when predictive coding between pictures is selected, a signal representing the selected predictive picture is output as a predictive mode. When the prediction mode output from the prediction mode selector 406 is the coding mode in the picture, the switch 407 outputs a zero signal. If the prediction mode is not the encoding mode in the picture, the prediction mode selector 406 outputs a predicted picture.

As can be seen from the above, when the video signal 321 output from the storage circuit 301 is an I picture, the motion compensation prediction circuit 310 always outputs a zero signal as the predicted picture 323, and the I picture is not subjected to inter-frame prediction but is subjected to intra-frame prediction. Convert encoding. Simultaneously, when the video signal output from the storage circuit 301 is the 6th P picture in Fig. 6, the motion compensation prediction circuit 310 executes the motion compensation prediction from the 3rd I picture in Fig. 6, and outputs the predicted image 323 . Furthermore, when the video signal 321 output from the storage circuit 301 is the 4th B picture shown in Figure 6, then the motion compensation prediction circuit 310 executes the process from the 3rd I picture and the 6th P picture shown in Figure 6 The motion compensation prediction of , and output the predicted image 323 .

The operation of the transmission buffer 306 will be described below. The transmission buffer 306 converts the video data variable-length-encoded by the variable-length encoder 305 into a bit stream of an MPEG video signal. Herein, the MPEG stream has a six-layer structure, as shown in FIG. 5 . Add header information as identification codes for the sequence layer 621, GOP layer 622, picture layer 623, slice layer 624 and block layer 626 to form a layer structure.

Next, the transmission buffer 306 restores the bit stream of the video signal and the bit stream of the audio signal into a plurality of information packets respectively, and multiplexes these information packets to include a synchronization signal, thereby forming an MPEG2-PS system stream (program stream ). In this paper, MPEG2-PS is composed of program packet layer and information packet layer, as shown in Figure 8. Then, the header information is added to the information packet layer and the program packet layer. In a typical example, the system stream should contain data for one GOP of video data.

Here, the packet layer has a structure in which the upper layers of the packet layer are "bundled" together. Each packet layer constituting the program packet layer is called a PES packet. In addition, the header information of the packet layer shown in FIG. 8 also includes a packet identification signal and a synchronization signal constituting the basis of the video signal and audio signal.

Meanwhile, among the packets constituting the packet layer, there are three kinds of PES packets as shown in FIG. 9 . Here, the second-level packet shown on FIG. 9 is a "video/audio/private 1" packet in which the front part for identifying the packet and time stamp information or the like is paired with each information as header information A code for the parts (PTS and DTS) required for the packet to be decoded is prepended to the packet data. However, the time stamp information PTS is a time control information for reproduction output, and is information for controlling the decoding order of the stream of each pack at the time of reproduction. On the other hand, DTS is time control information at the start of decoding, and is information for controlling the transmission order of decoded data.

The third level packet shown in FIG. 9 is a "private 2" packet in which user-specific data is written. The lowest level packet is a "filler" packet in which all packet data is marked with a "1". The header information of the "Private 2" packet and the "Padding" packet is composed of the start code of the packet and the packet length.

As described above, video data and audio data are converted into the system stream of MPEG2-PS through the transmission buffer 306, and this conversion is performed for each of the frame vectors. This information is error corrected, and at the same time, modulated to minimize interference between the codes on the disc, and recorded onto the disc 212 . At this time, for example, the amount of data per GOP unit is set to be approximately equal. It is thus clear that the editing work for each GOP unit can be accomplished by distributing data to sectors equal to an integral multiple of the frame period.

The operation at the time of playback will be described below. During playback, the video information recorded on the optical disc 212 is amplified by a playback amplifier 213 . The information is then restored to digital data at the demodulator 214 and error correction decoder 215, and then at the frame sector inverse conversion circuit 216 to pure raw video data, without elements such as address and parity. data. The data is then input into the data expansion circuit 217 whose composition is shown in FIG. 4 . A system stream composed of MPEG2-PS is input to the transmission buffer 501 .

At the transmission buffer 501, the input system stream is restored into a package unit. Then, each PES packet is restored according to the header information, thereby reorganizing the bit stream of video data and audio data restored in the PES packet unit. Furthermore, concerning video data, the bit stream is restored to the block layer shown in FIG. 5, thereby restoring and outputting block data and motion vector data.

The block data output from the transmission buffer 501 is sent to the variable-length decoder 502 to make the variable-length data into fixed-length data, and then it is reversed and summed by reverse DCT, and then output to the adder 506 . At the same time, the predictive data decoding circuit 505 decodes the predictive image from the transmission buffer 501 according to the motion vector, and then outputs it to the adder 506 .

In this case, the prediction data decoding circuit 505 provides a frame memory for storing the I picture and P picture data decoded by the adder 506, as in the motion compensation prediction circuit 310. Incidentally, as to the method of updating the reference image data, no explanation will be given because the method is the same as the method of encoding the data.

The adder 506 adds the output of the predictive data decoding circuit 505 and the output of the inverse DCT circuit 504 , and outputs it to the storage circuit 307 . Here, when data is coded, temporally continuous video information in frames is rearranged in accordance with the data coding order shown in FIG. 7 . Therefore, in the memory circuit 507, the data input in the order of FIG.

In the following, the process of image search and its high-speed playback will be described by taking the case where data with such a coding structure is recorded on an optical disc. When the coding structure is as shown in Figure 6, high-speed playback of images can be completed, and at this time the data is played back in I picture units. In this case, track jumping can be performed immediately after the playback of the I picture. Then, the subsequent or previous GOP is accessed, whereupon the I picture is reproduced. For the case shown in FIG. 6, high-speed forward playback and reverse playback can be realized by repeating this process.

However, since the GOP rate is a variable bit rate, it is impossible to fully distinguish where the front of the following GOP is located. Therefore, the optical head should be properly jumped to the position at the front of the GOP. Thus, it is impossible to determine which track should be accessed.

In addition, the I picture has a large amount of data, so when only the I picture is reproduced in a continuous manner, such as a dedicated reproduction, the picture cannot be reproduced at a frequency of 30 Hz like a general moving picture, because the Disk read speed is limited. Even if the optical head jumps after completing the playback of the I picture, its work also lacks stability because the gap for re-jumping to the next I picture is long.

A conventional digital video signal recording and reproducing apparatus is constructed as described above. In the case of high-speed playback at several times the speed of the I picture and the P picture, the I picture data and the P picture data are detected by the GOP in the bit stream recorded from the recording medium (such as an optical disc, etc.). After the front part is read. Therefore, when the amount of data of the I picture and the P picture is very large, or when it takes a long time to search the front of the GOP, the time for reading data from the recording medium is not enough, so there will be an unreadable There is a problem that high-speed reproduction cannot be realized by taking all the data of the I picture and the P picture.

In conventional recording and reproducing apparatuses of digital video signals, even when high-speed reproduction is performed only by using I pictures, it takes a lot of time to input I picture data if there is a large amount of data. Therefore, dedicated playback exceeding several tens of times cannot be realized. At this time, higher-speed dedicated playback can be realized by playing back only one I picture for several GOPs. The problem is that the interval for updating the playback image will increase and thus the content of the image will become blurred.

Since the recording and reproducing apparatus of the conventional video signal is coded in the above-mentioned manner, only the I picture having a large amount of data is decoded at the time of skipping the search (observing the data by fast playback). Therefore, it is possible to make the optical head jump without replaying enough data for decoding. Otherwise, when a sufficient amount of data is played back, the playback time of the data will be lengthened, and the position to which the GOP will jump must be set at a relatively far place, thereby causing the problem of a reduction in the number of pictures output on the screen.

In addition, since the sector address of the subsequent GOP cannot be recognized due to the variable rate, it is difficult to discriminate whether the front part of the GOP is located on the track reached by the "jump". Therefore, there is a problem that the front portion of the GOP needs to be positioned on the target track for many times of optical disk rotation, so that the output of picture data on the screen is too small during dedicated playback. Moreover, it also brings such a problem that even if the sector can be identified, there is no way to judge how much data can be replayed when the bald head jumps, so that no judgment can be made before passing through the video decoder, so the bald head The effect of jumping is reduced.

As other conventional recording and reproducing apparatuses for digital video signals, apparatuses such as those proposed in Japanese Patent Application Laid-Open No. 6-98314(1994), No. 6-78289(1994) and the like have been proposed. Figure 10 shows an example. Wherein the label 775 is a video signal generator, such as a video camera, VTR, etc.; 776 is an audio signal generator, such as a microphone, VTR, etc.; 762 is a video signal encoder; 763 is an audio signal encoder; 777 is a system layer bit stream Generator; 778 is an error correction encoder; 779 is a digital modulator; 780 is a disc; 756 is a playback amplifier; 786 is a detector; 781 is a digital demodulator; 758 is an error corrector; 759 is a system stream processor ; 782 is a video signal decoder; 783 is an audio signal decoder; 784 is a monitor; 785 is a speaker.

At present, the diameter of the commonly used optical disc is 120mm, and this optical disc can generally record data of 600M bytes or more. Recently, such optical discs can record video signals and audio signals for up to 74 minutes at a data rate of about 1.2 Mbps. At the time of data recording, a video signal is input from the video signal generator 775 to the video signal encoder 762 to encode the video signal. An audio signal is input from the audio signal generator 776 to the audio signal encoder 763 to encode the audio signal. The process of multiplexing headers etc. into these two coded signals is performed by the system layer bit stream generator 777 . After the error correction code is added by the error correction encoder 778, the error correction signal is digitally modulated by the digital modulator 779 to generate a bit stream for recording. A master disc with a recording device (not shown) is produced from the bit stream, and the contents of the master disc are copied to the optical disc 780, thus obtaining a commercially available floppy disc.

In the playback device for the user, the playback amplifier 756 is used to amplify the information from the video floppy disk of the optical disc, and then the playback signal is sent to the demodulator 786 . After the emphasis signal is detected in the detector 786, it is digitally demodulated by the digital demodulator 781, and then the error is corrected by the error corrector 758. Afterwards, the video signal part is taken out from the error-corrected signal, and the data taken out is decoded in the video signal decoder 782, and then together with the audio signal decoded by the audio signal decoder 783, are respectively output to the monitor 784 and the loudspeaker 785 in.

Typical methods for encoding such video signals are MPEG1 and MPEG2 called the MPEG (Moving Picture Private Grouping) method, which is an international standard encoding method. A specific example of the encoding method can be illustrated by the example of MPEG2.

Fig. 11 is a block diagram showing a video signal encoding section in a conventional digital signal recording and reproducing apparatus for explaining the encoding method of MPEG2. Fig. 12 is a block diagram of a video signal decoding unit in a conventional digital signal recording and reproducing apparatus for explaining a decoding method. Fig. 13 shows the concept of moving picture processing for video signal encoding in a conventional digital signal recording and reproducing apparatus for explaining the grouping of moving pictures according to the encoding method of MPEG2. Referring to Fig. 13, among the marks such as IBBPBBP... among the figure, I represents an I image, B represents a B image, and P represents a P image. For example, in FIG. 13A, in a limited number of frames, moving pictures from one to the immediately preceding one where another I appears are grouped. The number of frames of images constituting this group is usually 15 frames in most cases. However, the number is not limited to any particular number.

A GOP, the group of pictures making up the group, comprises at least one I-picture which can be completely decoded in one frame. The GOP also includes a P picture and a B picture, the P picture is coded based on the I picture by motion-compensated prediction by temporal system unidirectional prediction, and the B picture is encoded based on the I picture Images and P-pictures are coded by bidirectional prediction of the time system. Incidentally, the arrows in Fig. 13A and Fig. 13B show the prediction relationship.

In other words, B-pictures can only be encoded and decoded after I-pictures and P-pictures are ready. The first P picture in a GOP can be coded and decoded after the I picture preceding the P picture is ready. The second P-picture and subsequent P-pictures can be encoded and decoded when the P-picture immediately preceding the P-picture is ready. Therefore, neither the P picture nor the B picture can be encoded and decoded without the I picture.

Referring to Fig. 11, wherein reference numeral 787 is an image rearranger, 788 is a scan converter, 789 is an encoder buffer, 790 is a mode decision device, 702 is a motion vector detector, 706 is a subtractor, and 708 is a DCT circuit, which includes field memory, frame memory, and DCT calculator. Reference numeral 710 is a quantizer, 714 is an inverse quantizer, 761 is an inverse DCT circuit, 718 is an adder, 720 is a video memory, 722 is a rate controller, and 726 is a variable length encoder.

Referring to Fig. 12, reference numeral 733 is a variable length decoder, 736 is an inverse DCT circuit, 737 is an image memory, 788 is an adder, and 739 is an inverse scan converter. Incidentally, the motion vector detector 702 combined with the mode decider 790 represents a motion vector detection unit.

Next, the operation of the digital video signal recording and reproducing apparatus will be described with reference to FIGS. 11 to 13. FIG. Referring to FIG. 11, the image rearranger 787 rearranges the images so as to be encoded in the order shown in FIG. Scan converter 788 then converts the scan process from raster scan to block scan. This rearranging of images and conversion from raster scan to block scan is often referred to as preprocessing. And the image rearranger 787 and the scan converter 788 are collectively referred to as a preprocessor. The input image data is subjected to block scanning in encoding order. The picture passes through subtractor 706 when the picture is an I picture. When the picture is a P picture or a B picture, the picture is subtracted from the reference picture in the subtractor 706 .

At this time, the motion vector detector 702 determines the direction of motion and the amount of motion (the input of the original image sent to the motion vector detector 702 can be an image after the rearrangement of the image or an image after the original image block scanning image, and the circuit size is less in the latter case. In addition, the reference image must be imported from the image memory 720, and the arrow in the figure has been omitted), so this area can be read out from the image memory 720 Signals in , where direction and magnitude parts are taken into account. At this time, whether to use bidirectional prediction or unidirectional prediction is determined by mode determiner determination 790 .

A subtraction from the reference screen takes place at a subtractor 706 taking into account the motion vector. Even if the formed image has only a small electric power, its coding efficiency is high. The output of the subtractor 706 can be collected into either a field unit or a frame unit of the DCT circuit 708, thereby performing DCT processing and converting into data in a frequency component. This data is input into quantizer 710, where there is a different weight for each frequency. Scan the data in a zigzag manner on two scales of low frequency components and high frequency components, and make length coding and Huffman coding.

This length-encoded and Huffman-encoded data is controlled to be variable-length encoded so that the quantization table is scaled by the rate controller 722 so that the data conforms to the magnitude of the target code. Data that has been variable-length encoded is usually output via encoder buffer 789 . The quantized data is returned to the inverse quantizer 714, and then restored to the data in the original image space by the inverse DCT circuit 716, and finally the data in the original image space and the data referenced by the subtractor 706 are added by the adder 718, Thus, the same data as the decoded data is obtained.

Fig. 12 is a block schematic diagram of a decoder structure. The variable length decoder 733 decodes the image data including header information (such as motion vector, encoding method, image mode, etc.). After the decoded data is quantized, inverse DCT calculation is performed by the inverse DCT circuit 736 (by the way: the inverse quantizer located in the preceding stage of the inverse DCT circuit 736 is omitted in FIG. 12). Referring to the picture data from the video memory 737 and considering the motion vector, the picture data which has been collated is added to the inverse DCT data by the adder 738, thereby decoding the motion compensation prediction. The data is converted into a raster scan by an inverse scan converter 739, whereby an interlaced image is obtained and output.

In addition, according to the variable transmission rate disk system introduced in the article "Variable Transmission Rate Disk System and Code Size Control Method" published by Sugiyama et al. at the 1994 Annual Conference of the Television Industry, a higher quality digital Recommendations for video signal encoding methods. This method uses a program (such as the first set of programs) to make the encoding rate fixed, so each GOP is set at a certain rate, which is related to the difficulty of design and encoding. Fig. 14 is a block diagram of a video signal encoding unit in a conventional digital signal recording and reproducing apparatus. Wherein, reference numeral 791 is a motion compensation predictor, 792 is a code quantity memory, 793 is a GOP rate setting unit, 794 is a code quantity assignment unit, 795 is a subtracter, 796 is a code quantity counter, and 797 is a switch. The GOP setting unit 793 in FIG. 14 makes settings to change the setting of the quantization value according to the degree of difficulty of the design proposal. In other words, when the switch 797 is connected to the dummy coding side, the output of the variable length coder 726 is sent to the code size counter 796 to calculate the code size to be stored in the code size memory 792 .

The GOP rate setting unit 793 determines the virtual code size in the entire program according to the code size stored in the code size memory 792, so as to set and calculate the optimal coding rate in each GOP. At this time, the assignment of the code is calculated by the code quantity assignment unit 794 for real coding prediction. When the switch 797 is connected to the real coding side, the value of the assigned code amount is compared with the value of the code amount counter 796, so the quantizer 710 can be controlled by operating the switch 797 according to the real code amount. In this way, what is easy to design is given a small code size, and what is difficult is given a large code size, so that the gradually changing coding difficulties in the program are overcome. Therefore, it is reported that the image quality recorded at 3 Mbps by this method is close to the image quality encoded at 6 Mbps.

In consideration of the possibility of performing a skip search in a recording and reproducing apparatus of a digital video signal using an optical disc, when reproducing I pictures and P pictures so as to be accessed at high speed even at the front of the GOP, fast When rewinding, the P picture is positioned at an appropriate position of the GOP, so when searching for data on the bit stream, the optical head needs to be manipulated. However, this control cannot be completed on time, because the servo mechanism (such as actuator, etc.) has a certain time constant. A GOP usually includes 15 frames of images, and in the NTSC scan mode, 0.5 second can be used to find the front of the GOP. However, in order to detect the front of a certain GOP, it is necessary to read 1/2 or more of the bit stream at a certain frame rate even when trying to read only the I picture or the P picture when skipping the search. , so that 1/3 of the image is read out, as a result, when the optical head moving time is set at 200 milliseconds, the reading speed is set at 2.5 times faster than the usual speed or higher. This exceeds the response limit of the actuator. With the usual replay method, it is practically impossible to perform a skip seek.

For conventional digital video recording and playback equipment, the signals are programmed in this way. Therefore, when trying to perform a skip search as in a video tape recorder, it is impossible to obtain a perfect playback image. At this time, the playback data cannot be extracted from a picture data entry such as a B picture. Get all original images. In particular, during skip seeks, mutations (abnormal shifts) occur, which affect the processing of the output in frame units. When variable-rate reading is carried out with high-quality playback images, there will be a problem that the difficulty of accessing the front of the GOP increases. This is because the address position of the front of the GOP changes. A space is formed in the disk area due to the GOP having uneven cells.

Contents of the invention

The purpose of the present invention is to provide a kind of recording and reproducing equipment of digital video signal, it can carry out special-purpose reproduction with the I picture with large amount of data, and can obtain high-quality reproduction image, gives simultaneously Its recording and playback methods.

Another object of the present invention is to provide a recording and reproducing device of a digital video signal, which can carry out high-speed reproduction with I pictures and P pictures with a large amount of data, and can obtain high-quality reproduction images. The image, and its recording and playback methods are given at the same time.

Another object of the present invention is to provide a recording and reproducing apparatus for digital video signals, which can improve the access performance of GOP under the situation of adopting conceivable variable bit rates, thereby obtaining good skip search, and at the same time Give its recording and playback methods.

Another object of the present invention is to provide a recording and reproducing apparatus of a digital video signal, which can improve the access performance of the GOP under the conceivable variable bit rate coding situation, and effectively utilize the memory on the storage medium. Spatial area, thereby accomplishing the skip search, and at the same time giving its recording and replaying methods.

A method for recording an optical disc according to the present invention is used for recording on an optical disc a digital video signal encoded by using motion compensation prediction and DCT, said method comprising the following steps:

A system stream is arranged, the system stream including video packs formed of video blocks each including an I picture as an intra-coded picture, and a dedicated pack having a stream ID of a dedicated stream , the picture of the video data that is formed as the P picture of unidirectional predictive coding picture and the B picture as bidirectional predictive coding picture;

arranging the private information packets such that the video data blocks corresponding to the private information packets follow the private information packets; and

Address information corresponding to sectors of video blocks before and after the video block corresponding to the private pack is recorded in the private pack.

A kind of optical disc recording apparatus according to the present invention, is used for recording on the optical disc by using the digital video signal of motion compensation prediction and DCT encoding, said apparatus comprises:

A device for arranging a system stream including video packets formed of video blocks each including an I picture as an intra-coded image, and a dedicated packet having a stream ID of a dedicated stream, A picture of video data composed of a P picture which is a unidirectional predictive coded picture and a B picture which is a bidirectional predictive coded picture; and

a device for arranging private information packets so that said video data blocks corresponding to said private information packets follow said private information packets, and recording address information in said private information packets, the address information Corresponding to the sectors of the video data block before and after the video data block corresponding to the private pack.

A method of reproducing an optical disc according to the present invention for playing an optical disc comprising digital video signals encoded by using motion compensation prediction and DCT, wherein:

A system stream including video packs formed of video blocks each including video packets as intraframe coded images and private packs having stream IDs for private streams is arranged in the optical disc. A picture of video data composed of an I picture, a P picture as a unidirectional predictive coded picture, and a B picture as a bidirectional predictive coded picture;

arranging private information packets corresponding to video data blocks such that video data blocks corresponding to said private information packets follow said private information packets;

recording address information corresponding to sectors of video data blocks before and after the video data block corresponding to the private information pack in the private information pack; and

The reproducing method includes the step of: when reproducing the video data, referring to addresses of sectors corresponding to the video data blocks before and after the video data block recorded in the dedicated information pack information.

An optical disc reproducing apparatus according to the present invention for playing an optical disc comprising a digital video signal encoded by using motion compensation prediction and DCT, wherein:

A system stream including video packs formed of video blocks and private packs having a stream ID of a private stream each including a video data block consisting of an image coded as an intraframe A picture of video data composed of an I picture, a P picture as a unidirectional predictive coded picture, and a B picture as a bidirectional predictive coded picture;

arranging private information packets corresponding to video data blocks such that video data blocks corresponding to said private information packets follow said private information packets;

recording address information corresponding to sectors of video data blocks before and after the video data block corresponding to the private information pack in the private information pack; and

Said playback means includes: for obtaining address information corresponding to sectors of said video data block before and after said video data block recorded in said dedicated information pack, to obtain the address by referring to information, a device for replaying the video data.

According to the recording and reproducing apparatus of the digital video signal of the present invention, when the video signal is recorded on the GOP unit, one frame thereof is divided into n areas related to the I picture, so that each area is divided into n areas related to the I picture, so that each area is divided into n areas from the area located at the center of the screen. The starting sequence is coded and recorded at the front of a GOP. At the same time, the address information of each area of the I picture is also recorded as header information. At the time of exclusive playback, only the data of the I picture located in the central partial area of the screen is read out, and an exclusive playback image is output in which certain values related to the data in the unread data area are hidden. Therefore, exclusive playback can be performed at a faster speed than in the case of playing back all I pictures having a large amount of data.

In the recording and reproducing apparatus of the above-mentioned digital video signal, the individual video images are output by expanding the central area so as to be read out over the entire screen. Therefore, since the data in the center portion of the screen is expanded and integrated into the reproduced image, the area where the data cannot be read is not conspicuous, so that the reproduced image looks comfortable.

According to the recording and reproducing apparatus of the digital video signal of the present invention, its video signal is recorded on the GOP unit, and one frame is divided into n areas related to the I picture, so that each area starts from the central part of the screen. The sequence is coded and recorded at the front of a GOP. When reading and playing back video signals from a recording medium such as an optical disc (the address information of each area in the I picture is recorded as header information at the same time), if special playback is performed at this time, only the central part of the screen I picture data in the area is read. On the other hand, the area where no data is to be read is hidden in the form of data with a certain numerical value hidden when the dedicated playback image is output. Therefore, the exclusive playback can be realized at a faster speed than the case of playing back all the I pictures.

In the recording and reproducing apparatus of the above-mentioned digital video signal, by extending the read-out center portion area to the entire screen, an image for exclusive reproduction can be output. Therefore, the area where data cannot be read is not conspicuous, and the reproduced image looks comfortable.

According to another recording and reproducing device of digital video signal of the present invention, when the video signal is recorded on the GOP unit, its one frame is divided into n areas, so that each area is in order from the partial area located at the center of the screen Encoded and recorded at the front of a GOP. At the same time, the address information of each area of the I picture is also recorded as header information. At the time of exclusive playback, only the data of the I picture in the area unit is read, and the data is output as a playback image. When the entire area of the I picture cannot be read within a certain period of time, an exclusive playback picture can be output by interpolating the picture with the data on the previous screen. Therefore, the area of the center portion of the screen is given priority for reproduction, and as a result the reproduced image after interpolation looks comfortable.

According to still another recording and reproducing apparatus of digital video signal of the present invention, when video data is recorded on the GOP unit, one frame thereof is divided into n areas so that each area is coded and Recorded at the front of a GOP. At the same time, the address information of each area of the I-picture cannot be recorded as header information. During dedicated playback, only the data of the I picture in the unit of this area is read out, and the data in each area 1, 2, ... n is read out one by one from continuous n pictures. Range, as a result, images of one screen portion can be integrated and output as a playback image. When the entire area of the I picture cannot be read within a certain period of time, an exclusive playback picture is output by interpolating the picture with the data of the previous screen. Therefore, the area located in the center portion of the screen is given priority for playback. Since n pieces of I pictures are synthesized into one screen, the interpolated playback picture is not obvious.

According to another digital video signal recording and reproducing apparatus of the present invention, when video pictures are recorded on GOP units, one frame thereof is divided into n areas related to I pictures and each area is encoded. When I-pictures are comprehensively recorded for each area at the front of a GOP, the position of the area initially recorded in GOP units is "scrolled" for recording. At the same time, the address information of each area of the I picture is also recorded as header information. In exclusive playback, only the I picture data in the area unit is read and output as a playback image. When all the I pictures cannot be read out within a certain period of time, an exclusive reproduction picture is output by interpolating the picture with the data of the previous screen. Thus, the location of the area is "scrolled" in GOP units, so that one screen of images can be played back in a smooth manner.

According to another digital video signal recording and reproducing apparatus of the present invention, when video data is recorded on the GOP unit, one frame thereof is divided into n areas related to I pictures and each area is encoded, and divided into An error correction block unit, and recorded at the front of a GOP in order starting from the area located at the center portion of the screen. At the same time, address information of each area of the I picture is also recorded as header information. In exclusive playback, only I picture data in error correction block units is read and output as a playback image. When the entire area of the I picture cannot be read out within a certain period of time, an exclusive playback image is output by interpolating the image with the data of the previous screen. Thus, since the area located in the center portion of the screen is given priority for reproduction, the reproduced image looks comfortable.

According to another recording and reproducing apparatus of digital video signal of the present invention, when picture data is recorded on the unit of GOP, its one frame is divided into n areas related to I picture and P picture and for each The regions are coded and recorded at the front of a GOP in order starting from the region located at the center portion of the screen. At the same time, address information of each area of the I picture and the P picture is also recorded as header information. In exclusive playback, the I picture and P picture in the area unit are read out and output as playback pictures. When the entire area of an I picture or a P picture cannot be read out within a certain period of time, an exclusive reproduction picture is output by interpolating the picture with the data of the previous screen. Thus, since the area located in the center portion of the screen is given priority for reproduction, the interpolated reproduced image looks comfortable.

According to the recording and reproducing apparatus of the digital video signal of the present invention, when the video signal is recorded on the unit of the GOP, one frame thereof is divided into n areas related to the I picture and the P picture and each area is encoded, And are recorded at the front of one GOP in the order starting from the area located at the central portion of the screen. At the same time, address information of each area of the I picture is also recorded as header information. During dedicated playback, only the data of the I picture and the P picture in the area unit are read out, and the data of each area 1, 2...n are read out from the continuous n pieces of I pictures and P pictures. range in order to synthesize an image of a screen portion and output it as a playback image. When the entire area of an I picture or a P picture cannot be read out within a certain period of time, an exclusive reproduction picture is output by interpolating the picture with the data of the previous screen. Thus, since the area located in the center portion of the screen is given priority for reproduction, the interpolated reproduced image is not conspicuous.

According to another recording and reproducing apparatus of digital video signal of the present invention, when video signal is recorded on the unit of GOP, its one frame is divided into n areas related to I picture and P picture and the frame unit Each region code in . When the split frame is fixed for each area at the front of a GOP, the position of the area to start recording in the frame unit is "rolled". At the same time, the address information of each area in the I picture is also recorded as header information. In exclusive playback, only the I picture data in the area unit is read and output as a playback image. When all the I pictures cannot be read out within a certain period of time, an exclusive playback picture is output by interpolating the picture with the data of the previous screen. Thus, the recording order of the I picture and P picture areas is "scrolled" in GOP units, so that a reproduced image of one screen portion can be reproduced in a smooth manner.

According to another recording and reproducing apparatus of digital video signal of the present invention, when video signal is recorded on the unit of GOP, its one frame is divided into n areas related to I picture and P picture and for each area Encoding, and separate the error correction block unit. The divided frames are then recorded at the front of one GOP in order starting from the partial area located at the center of the screen. At the same time, address information of each area of the I picture is also recorded as header information. At the time of exclusive playback, only the data of the I picture in the error correction unit is read out and output as a playback image. When all the I pictures cannot be read out within a certain period of time, an exclusive playback picture is output by interpolating the picture with the data of the previous screen. Thus, since the area located in the center portion of the screen is given priority for reproduction, the reproduced image after interpolation looks comfortable.

According to the playback method (equipment) of a digital video signal, at least the intra-coded I picture is separated according to the frequency domain, quantization level or spatial resolution to form a bit stream of video data, wherein the separation involves at least I More important data such as images among the data of images are arranged at the front. And the address information in the divided data is arranged as header information at the front of the bit stream of the video data to form a packet. The data recorded on the recording medium must be rearranged in the order of the data according to the header information in the packet to be output before the data is divided when it is normally reproduced. When the dedicated playback is performed, the data arranged at the front is decoded and used as an output for the dedicated playback. Thus, by dividing the data by frequency domain, quantization level or spatial resolution, the data to be accessed at the time of exclusive playback is reduced, so that a smooth image for exclusive playback can be obtained. In addition, since the address of the divided data as header information of the system stream is recorded, it is possible to know the number of bytes that should be reproduced immediately at the time of reproduction, so that the optical head can jump efficiently in the exclusive reproduction. Also, at the time of normal playback, data is rearranged according to the address, so that a disadvantage caused by dividing data at the time of playback can be avoided.

According to the method (equipment) for recording and reproducing digital video signals, at least the intra-coded I pictures are separated according to frequency domain, quantization level and spatial resolution to form a bit stream of video data, wherein at least More important data such as pictures among data related to I pictures are arranged at the front. And the address information in the divided data is arranged as header information at the front of the bit stream of the video data to form a packet. The data recorded on the recording medium must be rearranged in the order of the data according to the header information in the packet to be output before the data is divided when it is normally reproduced. When the dedicated playback is performed, the data arranged at the front is decoded and used as an output for the dedicated playback. Thus, by dividing the data by frequency domain, quantization level or spatial resolution, the data to be accessed at the time of exclusive playback is reduced, so that a smooth image for exclusive playback can be obtained. In addition, since the address of the divided data as header information of the system stream is recorded, it is possible to know the number of bytes that should be reproduced immediately at the time of reproduction, so that the optical head can jump efficiently in the exclusive reproduction. Also, at the time of normal playback, data is rearranged according to the address, so that a disadvantage caused by dividing data at the time of playback can be avoided.

According to another recording and reproducing method (equipment) of a digital video signal, at least the I picture coded in the frame when recording is divided into n areas (n>1), thereby rearranging the divided into n areas in units of areas. I-pictures of three areas to form a bit stream of video data, in which the area located in the center of the screen is arranged at the front. Then, the address information of the I picture which has been divided into n areas is arranged at the front of the data bit stream to constitute a pack, and recorded on the recording medium. At the time of normal playback, the I picture data in the area unit is rearranged according to the header information arranged at the front of the pack, and output. In the case of exclusive playback, the exclusive playback can be accomplished by outputting only the data of the I picture that can be read from the front of the pack within a certain period of time. Thus, the data that should be accessed at the time of exclusive playback is reduced by dividing the data in the screen area at the time of recording. Since the address of the divided data is recorded as the header information of the system stream, it is possible to know the number of bytes that should be replayed immediately during replay, so that the optical head can effectively jump during dedicated replay, thereby in a certain time The address jump can be done in the unit. In addition, since the data is rearranged according to the address at the time of normal playback, as a result, disadvantages due to data division at the time of recording and playback can be avoided.

According to yet another recording and reproducing method (equipment) of a digital video signal, the I picture data through intra-frame coding is at least divided into n areas (n>1), thereby rearranging the divided n areas in units of areas I picture to form a bit stream of video data in which the area located at the center of the screen is arranged at the front. Then, the address information of the I picture which has been divided into n areas is arranged at the front of the video data bit stream as header information to constitute a packet. During normal playback, the I picture data rearranged according to the header information arranged at the front of the pack is rearranged in units of the area, and output from the recording medium on which the data is recorded. In exclusive playback, the exclusive playback is accomplished by outputting only data that can be read within a certain period of time. Therefore, by dividing the data in the screen area at the time of exclusive playback, jumping of addresses can be completed within a certain time unit, resulting in a reduction of data to be addressed at the time of exclusive playback. Furthermore, since the address of the divided data is recorded as the header information of the system stream, the number of bytes to be reproduced at the time of playback can be detected instantly, so that the optical head can efficiently jump at the time of exclusive playback. In addition, since the data is usually rearranged according to the address during playback, there is no disadvantage caused by division of data during data playback.

According to yet another method (apparatus) for recording and reproducing a digital video signal, at least an I picture that is intra-coded at the time of recording is divided into a low-frequency area, a high-frequency area, a quantization level, and a spatial resolution so that images from the divided The basic data of the I picture is rearranged in the unit of each area on the screen to constitute a bit stream of video data, wherein the area located at the central part of the screen of the I picture is arranged at the front. The divided area, the divided data, and the address information of the image are arranged at the front of the bit stream of the video data as header information, thereby constituting a pack, and recorded on the recording medium. At the time of normal playback, the data in the area unit is rearranged according to the header information arranged at the front part of the packet, and then the data is output. The divided data is rearranged in the order of the original data. In exclusive playback, only the I picture data from the front of the pack that can be read within a certain period of time is output for exclusive playback. Therefore, at the time of recording, data is divided according to frequency, quantization and spatial resolution, and divided into units of screen areas. Thus, at the time of private playback, the data to be accessed is reduced, so that a smooth private playback can be obtained by gradually reducing the amount of data to be accessed at the time of private playback. In addition, since the address of the split data is recorded as header information, and the number of bytes to be reproduced can be detected immediately during reproduction, jumping of the optical head can be effectively performed during exclusive reproduction. Furthermore, involving division of data in a plurality of divisions, the amount of data to be read out can be adjusted according to the specific playback speed, so that a wide range of specific playback speeds can be accommodated. In addition, since data is rearranged according to addresses during normal playback, there is no disadvantage of dividing data when playing back data.

According to another method (device) for reproducing a digital video signal, at least an intra-coded I picture is divided according to a low-frequency region, a high-frequency region, a quantization level, or a spatial resolution, so that Elementary data is rearranged in each area on the screen to constitute a video data stream in which the area located at the central portion of the screen of the I picture is arranged at the front. The divided area, the divided data, and the address information of the image are arranged at the front of the bit stream of the video data as header information, thereby constituting a pack, and recorded on the recording medium. During normal reproduction, the data is reproduced according to the header information arranged in the front portion of the packet, thereby outputting the data from the medium. The divided data is rearranged in the order of the original data. In exclusive playback, only the I picture data from the front of the pack that can be read within a certain period of time is output for exclusive playback. Therefore, the data is partitioned according to frequency, quantization and spatial resolution and divided into units of screen area. Thus, the data accessed at the time of exclusive playback is reduced by dividing the data into panel units. In addition, since the address of the split data is recorded as header information, and the number of bytes to be reproduced can be detected immediately during reproduction, jumping of the optical head can be effectively performed during exclusive reproduction. Furthermore, involving division of data in a plurality of divisions, the amount of data to be read out can be adjusted according to the specific playback speed, so that a wide range of specific playback speeds can be accommodated. In addition, since data is rearranged according to addresses during normal playback, there is no disadvantage of dividing data when playing back data.

According to another recording and reproducing method (apparatus) of a digital video signal (or reproducing method (apparatus) of a digital video signal), only the area located at the center portion of the screen of the I picture is read out. For the data in the non-read area, the reproduced image is synthesized with the masked data as a limit value. Therefore, the dedicated playback can be realized at a higher speed than in the case of playing back all I pictures having a large amount of data.

According to another recording and reproducing method (apparatus) of a digital video signal (or reproducing method (apparatus) of a digital video signal), only the area located at the center portion of the screen of the I picture is read out. The reproduced image is integrated by extending the readout area over the entire screen. Therefore, compared with the case of reproducing the entire I picture having a large amount of data, the dedicated reproduction can be realized at a higher speed, so that the area where the data cannot be read is not conspicuous.

The recording device of the digital video signal of the present invention comprises: a first coding device, is used for coding the video signal containing the coded picture, and this coded picture contains at least intraframe coded by using motion compensation prediction and orthogonal transform An image of the coded digital video signal; the second encoding means is used to encode the remaining components through encoding with the video signal output of the first encoding means; the data arrangement means is used for each image group data in Each output data from the first and second encoding means is arranged at a predetermined position of each pixel data. By encoding a substantial portion of the moving picture at least the area to be accessed can be reduced compared to the case where the first encoding means encodes the entire video signal. The second encoding means encodes the video information not encoded by the first encoding means, so that the entire video information is encoded by the two encoding means. In addition, the data arrangement means rearranges the data obtained by the two encoding means so that the data can be easily accessed by the optical head. Thus, coding is done such that the amount of code that should be accessed is reduced, at least during dedicated playback. Thus, arrangement of data to be accessed can be efficiently performed at least at the time of exclusive playback.

In the recording apparatus of the above digital video signal, the video information is encoded to be thinner at a predetermined interval with respect to the video image including the encoded image including the encoded image in at least one frame. The first encoding means then encodes the thinned video image so that at least the area to be accessed is reduced. When accessing only the data of the first encoding means, the video image can be encoded such that the picture is sufficiently comprehensible for a human being after its decoding.

In the recording apparatus of the above-mentioned digital video signal, the first encoding means encodes only the low-frequency region of the orthogonal transformation. The first encoding means encodes image data belonging to a part of the frequency, thus at least reducing the area to be accessed. When only the data of the first encoding means is accessed, the video image can be encoded such that the picture is sufficiently comprehensible for a human being after its decoding.

In the recording apparatus of the above-mentioned digital video signal, the first encoding means coarsely quantizes at the quantization level to be encoded. The first encoding means encodes the upper-bit data which greatly affects the image by coarse quantization, thus at least reducing the area to be accessed without lowering the resolution. When accessing only the data of the first encoding means, the video image can be encoded such that the picture is sufficiently comprehensible for a human being after its decoding.

Another digital video signal recording device of the present invention extracts the data in the low-frequency component from a data row, the video signal in the data row is segmented with predetermined bits, and the video signal includes at least one frame Intra-coded, coded picture from a picture of a video signal coded using motion compensated prediction and orthogonal transform. The data is segmented by predetermined bits for each block, thereby segmenting the low-frequency region of the video signal. Therefore, it is easy to limit the code size to a fixed length. Furthermore, when performing decoding of data in the low-frequency region, the data can be encoded so that the content of the image can be roughly understood.

The playback device of the digital video signal of the present invention rearranges the data in the low-frequency region and the data in the high-frequency region to become a predetermined order, so that the manner of decoding the rearranged data or selectively selecting the data The way of decoding is in the low frequency region. During normal playback, a complete decoded image can be obtained by concatenating the coded data of two segments. During dedicated playback, only data in the low frequency region is decoded. Therefore, the data can be decoded according to the working state of the device, so that the obtained image can be roughly understood in its content.

In the above-mentioned playback device of a digital video signal, when data is decoded in such a manner that only data in a low frequency region is decoded, only data that can be decoded is decoded. Data that cannot be decoded near the edge of a predetermined number of bits are discarded, so that data in the high-frequency region is replaced by a fixed value for inverse orthogonal transformation. When decoding the low-frequency region from two segmentally encoded data at the time of dedicated playback, only the data that can be decoded is decoded and the bits that cannot be decoded are discarded. Decoding of abnormal data can thus be avoided. For the remaining high-frequency region, it is replaced by a fixed value and decoded, so that a decoded image can be obtained without distortion.

In another recording device for digital video signals, one end of the block code is added to the coded data of each block of the video signal, which includes at least one frame coded image from when obtained as a low frequency area A predetermined number of bits of data is an encoded image of an image of an encoded digital signal using motion compensation prediction and orthogonal transformation. The above coded data exceeding a predetermined number of bits is coded as high frequency range data. Both the low frequency region and the high frequency region of the information block are coded in such a way that the information block appears to terminate at the end of the information block (EOB) code. Therefore, when only low-frequency data is decoded, coded data that can be decoded can be obtained without redundant circuits such as for discarding data.

Another digital video signal recording apparatus of the present invention reassembles data based on data in a low frequency region, data in a high frequency region, and an EOB code. Then, a method of decoding the recombined data or a method of selectively decoding only the data in the low-frequency region is selected, so as to decode the coded data recombined according to the selection result. For the high-frequency region, replace its data with a fixed value to perform inverse orthogonal transformation. During normal playback, a completely decoded picture can be obtained from coded data segmented separately by EOB. During dedicated playback, only the data in the low-frequency region is decoded from the encoded data, so that whether it is normal playback or dedicated playback, it can operate according to the working status of the device, so we can get a picture that we can understand coarse image. In addition, when the low-frequency area is decoded from the coded data, the remaining high-frequency information block area is replaced with a fixed value and decoded, so that the area can be decoded to avoid data distortion. Both the high and low frequency regions of the block can be decoded as if the block appeared to end at the EOB.

Another digital video signal recording device of the present invention includes: a low-resolution encoding device, used to encode the data of the low-resolution component, and the pixels in this component are thinned out relative to the video signal, so Said video signal comprises a coded picture comprising at least one picture in a frame from a coded digital picture using motion-compensated prediction and orthogonal transformation; differential component coding means for interpolating low The output data of the encoding means of the resolution encodes the differential component of the image before pixel thinning; and the information adding means is used to divide the output of the encoding means of low resolution and the differential component encoding means into predetermined area to add an error correction code to form data. When spatially thinned image data is encoded and only such encoded data is accessed, the image data can be encoded so that the image is decoded so that its picture is sufficiently comprehensible for humans. The decoded data from the low-resolution encoding device is interpolated so that a differential component is obtained by comparing the image with the image before the low-resolution conversion, and thus cannot be obtained by the low-resolution encoding device A high-resolution portion of image data can be encoded. In this way, not only low-resolution image information can be coded.

Another digital video signal reproducing apparatus of the present invention integrates decoded low-resolution component data and differential component data. At the time of normal playback, the coded data of the low-resolution component is integrated with the coded data of the high-resolution component (which is a differential component between the low-resolution part and the data before being thinned out to low-resolution), and then Pictures with full resolution components can be decoded.

In the above-mentioned digital video signal reproducing apparatus of the present invention, the method of decoding the image by integrating the data of the low resolution and the data of the differential component is switched to a method of decoding only the low resolution component. At the time of normal playback, the coded data of the low-resolution component and the high-resolution component (which is the difference component between the data before being thinned out to low-resolution and the data of the low-resolution part) segmented into two The coded data is combined so that images with full resolution can be decoded. During dedicated playback, the decoding method is switched according to the working state of the device, so that the coarse image can be decoded only by decoding low-resolution encoded data.

In the reproducing apparatus of the above-mentioned digital video signal, when a picture of low resolution is decoded, only a picture interpolated after decoding is produced. At the time of exclusive playback, when only low-resolution coded data is decoded, the video data of the low-resolution component is interpolated to change the size of the picture back to its original size.

Another digital video signal recording device of the present invention includes: judging means for judging the degree of distortion when the image is encoded and decoded according to motion compensation prediction and orthogonal transformation; adaptive coding means for judging according to the judging means Output, by adaptively changing the rate, encode the data rate; the information addition device is used to add audio signals, additional information such as headers, and error correction codes together; the data rate setting device is used to set the discrete value to adaptively change the data rate. In a variable rate coding arrangement, the rate is limited to only a finite value. Therefore, the data rate information of the GOP (equivalent to the code amount of the GOP) can be represented with a small number of bits.

Another digital video signal recording device of the present invention includes: judging means for judging the degree of distortion when the image is encoded and decoded according to motion compensation prediction and orthogonal transformation; adaptive coding means for judging according to the judging means Output, by adaptively changing the rate, encode the data rate; information adding means, for adding audio signal, header and other additional information, and error correction code together, wherein the device is constituted to be available in optical head etc. The data rate information is multiplexed on the recording medium or written on a predetermined area of the recording medium. In addition to video data, when image data is encoded at a variable rate, the data rate setting information is also recorded on the recording medium, so that the data rate information can be read together, so when recording the information, it is also necessary to record the data rate setting information on the recording medium. The position of the predetermined GOP on the disc can be recorded immediately.

Another recording device for digital video signals includes: judging means for judging the degree of distortion when the image is encoded and decoded according to motion compensation prediction and orthogonal transformation; information adding means for adding additional information such as audio signals and headers , and an error correction code are added together; the first encoding means is used for encoding a video signal that has been thinned out at predetermined intervals with respect to a video signal, and the aforementioned video signal includes a frame containing an intra-coded picture The encoded image of the image; the second encoding means is used to encode the components of the remaining video signal after being encoded by the first encoding means and outputted, wherein said device is composed to make the first or The data rate of at least one of the second encoding means is adaptively changed and encoded. With variable rates, high-quality encoding can be achieved. In greatly increased rate GOPs, spatially thinned video data is coded in such a way that at least the area being accessed is reduced.

Another digital video signal recording device of the present invention includes: judging means for judging the degree of distortion when the image is encoded and decoded according to motion compensation prediction and orthogonal transformation; information adding means for adding audio signals, headers, etc. The additional information and the error correction code are added together; the first encoding means is used to encode only the low-frequency region that has been orthogonally transformed with respect to the video signal, and the video signal includes an intra-frame encoded image the coded image of the coded image; the second coding means is used for coding the components of the remaining video signal after the signal is coded by the first coding means, wherein said device is constituted to be able to use according to the judgment output of the designed judging means The data rate of at least one of the first or second encoding means is adaptively changed to encode the video signal. With variable rates, high-quality encoding can be achieved. For greatly increased rate GOPs, the video data in local frequency regions of each block is coded such that at least the area to be accessed is reduced.

Another recording device for digital video signals includes: judging means for judging the degree of distortion when the image is coded and decoded according to motion compensation prediction and orthogonal transformation; information adding means for adding audio signals, headers, etc. The information is added together with an error correction code; the first encoding means is used to encode the video signal by setting the coarse quantization at a quantization level with respect to the video signal; the video signal includes a frame containing coded picture of the coded picture; second encoding means for encoding the components of the remaining video signal after the signal is encoded by the first encoding means, wherein said video signal is judged according to the designed judging means The output is encoded by adaptively changing the data rate of at least one of the first or second encoding means. With variable rates, high-quality encoding can be achieved. In a GOP whose rate is greatly increased by means of a variable rate, the data in the upper bits, which greatly affect the picture, are encoded, and the encoding is performed so that at least the area to be accessed can be reduced.

Another digital video signal playback device of the present invention switches the playback mode between the normal playback mode and the dedicated playback mode to extract data rate information. When the dedicated playback mode is used, the location of the dedicated playback data on the recording medium is calculated based on the data rate information in the dedicated playback mode. When reproduction is performed at different data rates by extracting the data rate information of each GOP, in the case of normal reproduction, coded data divided into two are integrated and decoded. In exclusive playback, the position of the GOP on the recording medium to be accessed is calculated. Then, at least the data to be accessed is replayed, so that the next target GOP is accessed again. At this time, the location information of the GOP to be accessed on the recording medium is calculated to facilitate dedicated playback and obtain high-quality variable rates.

In the playback apparatus of the above digital video signal, the position of the optical head to the recording medium is controlled based on the result of the position calculation and the specific playback rate. The position information on the disc of the GOP that is the access target is calculated based on the specific playback rate. The position of the optical head can be controlled to the position of the target GOP according to the exclusive playback rate, so high-quality variable-rate playback can be performed at various speeds in the exclusive playback mode.

In another digital video signal recording apparatus of the present invention, the code amount is controlled so as to correspond to an area assigned to a picture group formed by a digital video signal encoded according to motion compensation prediction and orthogonal transformation. , the apparatus of the present invention includes: encoding means for encoding; code amount comparing means for comparing the output of the encoding means with a predetermined amount of data; and data feeding means for burying redundant data into a blank In the blank area of the image group of the area. In the case of encoding and recording data at a variable rate, the access time can be shortened by positioning the GOP at a position where the optical head can easily access, thereby increasing the data read out during data encoding and recording during dedicated playback quantity. In addition, identify unnecessary portions on the disc, such as blank portions, thereby using these features to improve image quality, or to help extend recording times with these features.

Another digital video signal playback device of the present invention includes data reconstruction means for reconstructing video signal coded data embedded in the original group of images, and data decoding means for decoding the data reconstructed by the data reconstruction means . Coded data in which other GOP data is embedded in blank portions can be reconstructed so that the data can be decoded without distortion. Also, at the time of specific playback, the amount of data still increases, and high-quality playback images can still be obtained.

Still another digital video signal reproducing apparatus of the present invention switches which of the three decoding means according to a specific reproducing speed. Among the three decoding devices, the first decoding device is used to decode the first and second coded data and obtain a reproduced image, and the second decoding device is used to decode the first coded data and obtain such a reproduced image. The image corresponds to a low-frequency region of an intra-coded image, the number of pixels is thinned out or coarsely quantized, and the third decoding means is used to decode the first coded data and obtain a reproduced image corresponding to In the low-frequency area of the inter-predicted picture and the intra-coded picture, the number of pixels becomes sparse, or coarsely quantized. Since mode switching is performed between a mode for decoding and displaying only I pictures and a mode for displaying I pictures and P pictures, I pictures and P pictures can be realized with slower dedicated playback (for example, double-speed playback). The dedicated playback of pictures results in precise dedicated playback without frame skipping compared to dedicated playback of only I pictures. Also in high-speed dedicated playback, various playback speeds such as dedicated playback I pictures can be handled.

Still another digital video signal playback device of the present invention includes video data extraction means for extracting data corresponding to video signals from playback codes, a map for decoding and playback of video data output from the video data extraction means. Like data decoding and playback means, and mode switching means. The mode switching means switches between a normal playback mode, a mode of playing back and displaying an odd field or an even field, and a mode of displaying an odd field or an even field by inverting a field structure. In exclusive playback, the field structure is optimized according to the mode. During reverse playback, the device works in reverse to give the display until the field is displayed. For example, during frame-skip playback such as fast rewinding, by outputting the same video image in even and odd fields, the number of fields is set at a certain value to obtain a dedicated playback image that is easy to watch.

Description of drawings

The above and other objects and features of the present invention will become more apparent from the following detailed description together with the accompanying drawings.

Fig. 1 is a block diagram of a conventional optical disc recording and reproducing apparatus.

Fig. 2 is a block diagram of a video signal encoding unit in conventional MPEG.

Fig. 3 is a block diagram of a conventional motion compensation prediction circuit.

Fig. 4 is a block diagram of a conventional MPEG video signal decoding unit.

FIG. 5 is a schematic diagram showing a data arrangement structure of a conventional MPEG video encoding algorithm.

Fig. 6 is a diagram showing an example of a GOP structure of a conventional MPEG video encoding algorithm.

7A and 7B are diagrams showing an example of a conventional MPEG video bit stream.

FIG. 8 is a diagram showing an example of a system stream in a conventional MPEG PS.

Fig. 9 is a diagram showing an example of a conventional MPEG PES packet flow.

Fig. 10 is a block diagram of a conventional digital signal recording and reproducing apparatus.

Fig. 11 is a block diagram of a video signal encoding unit in a conventional digital video signal recording and reproducing apparatus.

Fig. 12 is a block diagram of a video signal decoding unit in a conventional digital video signal recording and reproducing apparatus.

13A and 13B are diagrams illustrating the principle of dynamic image processing in a conventional digital signal recording and reproducing apparatus.

Fig. 14 is a block diagram of a video signal encoding unit in a conventional digital signal recording and reproducing apparatus.

15 is a block diagram of a recording system in the digital signal recording and reproducing apparatus according to Embodiment 1.

16 is a block diagram of a playback system in the digital signal recording and playback apparatus according to Embodiment 1.

Fig. 17 is a schematic diagram illustrating a macroblock.

FIG. 18 is a schematic diagram illustrating screen division.

Fig. 19 is a schematic diagram illustrating data arrangement.

20A to 20D are conceptual diagrams illustrating a playback method of dedicated playback.

21A to 21B are conceptual diagrams of a method of dedicated playback in the case of data expansion.

FIG. 22 is a schematic diagram illustrating screen division according to Embodiment 3. FIG.

FIG. 23 is a conceptual diagram illustrating data arrangement according to Embodiment 3. FIG.

24A to 24E are conceptual diagrams illustrating a method of performing exclusive playback according to Embodiment 3. FIG.

25A to 25B are schematic diagrams illustrating an arrangement of error correction blocks according to Embodiment 3. FIG.

26A to 26D are conceptual diagrams illustrating a method of performing exclusive playback according to Embodiment 4.

27A to 27F are schematic diagrams illustrating a method of specific playback in the case of data interpolation in Embodiment 4.

FIG. 28 is a conceptual diagram illustrating data arrangement in Embodiment 5. FIG.

29A to 29E are conceptual diagrams illustrating a playback method of dedicated playback according to Embodiment 5. FIG.

FIG. 30 is a conceptual diagram illustrating data arrangement in Embodiment 6. FIG.

31A to 31F are conceptual diagrams illustrating a method of performing exclusive playback in Embodiment 6. FIGS.

32A to 32B are diagrams illustrating the arrangement of error correction blocks in Embodiment 6.

33A to 33G are schematic diagrams illustrating a playback method in Embodiment 7.

34A to 34F are conceptual diagrams illustrating a method of performing specific playback in the case of performing data interpolation in Embodiment 7.

FIG. 35 is a conceptual diagram illustrating data arrangement in Embodiment 8. FIG.

36A to 36F are conceptual diagrams illustrating a method of performing exclusive playback in Embodiment 8. FIGS.

Fig. 37 is a block diagram of a digital video signal encoding unit in Embodiment 9.

Fig. 38 is a diagram showing the principle of frequency division in Embodiments 9 and 11.

Fig. 39 is a flowchart of digital video signal encoding processing in Embodiment 9.

Fig. 40 is a diagram illustrating a header of a bit stream in Embodiment 9.

41A to 41D are diagrams showing rearrangement of bit streams in Embodiment 9.

Fig. 42 is a diagram showing an example of address information of a system stream in Embodiment 9.

Fig. 43 is a block diagram of a digital video signal decoding unit in Embodiment 9.

Fig. 44 is a schematic diagram showing the principle of decoding processing in the ninth embodiment.

Fig. 45 is a flowchart of decoding processing in Embodiment 9.

Fig. 46 is a block diagram of a digital video signal encoding unit in Embodiment 10.

Fig. 47 is a block diagram of a digital video signal decoding unit in Embodiment 10.

Fig. 48 is a schematic diagram showing an example of a screen area in the tenth embodiment.

Fig. 49 is a diagram showing an example of a bit stream when video data is rearranged in units of picture areas in Embodiment 10.

Fig. 50 is a flowchart of digital video signal encoding processing in Embodiment 10.

Fig. 51 is a diagram showing an example of address information of a system stream in Embodiment 10.

Fig. 52 is a diagram showing an example of a system flow in Embodiment 10.

53A to 53E are diagrams showing examples of reproduced screens among reproducible images at the time of reproduction.

54A to 54E are diagrams showing examples of reproduced screens in which only the center portion of the screen is output at the time of reproduction in Embodiment 10. FIGS.

55A and 55b are diagrams showing examples of playback screens in which the central portion of the screen is enlarged and displayed in Embodiment 10. FIGS.

Fig. 56 is a flowchart of digital video decoding processing in Embodiment 10.

Fig. 57 is a block diagram of a digital signal encoding unit in Embodiment 11.

Fig. 58 is a flowchart of digital video signal encoding processing in Embodiment 11.

59A to 59D are diagrams showing examples of system flow in Embodiment 11.

Fig. 60 is a diagram showing an example of address information of a system stream in Embodiment 11.

Fig. 61 is a block diagram of a digital video signal decoding unit in Embodiment 11.

Fig. 62 is a flowchart of digital video signal decoding processing in Embodiment 11.

Fig. 63 is a block diagram of a digital video signal encoding unit in Embodiment 11.

Fig. 64 is a schematic diagram showing the principle of resolution transition on a screen in the twelfth embodiment.

FIG. 65 is a schematic diagram illustrating an example of data composition results in Examples 12, 13, and 14.

Fig. 66 is a block diagram of a digital video signal encoding unit in Embodiment 13.

Fig. 67 is a diagram illustrating an example of data arrangement of DCT coefficients within a DCT block.

Fig. 68 is a block diagram of a digital video signal encoding unit in Embodiment 14.

FIG. 69 is a diagram illustrating an example of statistics of encoded data in Embodiments 12, 13, and 14.

FIG. 70 is a diagram showing an example of the processing sequence in Embodiments 12, 13 and 14.

71A to 71D are diagrams illustrating distribution shapes of frequency components of bit streams of DCT blocks in one block in Embodiment 15. FIGS.

Fig. 72A is a block diagram of a digital video signal decoding unit in Embodiment 15.

Fig. 72B is a diagram illustrating the operation principle of digital video signal decoding processing in Embodiment 15.

Fig. 73A is a block diagram showing a digital video signal encoding unit in Embodiment 16.

Fig. 73B is a diagram illustrating the operation principle of digital video signal encoding processing in Embodiment 16.

Fig. 74 is a block diagram of a digital video signal decoding unit in Embodiment 16.

Fig. 75 is a block diagram of a digital video signal decoding unit in Embodiment 17.

Fig. 76 is a block diagram of a GOP address generator and an optical disk controller in Embodiment 18.

Fig. 77 is a block diagram of a GOP address generator including playback processing and an optical disk controller in Embodiment 18.

Fig. 78 is a block diagram of a digital video signal decoding unit when division by frequency and division by quantization are performed in Embodiment 19.

Fig. 79 is a block diagram of a digital video signal decoding unit when division by bit length is performed in Embodiment 19.

Fig. 80 is a block diagram of a digital video signal decoding unit when division by resolution is performed in Embodiment 19.

Fig. 81 is a block diagram of a digital video signal encoding unit in Embodiment 20.

Fig. 82 is a block diagram of a digital video signal decoding unit in Embodiment 20.

83A and 83B are diagrams illustrating the principle of processing with the digital video signal recording and reproducing apparatus in Embodiment 20.

Fig. 84 is a block diagram of a digital video signal decoding unit when division by frequency or division by quantization is performed in Embodiment 21.

Fig. 85 is a block diagram of a digital video signal decoding unit when division by bit length or division by quantization is performed in Embodiment 21.

Fig. 86 is a block diagram of a digital video signal decoding unit when division by resolution and division by quantization are performed in Embodiment 21.

Fig. 87 is a block diagram of a digital video signal decoding unit when division by frequency and division by quantization are performed in Embodiment 22.

Fig. 88 is a block diagram of a digital video signal decoding unit when division by bit length and division by quantization are performed in Embodiment 22.

89A and 89B are diagrams showing the principle of processing at the time of jump search in Embodiment 22.

90A and 90B are diagrams showing the principle of processing at the time of reverse playback in Embodiment 22.

Detailed ways

The present invention will now be explained in detail based on the drawings showing various embodiments.

First, Embodiment 1 of the present invention is explained. FIG. 15 is a block diagram showing the recording system of the digital video signal recording and reproducing apparatus in Embodiment 1. FIG. Referring to FIG. 15 , the digital video signal output from the input terminal 1 is input to the formatting circuit 3 . The electrical video signal output from the formatting circuit 3 is input to the first input terminal of the subtractor 4 and the second input terminal of the motion compensation prediction circuit 11 . The output signal of the subtractor 4 is input to the quantizer 6 via the DCT circuit 5 . The output signal of the quantizer 6 is input to the first input terminal of the buffer memory 12 via the variable length coder 7 . At the same time, the output signal of the quantizer 6 is also input into the inverse DCT circuit 9 through the inverse quantizer 8 . The output signal of the inverse DCT circuit 9 is fed to the first input terminal of the adder 10 .

The output signal of the adder 10 is fed to a first input terminal of a motion compensation prediction circuit 11 . The first output of the motion compensation prediction circuit 11 is sent to the second input terminal of the adder 10 and the second input terminal of the subtractor 4. In addition, the second output of the motion compensation prediction circuit 11 is sent to the second input terminal of the buffer memory 12 . The output of the buffer memory 12 is input to the modulator 14 via the format encoder 13 . The output of the modulator 14 is recorded via the output terminal 2 on a recording medium such as an optical disc.

FIG. 16 is a circuit block diagram showing a reproduction system in the digital video signal recording and reproduction apparatus according to Embodiment 1. FIG. Referring to FIG. 16 , video information read from a recording medium is input to a modulator 21 from an input terminal 20 . The output of the demodulator 21 is input to the format decoder 23 via the buffer memory 22 . The first output of the format decoder 23 is input to the variable length decoder 24 and inversely quantized in the inverse quantizer 25 . The output signal is then subjected to an inverse DCT in an inverse DCT circuit 26 and fed to a first input of an adder 28 . Meanwhile, the second output of the format decoder 23 is input into the predictive data decoding circuit 27 . The output of the prediction data decoding circuit 27 is then supplied to a second input terminal of the adder 28 . The output of the adder 28 is output from an output terminal 30 via a non-formatting circuit 29 .

Then explain how the device works. A digital video signal is input from the input terminal in line units, and sent to the formatter circuit 3 . Here at the time of motion compensation prediction, as shown in FIG. 6 like a conventional example, one GOP is set to 15 frames, and predictive coding is performed, 1 frame of I picture, 4 frames of P picture (P1 to P4), and 10 frames of B pictures (B1 to B10). In this case, video data input in a continuous manner is rearranged in the formatter circuit 3 and output in the order shown in FIG. 7 in units of frames.

In addition, the data input in units of rows is rearranged in units of 8×8 pixel blocks to form a macroblock (a total of 6 blocks, such as four adjacent luminance signal Y blocks and the position corresponding to the Y block Two color-difference signal Cr and Cb blocks. The data is output in units of macroblocks. Here, macroblocks are determined by the minimum unit of motion compensation prediction, and the motion vector of motion compensation prediction is determined by macroblocks definite.

Also, for an I picture, one frame of video data is divided into three areas by the formatter circuit 3, and therefore blocks are performed in units of 8 x 8 pixels in this area to form and output macroblocks. Here, three divided areas are set as areas 1, 2, and 3 from the top of the screen as shown in FIG. 18 . In FIG. 18, the area 2 located in the central part of the screen has a size of 720 pixels x 288 lines, and the areas at both ends of the screen have a size of 720 pixels x 96 lines. Meanwhile, the P picture and the B picture are divided into blocks without being divided into regions, and are output in units of macroblocks.

The output of the formatting circuit 3 is input to the subtractor 4 and the motion compensation prediction circuit 11 . Subtractor 4, DCT circuit 5, quantizer 6, variable length coder 7, inverse digital converter 8, inverse DCT circuit 9, adder 10 and motion compensation prediction circuit 11 work conditions are the same as conventional embodiment Likewise, descriptions of their working conditions are omitted.

The video data output from the variable length encoder 7 and the motion vector output from the motion compensation prediction circuit 11 are input to the buffer memory 12 . Motion vectors and video data for one GOP portion are recorded in the buffer memory 12, and the data are then sequentially output to the format encoder 13. The output of the format encoder 13 is input to a modulator 14, and an error correction code, etc. is applied to and recorded on a recording medium such as an optical disc.

In the format encoder 13, the video data of one GOP portion is rearranged in the data arrangement shown in FIG. 19, and then output to the modulator 14. Here, the I picture is divided into three areas as shown in FIG. 18 . When the data of the I picture corresponding to these regions 1 to 3 is set to I(1), I(2) and I(3), the I picture data is constituted, so that the data is set as I(2) The order of , I(1) and I(3) is recorded in the front part of the data string of one GOP part.

In addition, the data of each picture area is stored at the address at the front end of the GOP as header information. The number of bytes occupied by the data of each image area divided into three in the format shown in Fig. 19 is also stored as header information. Therefore, based on the number of bytes occupied by each area recorded in the header information, the end position of each area can be regarded as a relative position from the head of the GOP at the time of playback. Therefore, at the time of exclusive playback, the optical head jumps to the head address of the GOP in units of a certain time, so that the data of each area can be read out based on the header information from the head of the GOP.

Using general video signal recording and reproducing equipment, I pictures are recorded in units of frames according to the recording format at the time of data recording. In contrast, in FIG. 19, priority is given to an area located at the center portion of the picture in the three-part I-picture data so that this area is located at the head of a GOP. Therefore, in the case where only a part of this area of the I picture can be decoded within a certain time at the time of high-speed playback, at least the playback picture of the central part of the picture can be output.

Next, the operation of the playback time will be described with reference to FIG. 16 . The demodulator 21 performs error correction processing so that the video signal recorded in the buffer memory 22 in the format shown in FIG. length decoder 24 . Here, the operation at the time of normal playback is the same as that of the conventional embodiment, and the explanation about the operation is omitted.

During high-speed playback, for data recorded on a recording medium such as an optical disc in a GOP unit, the optical head jumps to the front end of a GOP in units of a certain time, thereby reading out the data in units of areas according to the header information recorded in the front end. The data part of the I picture, the data is decoded in the decoder 21 and input into the buffer memory 22. Here, in the case of reading data from a recording medium such as an optical disc during high-speed playback, even if the address of the head of the GOP recorded on the optical disc is known, the time to wait for the rotation of the disc when jumping to the head of the GOP increases.

Therefore, when the speed is increased during high-speed playback, the time for reading data on the optical disc is shortened. Since the time to wait for the disc to rotate varies, it is impossible to read out all the I picture data in a stable manner. Therefore, when the speed of high-speed playback increases, after only the data in the area 2 located in the center of the picture is read, the optical head jumps to the front end of the following GOP, so that only the data in the area 2 that can be read is input into the buffer memory. twenty two. In this case, the format decoder 23 decodes only the area 2 of the I picture which can be read. On the other hand, areas 1 and 3 whose data are not read out are masked with gray data, and a high-speed playback image is output, therefore, when one GOP is set to 15 frames, a dedicated playback at 15 times speed can be obtained. Play the image.

Fig. 20 shows a reproduced image in the case of high-speed reproduction by reproducing only the area 2 of the I picture from the nth GOP to the n+3GOP of one GOP. In FIG. 20, areas 1 and 3 at both ends of the frame in the figure are masked by gray data. In addition, when the information amount of the I picture is small, the disc rotation waiting time is short, and the time to read the data in the areas 1 and 3 is available, the data in the areas 1 and 3 are not decoded. This is because at high-speed playback, all the data in one picture portion cannot be read out in a stable manner, and if the mask is output only when the data in areas 1 and 3 can be read out, these areas cannot be read in a certain range. Output within a time interval, so that the high-speed playback image appears unnatural.

As described above, since I pictures for exclusive playback are arranged to give priority to an area located at the center of a picture as shown in FIG. 19, I pictures are recorded at the head of a GOP on the recording medium. Even if the speed of high-speed playback is increased, only the data of the area 2 located at the center is read for high-speed playback, with the result that the content of the playback image is easy to see. In addition, since only the data within the region 2 is read from the recording medium, higher-speed dedicated playback can be realized compared to the case of reading the entire I picture.

In the above-mentioned Embodiment 1, the I picture is recorded vertically divided into three areas as shown in Fig. 18, but the image does not have to be divided into three areas. It can be divided into n areas (n<1) in units of slices prescribed by the international standard MPEG for recording data. Here, a slice has a one-dimensional structure of an arbitrary number (more than one) of macroblocks in length, so when the right end of the screen is reached, the left end of the next line is displayed continuously.

Embodiment 2 of the present invention will be described below with reference to the drawings. Fig. 21 is a conceptual diagram for explaining a dedicated reproduction method in the case of performing data expansion in the second embodiment. In Embodiment 1, the I picture is divided into three areas as shown in FIG. 18 so that only the data of the area 2 located at the center of the screen is read out for playback. So for regions 1 and 3, only the masked data is output. However, the data of area 2 is expanded to the size of one screen as shown in FIG. 21 .

In this case, when the video signal is converted into data in line units by the unformatting circuit 29, the value of the area 2 is interpolated, expanded to the size of one screen portion, and output. In the case of FIG. 21, the area 2 has a size of 720 pixels x 288 lines, and is formed symmetrically with 144 lines vertically from the center of the screen.

Here, during exclusive playback, if the upper half of area 2 is set as AR2a and the lower half as AR2b as shown in FIG. The reproduced images AR2a' and AR2b' are symmetrical as shown in Fig. 21B. For the method of image expansion, when the data of each line unit of AR2a is defined as AT(l) (l: line number, 1≤l≤144), and the line data of the upper half of the expanded picture is set as DT (m)(1≤m≤240), the expansion is expressed by the following formula:

DT(3n-2)＝AT(2n-1)

DT(3n-1)＝AT(2n-1)

DT(3n)=AT(2n)(n=1 to 80)

At the same time, when the data of each line unit of AR2b is defined as AB(l) (l: the number of lines, 1≤l≤144), and the line data of the lower half of the expanded screen is set as DB(m) (1≤ m≤240), the expansion can be expressed by the following formula.

DB(3n-2)=AB(2n-1)

DB(3n-1)=AB(2n-1)

DB(3n)=AB(2n) (n=1 to 80)

As described above, at the time of high-speed reproduction, only the data of the area 2 located at the center of the screen is read, expanded to the size of one screen, and output as a reproduced image. Therefore, since both ends of the image are not masked during high-speed playback, the reproduced image is no longer ugly.

In the above-described Embodiment 2, the screen is expanded in the vertical direction by simply inserting data in units of rows. Row data can be linearly interpolated with respect to the vertical direction.

Embodiment 3 of the present invention will now be described. The configurations of the recording system and the reproducing system of the digital video signal recording and reproducing apparatus in Embodiment 3 are the same as in Embodiment 1 (see FIGS. 15 and 16).

The working condition of the equipment is introduced below. A digital video signal is input from the input terminal 1 in line units, and sent to the formatter circuit 2 . Here, at the time of motion compensation prediction, one GOP is set to 15 frames as in the conventional example shown in FIG. then. GOP is predictively coded as one frame of I picture, 4 frames of P picture (P1 to P4), and 10 frames of B picture (B1 to B 10). In this case, in the formatter circuit 3, video data input in a continuous manner as in the conventional example is rearranged in units of frames in the order shown in FIG. 7, and then output. In addition, the data input in row units is rearranged in units of blocks having 8×8 pixels to constitute a macroblock as shown in FIG. 17 (adjacent four luminance signal Y blocks and two color difference signal Cr and Cb blocks, a total of six blocks), so that the data is output in units of macroblocks. Here, a large block is the smallest unit of motion compensation prediction, and the motion vector of motion compensation prediction is determined in units of macroblocks.

Furthermore, in the formatter circuit 3, the I picture is divided into five areas in the vertical direction of one frame of video data, each area being 720 pixels x 96 lines. In this area, blocks are divided into 8×8 pixels to form macroblocks and output. In this case, the divided five regions are defined as regions 1, 2, 3, 4, and 5. At the same time, the P picture and the B picture are divided into blocks, but are not divided into regions, and are output in units of macroblocks.

The output of the formatting circuit 3 is input to a subtracter 4 and a motion compensation prediction circuit 11 . Subtractor 4, DCT circuit 5, quantizer 6, variable length coder 7, inverse quantization converter 8, inverse DCT circuit 9, adder 10 and motion passively compensated prediction circuit 11 work conditions are the same as traditional implementation For the same example, the description of these devices will be omitted.

The video data output from the variable length encoder 7 and the motion vector output from the motion compensating circuit 11 are input to the buffer memory 12 . Video data and motion vectors for one GOP portion are recorded in the buffer memory 12, and the above data are sequentially output to the format encoder 13. The output of the format decoder 13 is input to a modulator 14 to add an error correction code and the like and record on a recording medium such as an optical disc.

In the format encoder 13, the data of the GOP section is output to the modulator 14 by rearranging the video signal in the data arrangement as shown in FIG. The I picture is divided into five areas as shown in FIG. 22, so that the data corresponding to the I picture of areas 1 to 5 are defined as I(1), I(2), I(3), I(4) and I (5). In Fig. 23, the data structure of the I picture is recorded in the data stream front end of a GOP in the order of I(1), I(2), I(3), (I4) and I(5), so that the priority Gives the area at the center of the frame.

Furthermore, in Fig. 23, an address for storing data of each I picture is written as header information. As header information, the number of bytes occupied in the data format by the data of each area obtained by dividing the I picture into five parts is recorded. Therefore, during playback, according to the number of bytes occupied by each area recorded in the header information, the number of bytes occupied by each area can be recorded during playback, and the end position of each area can be recorded during playback. Considered as a relative address relative to the head of the GOP. As a result, the optical head jumps to the head address of the GOP in units of a certain time, so that the data of the I picture can be read based on the header information from the head of the GOP.

With the usual general video signal recording and playback equipment, the I picture is recorded in units of frames in the data format at the time of recording. In contrast, in FIG. 23, priority is given to the area located in the central part of the picture among the five areas obtained by dividing the I picture to be arranged at the head of one GOP, with the result that even if only the part in the I picture The area can be decoded, and at least the reproduced image of the center portion of the picture can be output.

Next, the operation during playback will be described with reference to FIG. 16 . After the error correction process in the demodulator 21 and the video signal recorded in the buffer memory 22 in the format in FIG. length decoder 24 . Here, the operation during normal playback is the same as that of the conventional embodiment, and its detailed description will be omitted.

When replaying at high speed, for recording data on recording media such as optical discs, the optical head jumps to the front end of the GOP in units of a certain time, reads the data part of the I picture according to the header information, and the data is in the demodulator 21 Demodulation is performed to input into the buffer memory 22 . Yet, under the situation that the amount of information of I picture is too large to be read out in a certain period of time, half of the data has been read out until the last item of the data is read out, and the optical head jumps to the front end of the lower back GOP, only will be able to read The fetched data is input into the buffer memory 22. In this case, in the format decoder 23, only the readable I-picture area is decoded and output as a high-speed playback picture. Therefore, when the GOP is set to 15 frames, a 15-times-speed dedicated playback image can be obtained.

Fig. 24 shows reproduced pictures in the case of reproducing I pictures of one GOP. In this case, the amount of information in all areas of the I picture is too large to be read from the recording medium, and the unreadable area is output with the front area as it is, thereby synthesizing a high-speed playback image. In FIG. 24, when the n+1-th GOP area 5 and the n+3-th GOP areas 1 and 5 cannot be read, the reproduced picture preceding the reproduced picture remains as it is.

In this way, the I picture for exclusive playback as shown in FIG. 23 is set so that priority is given to the area located at the center portion of the picture at the head of one GOP recorded on the recording medium, and therefore, even if the entire In the case of the I picture, priority is also given to the center portion of the screen during playback, so that the contents of the playback image can be easily understood.

In Embodiment 3 above, when the entire I picture cannot be read, the reproduced image is interpolated in units of regions. If interpolation is not performed in units of regions, interpolation can be performed in units of error correction blocks.

In this case, according to the data arrangement shown in FIG. 23, the demodulator 21 divides the data into packets of several bytes long and adds an error correction code to each packet. Fig. 25 shows an example in the case where data input in a continuous manner in five areas is divided into several packets in units of error correction blocks. Fig. 25A shows a data string before packet division. Fig. 25B shows a data string after packet division. Five areas of I picture are divided into some data packets of certain capacity, and area I (3) is divided into the data packet from I to i, and area I (4) is divided into the data packet from i to j, so that input.

During high-speed playback, for data recorded on a recording medium such as an optical disc in units of GOP, the optical head jumps to the front end of the GOP in units of a certain time, and reads the data part of the I picture in units of areas according to the header information . The data portion is demodulated by the demodulator 21 and input to the buffer memory 22 . However, in the case where the entire I picture cannot be read within a certain time due to the large amount of information of the I picture, even when an area of data is being read, the optical head jumps to the front of the following GOP. In addition, the data that can be read is subjected to error correction processing, so that the data that can be error corrected is input to the buffer memory 22 . In this case, the format decoder 23 can recognize the address of the I picture area in the middle of decoding so that the data that can be read is decoded in units of macroblocks and output as a high-speed playback picture. In this case, for a macroblock that cannot be decoded, the previous picture is output as it is.

In Embodiment 3 above, the data in each area of the I picture is divided into packets in a continuous manner. However, data may be divided such that data of two or more areas may not be included in one packet. In this case, the data of a part of an area is enclosed in integral multiples of error correction blocks, with the result that data can be rearranged in units of areas immediately after error correction processing. When the data of each area is divided into packet units, half of the data is input to the last packet of each area, so that the remaining data needs to be placed in a data mask (for example, all data is masked as "1").

Furthermore, in the above-mentioned Embodiment 3, priority is given in the order of 3, 2, 4, 1, and 5. However, it is not limited to this order. The sequence could be 3, 4, 2, 5 and 1, for example.

Furthermore, in the above-mentioned Embodiment 3, the I picture is divided into five areas in the horizontal direction, and recorded as shown in FIG. 22 . The data is not necessarily divided into five areas, but may be divided into n areas (n>1) in units of slices defined by the international standard MPEG. Here, a slice has a one-dimensional structure of a number (one or more) of macroblocks of an arbitrary length. A slice is a band that continues to the left end of the next line when it reaches the right end of the frame.

Embodiment 4 of the present invention will be explained below with reference to the drawings. Fig. 26 is a schematic diagram showing an exclusive playback method in Embodiment 4. In Embodiment 3, exclusive playback is performed by the playback method shown in FIG. 24 . However, it is also possible to perform exclusive playback and output a playback image as shown in FIG. 26 . In this case, the format decoder 23 synthesizes one picture by reproducing each area of I pictures of five consecutive GOPs as shown in FIG. 26 . For example, in FIG. 26A, a frame portion of the reproduced image is synthesized from the I-picture of the nth to the n+4GOP, so that the I-picture of the n+4GOP is reproduced in area 1, and the I-picture of the n+3GOP is reproduced in area 1. The I picture is reproduced in the zone 2, the I picture of the n+2 GOP is reproduced in the zone 3, the I picture of the n+1 GOP is reproduced in the zone 4, and the I picture of the n+1 GOP is reproduced in the zone 5. In addition, referring to FIG. 26, when attention is paid to area 5, I pictures of nth, n+1, n+2... GOPs are reproduced as reproduction video data.

Furthermore, when the entire I picture cannot be read out within a certain time due to the large amount of information of the I picture, the data of the previous frame is output as it is to synthesize a high-speed playback image.

Fig. 27 is a playback image when the n+1th GOP area 5 and the n+3th GOP areas 1 and 5 cannot be read. In this case, since the data arrangement is recorded on the recording medium by giving priority to an area located in the center portion of the screen as shown in FIG. 23, even if the entire I picture cannot be read out in time , Also give priority to the central part of the picture during playback, as a result, the ugly scene of the playback image will not take place. Furthermore, even when data of two or more areas cannot be read, one screen is divided into five areas. Since the reproduced frames are different in each area, it is difficult to detect missing data in the reproduced image.

Next, Embodiment 5 of the present invention will be described with reference to the drawings. Fig. 28 is a diagram showing an arrangement structure of digital video signal data according to Embodiment 5. In Embodiment 3, the number arrangement is written in the order of areas 3, 2, 4, 1 and 5 for the I picture as shown in Fig. 23 . The number arrangement may also have the structure shown in FIG. 28 . In FIG. 28, when data of an I picture is recorded in the data arrangement head portion of a GOP portion, the area numbers at the head of each GOP are rolled up. In other words, as shown in FIG. 28, when I picture data is recorded on the n-th GOP in the order of I(5), I(1), I(2), I(3), and I(4), the I picture Image data is recorded in the n+1th GOP in the order of I(1), I(2), 1(3), I(4) and I(5). Also, I(2) first enters the n+2th GOP. When the GOP number becomes n+3, n+4, etc., the front-end area is rolled up sequentially, with I(3), I(4), I(5), I(1), etc... ...to be recorded.

In addition, at the head of the GOP, an address for storing data of each I picture and information for identifying the type of the head area are written as header information. As the header information, the area number recorded at the front end and the number of bytes representing the amount of data occupied in the data format of each area as shown in FIG. 27 are included. Therefore, when playing back, the data sequence of the I picture area and the data sequence of each area on the recording medium can be recorded with the front E field number recorded in the header information and the number of bytes occupied by each area on the recording medium. The end position of is regarded as the relative position with respect to the front end of the GOP. Therefore, during special playback, the optical head jumps to the front address of the GOP with a certain time as a unit, and the I image data can be read according to the header information for each area from the GOP front end.

In this case, the position where the I picture area divided into five areas is recorded is rolled up in units of GOPs, so even if only a part of the I picture area can be decoded at the time of exclusive playback, the undecodable area cannot be decoded. Focus on a fixed position in the screen.

During high-speed playback, for the data recorded on recording media such as optical discs, the optical head jumps to the front end of the GOP in units of a certain time, reads out the data part of the I picture in units of areas according to the header information, and demodulates the data. Demodulated in the device 21, and then input to the buffer memory 22. Yet when the amount of information of the I picture is too large to read the whole I picture in a certain period of time, after the last data of that area on the way to be read is read, the optical head just jumps to the front end of the next GOP, so that only The data of the area that can be read is input into the buffer memory 22 . In this case, the format decoder 23 decodes only the area of the I picture that can be read out as a high-speed playback picture. Therefore, in the case where one GOP is set to 15 frames, a 15-times-speed dedicated playback image is obtained.

FIG. 29 shows reproduced images in the case of reproducing I pictures of one GOP at high-speed reproduction. In this case, I pictures are recorded on the recording medium in the order shown in FIG. Here, when the I picture has a large amount of information and the entire I picture cannot be read out in time, the data of the previous frame is output as it is to synthesize a high-speed playback image. Fig. 29 shows the case where the n+1th GOP area 5 and the n+3th GOP areas 1 and 2 cannot be completely read. In this case, the data of the previous frame remains as it is.

As described above, the order of recording I pictures for exclusive playback as shown in FIG. 28 is scrolled in units of GOPs. Therefore, even in the case where only some areas of the I picture can be decoded at the time of exclusive playback, the areas that cannot be decoded are not concentrated at fixed positions on the screen.

Embodiment 6 of the present invention will be described below with reference to the drawings. Fig. 30 is a diagram showing a data arrangement structure of digital video data according to Embodiment 6. In this case, the I picture and the P picture are divided into five areas, each of 720 pixels x 96 lines, so that each area is divided into blocks in units of macroblocks, and as shown in FIG. 22 coding. However, the P picture is divided into five areas. Motion compensation prediction and coding are performed so that the search range of the reference pattern of motion compensation prediction is closed within the region. Here, the divided five regions are defined as 1, 2, 3, 4, and 5 from the top. Also, for B pictures, motion compensation prediction is performed without partitioning and coding is performed.

With the format encoder 13, the video signal is rearranged with the data of one GOP portion in the data arrangement shown in FIG. 30, and output to the modulator 14. Here, for the I picture and the P picture, one picture is divided into five areas as shown in FIG. 22 . The I picture and the P1, P2, P3 and P4 pictures are defined as I(1) to I(5) and Pi(1) to Pi(5) (i=1 to 4). In FIG. 30, the I picture, The data of the P1, P2, P3 and P4 pictures are structured to be recorded in the order of 3, 2, 4, 1 and 5 from the front of the data string of a GOP section so that priority is given to the area located in the central part of the picture. Furthermore, in FIG. 30, the amount of data of each area is recorded as header information at the head of the GOP, so that the address of data in each I picture and P picture area can be identified.

During high-speed playback, for the data recorded on a recording medium such as an optical disc in units of one GOP, the optical head jumps to the front end of the GOP, thereby reading out the data part of the I picture and the P area in units of areas, by the solution The modulator 21 is demodulated and input to the buffer memory 22. However, when the amount of information of the I picture and the P picture is too large to read out the entire I picture and the P picture within a certain time, an area in the middle of reading until the last data is read out. At this time, the optical head jumps to the front of the GOP in units of a certain time so that only the read data is input to the buffer memory 22 . In this case, the format decoder 23 decodes only the areas of the I picture and the P picture that can be read, and then outputs the data as a high-speed playback picture. Therefore, in the case where one GOP is set to 15 frames, a three-times-speed dedicated playback image can be obtained.

In addition, since priority is given to the area located in the center portion of the picture so as to be arranged at the front of a GOP in the I picture divided into five areas, even in the case where only a part of the I picture or the P picture can be decoded , at least the center part of the picture can also be output. In addition, pictures are recorded on the recording medium in the order of I picture, P1 picture, P2 picture, P3 picture and P4 picture. Therefore, even when all data cannot be read, it does not happen that the test data decoding circuit 27 cannot reproduce the reference data.

Fig. 31 shows reproduced pictures in the case of high-speed reproduction of pictures by playing back only I pictures and P pictures in one GOP. In this case, when the amount of information of the I picture and the P picture of the picture is too large to read the entire I picture and the whole P picture from the recording medium within a certain period of time, the data of the last frame will remain It is output as it is to synthesize a high-speed playback image for an area that cannot be read. Fig. 31 shows the case where the areas 3, 4 and 5 in P4 of the n-th GOP cannot be read. In this case, the data of the previous frame remains as it is.

As described above, as shown in FIG. 30, by collecting and arranging I pictures and P pictures used in exclusive playback in units of the area at the head of one GOP, I pictures and P pictures are played back at the time of dedicated playback. Image to output a dedicated playback image. Furthermore, in the case where the entire I picture and the entire P picture cannot be read out due to time constraints, the value of the previous frame is interpolated to allow output of the playback image.

In Embodiment 6 above, in the case where the entire I picture and P picture cannot be read, interpolation is performed on the reproduced picture in units of regions. However, interpolation may be performed not in units of regions but in units of error correction codes.

In this case, with the data arrangement shown in FIG. 30, the demodulator 21 divides the data into packets of several bytes so that an error correction code is added to each packet. Fig. 32 shows that the data of five regions input in a continuous manner is divided into packets consisting of error correction blocks. Fig. 32A shows a data string before packet division. Fig. 32B shows the data after packet division. In FIG. 32, data is divided into i-th to j-th packets in the area P1(3).

During high-speed playback, for the data recorded on a recording medium such as an optical disc in units of GOPs, the optical head jumps to the front end of the GOP in units of a certain time, thereby reading out the data of the I picture in units of areas according to the header information The part is demodulated by the demodulator 21, and then input into the buffer memory 22. However, when the amount of information in the I picture is so large that the entire I picture and the whole P picture cannot be read out within a certain period of time, the optical head will jump to the bottom even in the middle of reading the data of a part of the area. The front end of a GOP. Furthermore, the data that has been read is subjected to error correction processing, and the error-corrected data is input to the buffer memory 22 . In this case, the format decoder 23 recognizes the addresses of the I picture and the P picture in the process of decoding, so that the data that can be read is decoded in units of macroblocks and output as a high-speed playback picture. In this case, for a macroblock that cannot be decoded, the data of the previous picture is output as it is.

In the above-described Embodiment 6, the detection range is set to be closed in each area at the time of motion compensation prediction, but it does not always need to be closed.

Embodiment 7 of the present invention will be described below with reference to FIG. 33 . Fig. 33 is a diagram showing a specific playback method of the seventh embodiment. In Embodiment 6, ad hoc playback is performed by the playback method shown in FIG. 31 . However, exclusive playback can be performed so that the playback image is output as shown in FIG. 33 . In this case, the format encoder 23 synthesizes one picture by replaying regions one by one from consecutive five frames as shown in FIG. 33 . In Fig. 33A, a reproduced picture of one picture portion is synthesized from I picture, P1 to P4 pictures. In addition, in FIG. 33A, the P4 image is played back in the area 1, the P3 image is played back in the area 2, the P2 image is played back in the area 3, the P1 image is played back in the area 4, and the I image. Also, in FIG. 33, attention is paid to the elapsed time of area 5. The reproduced video data includes the I picture of the nth GOP, P1, P2, P3, and P4, and the I picture of the n+1th GOP, and the P1 picture.

In addition, when the amount of information of I pictures and P pictures is so large that all I pictures and P pictures cannot be read out within a certain period of time, the data of the previous frame is output as it is, and the high-speed playback picture is synthesized. elephant. Fig. 34 shows a reproduced image in the case where the regions 1, 4 and 5 of the n-th GOP cannot be read. In this case, as shown in FIG. 30, for the data string, priority is given to the area located at the central part to be recorded on the recording medium, so that reproduction does not occur due to priority being given to the central part of the screen. If the image is ugly. In addition, even if the data in more than two areas cannot be read, a picture is divided into five areas, and the frames reproduced in each area are different, and the missing data in the reproduced image It's hard to see.

Next, Embodiment 8 of the present invention will be described with reference to the drawings. Fig. 35 is a diagram showing the arrangement structure of digital video data in the eighth embodiment. In Embodiment 6, the data arrangement is written in the order of areas 3, 2, 4, 1 and 5 as shown in FIG. 30, but this arrangement may also have a structure as shown in FIG.

In Fig. 35, when the data of I picture and P picture is recorded at the head of the data arrangement of one GOP section, the area number at the head is rolled up for each frame. In other words, as shown in FIG. 28, I picture data is recorded in the order of P1(2), P1(3), P1(4), P1(5) and P1(1) in the nth GOP. Also in the P2 image, P2(3) becomes the front end. In the P3 picture and the P4 picture, the leading area is rolled up and recorded sequentially like P3(4) and P4(5).

In addition, at the head of the GOP, addresses where data of I pictures and P pictures are stored, and type information for identifying an area at the head of each frame are recorded as header information. Here, the area number recorded at the head of each area and the number of bytes indicating the amount of data in each area divided into five parts are recorded as header information. Therefore, at the time of dedicated playback, the optical device jumps to the head of the GOP in units of a certain time, so that data can be read out in units of areas based on header information.

In this case, since the positions where the I picture and the P picture are divided into five parts are rolled up in units of frames, even in the case where only a part of the I picture and the P picture can be decoded, no It happens that undecoded areas are concentrated at fixed positions on the screen.

In high-speed dedicated playback, data recorded on a recording medium such as an optical disk in units of one GOP is read out in units of areas based on header information. At this time, the data is demodulated by the demodulator 21 and input to the buffer memory 22 . However, when the information amount of the I picture and the P picture is so large that the entire I picture and the P picture cannot be read within a certain time, for the area in the middle of reading, the last data is read. Then, the optical head jumps to the head of the GOP, and only data that can be input into the area of the buffer memory 22 is input. In this case, the format decoder 23 decodes only the areas of the I picture and the P picture, and outputs it as a high-speed playback picture.

Fig. 36 shows a reproduced picture of high-speed reproduction in the case where only I pictures and P pictures in one GOP are reproduced. In this case, when the amount of data of the I picture and the P picture is so large that the entire I picture and the P picture cannot be read out within a certain period of time, it is possible to output the data of the last frame as it is. Composite high-speed playback patterns. Fig. 36 shows the case where the areas 3, 4 and 5 in P4 of the n-th GOP cannot be read. In this case, the data of the previous frame remains as it is.

As described above, since the order of recording I pictures is rolled up in GOP units as shown in FIG. It will not focus on a fixed position on the screen.

Next, Embodiment 9 of the present invention will be described with reference to FIG. 37. FIG. Fig. 37 is a block diagram showing a recording side of a digital video signal encoding processing unit in a digital video recording and playback apparatus, in which DCT blocks are divided into low-frequency area and high-frequency area levels so that only the low-frequency area is located at the front of a GOP. In FIG. 37, numeral 51 denotes a buffer memory, 52 denotes a subtractor, 53 denotes a DCT circuit, 54 denotes a digital conversion circuit, 55 denotes a variable length encoder, 56 denotes an inverse digital conversion circuit, 57 denotes an inverse DCT circuit, 58 denotes an adder, 59 denotes a motion compensation prediction circuit, 60 denotes a counter for calculating the number of events and the code amount, 61 denotes a format encoder, and 65 denotes an input terminal.

The following describes the working conditions of the equipment. The video data to be input is an interlaced image with an effective frame size of 704 pixels horizontally and 480 pixels vertically. Here, the operations of the subtractor 52, the DCT circuit 53, the digital conversion circuit 54, the variable length encoder 55, the inverse quantizer 56, the inverse DCT circuit 57, the adder 58 and the motion compensation prediction circuit 59 are the same as those of the conventional implementation. If the corresponding devices in the examples are the same, their descriptions will be omitted.

Referring now to Fig. 38, the operation of the variable length encoder 55 will be described. Fig. 38 shows the data arrangement of DCT coefficients in a DCT block. In FIG. 38, the low-frequency components are located in the upper left portion, and the data of the DCT coefficients of the high-frequency components are located in the lower right portion. Of the data of the DCT coefficients arranged in the DCT block, the data of the coefficients up to a specific position (the end of the event) (for example, the shaded portion in FIG. 38 ) is coded into a variable-length code as a variable-length region, and output to format encoder 61 . Then, variable-length coding is performed on the DCT coefficient data subsequent to the DCT coefficient data at the above position. In other words, data in the spatial frequency region is coded by segmentation.

The boundary between the low frequency region and the high frequency region is called the inflection point. The inflection point is set as a predetermined code amount of the low-frequency region that the head can reach at a given playback time. The variable length encoder 55 divides the DCT coefficients into a low-frequency region and a high-frequency region according to the inflection point to output to the format encoder 61 .

Identify coding regions at the boundaries of the event. It goes without saying that determination can be made by other methods. For example, coding regions can be identified at the boundaries of a certain number of events. The data is divided using the quantized data coarsely quantized by the quantizer 54 and the difference value between the fine quantization and the coarse quantization. In addition, it is possible to encode a picture whose resolution has been thinned out to about half using the buffer memory and to encode a difference picture between an image whose resolution has been restored from half the resolution and an image of the original resolution to encode The data is divided. In other words, data division is not limited to division of frequency regions. Needless to say, the high-efficiency coded data of an image can be divided by quantization and division of spatial resolution.

At this time, more important data as an image when divided by frequency is low frequency data. When division is by quantizer, data is subjected to coarse quantization for encoding. When the data is divided by spatial resolution, the thinned image is coded. By decoding only those important data, it is possible to obtain a decoded image which is easy to perceive by human beings. In this way, an efficiently encoded data can be separated into more fundamental important data and other data (a process called layering), plus error correction, and modulated; for recording on an optical disc.

In this way, since only the low-frequency components of I pictures and P pictures are divided, reading and playing back only these low-frequency components at the time of special playback greatly reduces the amount of data read out at the time of special playback. As a result, the time required to read data from the medium is short, and high-speed playback with smooth movement can be realized during skip seek. Furthermore, when only the I picture and the P picture are arranged in a continuous manner, the data of the low frequency components of the I picture and the P picture can be easily read from the optical disc for decoding. In this case, a more efficient data structure can be formed only by extracting and arranging low-frequency components than arranging the entire areas of I pictures and P pictures at the front of the GOP.

The operation of the format encoder 61 will be introduced below. Fig. 39 is a block diagram showing the operation of the format encoder. At the beginning, when the encoding job starts, it is judged whether the encoding mode is in the hierarchical mode. When the schema is not hierarchical, information is inserted into the system stream that indicates that the schema is non-hierarchical in order to obey the traditional stream structure. In case of layered mode, make sure that the sequence header is set. Specifically, data that confirms the measurable extension of the sequence. When the data is correctly written, the front end of the picture is recognized so that the I picture and the P picture are divided into low frequency component data and high frequency component data to detect the respective data lengths.

At the same time, the length of the data of the B picture is detected for each picture. In addition, when the low-frequency component data of the I picture and the P picture are arranged immediately after the top of the GOP, a packet in which only address information is recorded is created. In this data packet, the low frequency components of the I picture and the P picture, the high frequency components of the I picture and four P pictures, and the address information of ten B pictures are included, so as to record the data length of each data.

Therefore, based on this data length, the leading position of each data stream can be obtained as a relative address with respect to the leading end of the GOP header. Packets containing this address information, low frequency components of the I picture and four P pictures, and the remaining data are sequentially arranged for formatting.

Wherein, the confirmation of the measurable mode on the measurable extension part of the sequence header refers to the confirmation of the measurable mode setting determined in the syntax of MPEG2 in FIG. 40 and the confirmation of the description of the priority inflection point on the slice header. The priority inflection points are located at a predetermined number of events in FIG. 40 (corresponding to the above inflection points), and refer to data representing boundaries between the divided low frequency components and high frequency components.

When the scrambling mode is assumed to be "00", it means that the subsequent bit stream is a data-divided bit stream, and also means that the bit stream divided into low-frequency components and high-frequency components is continuous. When the B picture is composed of low frequency components so that high frequency components are not generated, the B picture is not divided.

An example of a bitstream generated in this way is shown in Figure 41. Fig. 41A shows an unlayered bitstream. When the bit stream is layered with the circuit shown in Fig. 37, the bit stream is divided and layered as shown in Fig. 41B. When data is arrayed in consideration of such exclusive reproduction, low frequency components of I pictures and P pictures are arrayed at the front of the GOP as shown in FIG. 41C.

FIG. 41D shows a data arrangement in the case where address information is included in a dedicated packet as shown in the flowchart in FIG. 39 . In this case, as described above, the address information can be expressed by a relative address with respect to the front end of the GOP header. However, the address information can be represented by which byte of which packet is the head of each picture. Needless to say, address information can also be represented by sector addresses on the optical disc.

Fig. 42 shows an example in which address information is included in a dedicated packet. When a packetized elementary bitstream (referred to as PES) packet is used as a private packet, the bitstream ID is expressed as BF (hexadecimal notation). After describing the packet length, the MSB of the byte is set to 1, and the following bits are set to 0, so that the code is indistinguishable from all start codes (start code of data packet and start code of bit stream) same. Then, the remaining six bits are used to describe the layering mode, the layering type, the image type to be used in the exclusive playback, and the number of start addresses.

Thereafter, 21-bit long address information is described so that the maximum length of GOP data amount up to 2M bytes can be expressed. However, 100 (in binary notation) is inserted into the first three bits of the 21-bit data, making the data appear different from the first 24 bits of the above start code, 000001 (in hexadecimal notation). Here, the starting address includes the starting address of the low frequency component of the I picture, the starting address of the low frequency component of the four P pictures, the high frequency component of the I picture, the starting address of the high frequency component of the four P pictures , and the starting addresses of ten B pictures. In addition, the sector addresses on the optical disk on which the data of the previous and subsequent GOPs are recorded are added to make the optical head jump at the time of specific playback.

When a 1-bit parity bit is added to a 21-bit address, the reliability of the data is improved. In this case, 10 (binary notation) can be added to the front relative to 21 bits + 1 bit. In addition, considering the high-speed multiple of dedicated playback, when adding the sector areas of several GOPs before and after and the addresses of the previous and next GOP, the high-speed multiple of dedicated playback can be expanded. Furthermore, this indicates that address information is described in the dedicated 2 packet of the packet of PES. Needless to say, address information may be written on other user areas or the like of dedicated descriptors such as program flow graphs or the like.

Now, the playback side of Embodiment 9 will be described based on FIGS. 43 and 44. FIG. Fig. 43 is a block diagram of a digital video signal decoding processing section. In Fig. 43, numeral 71 denotes a program stream header detector, 72 denotes a PES packet header detector, 73 denotes a video bit stream generator, 74 denotes a data rearranger, 75 denotes an address memory, 76 denotes a mode switching switch, 77 78 denotes a switch, 79 denotes an inverse quantizer, 80 denotes an inverse DCT circuit, 81 denotes an adder, 82 denotes a predictive data decoding circuit, 83 denotes a frame memory, and 84 denotes a decodable determiner. FIG. 44 is a schematic diagram showing the operating principle of FIG. 43 .

Next, the working condition of Fig. 43 will be described according to Fig. 45 . Fig. 45 is a flowchart showing the operation of the format decoder during playback. The bit stream output from the ECC is subjected to inspection of the program stream header to be divided into individual PES packets. In addition, the header of the PES data packet is detected to distinguish the private data packet containing address information from the video data packet.

In the case of a dedicated packet, the address information contained in the packet is extracted and stored. Meanwhile, in the case of video packets, the bit stream of video data is extracted. Here, in the case of normal playback, for I picture and P picture, data of low frequency component and high frequency component are extracted from the bit stream of video data, thereby rearranging data and outputting a playback picture. At the same time, only the low-frequency components of the video data are extracted for playback during exclusive playback. Here, after reproducing the low frequency components, the optical head is allowed to jump to the front of the following GOP.

In this case, when describing these addresses in the video stream, the address information is extracted and stored after being converted into a bit stream. Therefore, in the case where the address information is described by the dedicated descriptor of the program flow graph, the address information is extracted and stored at the detection level of the header of the program flow graph. Needless to say, address information may be a relative address or an absolute address.

Actually, mode signals such as skip search and normal continuous playback are input to the mode changeover switch 76 . Simultaneously, a reproduction signal from an optical disk or the like is amplified by an amplifier, thereby reproducing the signal with a phase-synchronized clock output from a PLL or the like. Discrimination operations are then performed for digital demodulation. Therefore, after the error correction process is performed, the program stream header detector 71 obtains the data information after the header.

Also, the PES packet header detector 72 detects address information of each image described in a private 2 packet (private2 packet) of the PES packet and address information of data for exclusive playback, and stores the information in the address memory 75 . Here, PES packets for audio, PES packets for characters, etc., and PES packets for video are classified so that only video packets are output to the video bitstream generator 73 .

Here, the video bitstream generator 73 erases the additional information from the PES packets to form a bitstream. Specifically, data such as various control codes and time stamps are erased. Thereafter, based on the address information obtained from the address memory 75, the bit stream is rearranged by the data rearranger 74 at the time of normal reproduction using the output of the mode changeover switch 76.

The output (control signal) of the mode changeover switch 76 is sent to the data rearranger 74 and the decodability determiner 84 . By obtaining the control signal, the data rearranger 74 reconstructs data before division from the divided and layered low frequency components and high frequency components. Otherwise, only the low frequency components are output to the variable length decoder 77 . In other words, during normal playback, each low frequency component is synthesized with a high frequency component so that the device operates in a manner of rearranging data in the order of ordinary pictures. At the time of exclusive playback, only the low-frequency components of the I picture or the low-frequency components of the I picture and the P picture are output depending on the high-speed multiplier.

Time stamps are not used when only dedicated playback by low frequency components is allowed. In contrast, the variable length decoder 77 together with the decodable determiner 84 extracts the boundary of an event in the low frequency component area represented by the priority inflection point of the tile header, so that data up to the boundary is decoded and output to the switch 78 . This switch 78 is connected so that no zeros are inserted during normal playback. At the same time, at the time of exclusive playback, the decodability determiner 84 controls the switch 78 so that zeros are inserted into the high-frequency component area after the priority inflection point at the time of specific playback.

The above operation will now be described with reference to FIG. 44. FIG. Referring to FIG. 44, when the division inflection points are E1 to E3, E1 to E3 are stored in the low-frequency component stream. E4 to EOB are stored in the high-frequency component stream. In the low-frequency component stream, low-frequency component data in subsequent DCT blocks after E3 are stored.

Here, at the time of normal playback, the data rearranger 74 extracts data E1 to E3 from the low-frequency component stream, and data E4 to EOB from the high-frequency component stream. Also, the data rearranger 74 extracts data separately, and sequentially reconstructs DCT data. In contrast, at the time of dedicated playback, the data rearranger 74 extracts the data E1 to E3, followed by variable-length decoding by the variable-length decoder 77, and the decodable determiner 84 detects a priority inflection point so that zeros are inserted into the graph. 44, so that only low frequency components are used to form DCT blocks.

Data converted into DCT blocks is decoded based on motion vectors. Here, since the decoding is the same as the conventional example, the explanation about the motion vector decoding is omitted. However, a reference for decoding a P picture at the time of specific playback is decoded by using an I picture or a P picture decoded with only low-frequency components.

Data decoded in units of data blocks is input to the frame memory 83 . Here, the frame memory 83 restores images in the original order of the GOP structure, and outputs them by switching from block scanning to raster scanning. Incidentally, the frame memory 83 can generally be used in conjunction with the memory incorporated in the predictive data decoding circuit 82.

The encoding area is defined at the boundary of the event, but it goes without saying that the boundary can also be defined by other methods. In other words, in addition to the frequency area division, the high-efficiency encoding data of the image can also be divided by quantization or spatial resolution.

At this time, the more important data as an image in the case of frequency division is the data of the low frequency area. In the case of quantized division, the data refers to data encoded by coarse quantization. In the case of dividing data by spatial resolution, the data refers to data obtained by encoding a thinned-out image. In this case, in the reproduced image decoded using only these data items, areas easily perceived by people are defined as important data. In other words, a high-efficiency coded data is divided into basic important data and less important data (this process is called layering), so that when the data is played back from the disc, only the basic important data can be played back during dedicated playback.

Embodiment 9 describes the case of the recording side corresponding to the playback side. This also considers the case where recording and reentry are combined into one set, such as a hard disk or the like, and only the playback side is considered, presupposing that data is recorded according to the principle of a generally available laser disc or the like.

Embodiment 10 of the present invention will be described below. Fig. 46 is a block diagram showing a recording system of a digital video signal recording and reproducing apparatus according to Embodiment 10 of the present invention. The same numerals in FIG. 46 denote the same or corresponding parts in FIG. 37 . Number 65 denotes an input terminal, 51 denotes a buffer memory, 52 denotes a subtractor, 53 denotes a DCT circuit, 54 denotes a quantizer, 56 denotes an inverse quantizer, 57 denotes an inverse DCT circuit, 58 denotes an adder, and 59 denotes a motion compensation prediction A circuit, 55 denotes a variable length encoder, 62 denotes a region rearranger, and 61 denotes a format encoder.

Fig. 47 is a circuit block diagram showing a reproducing system of a digital video signal recording and reproducing apparatus according to Embodiment 10 of the present invention. Like numbers in FIG. 47 denote like or corresponding parts in FIG. 43 . Number 71 denotes a program stream header detector, 72 denotes a PES data inclusion detector, 73 denotes a video bit stream generator, 85 denotes an area rearrangement device, 75 denotes an address memory, 76 denotes a mode switching switch, and 77 denotes a variable length decoder , 79 denotes an inverse quantizer, 80 denotes an inverse DCT circuit, 81 denotes an adder, 82 denotes a prediction data decoding circuit, and 83 denotes a frame memory.

The working situation of embodiment 10 is introduced below. The digital video signal from the input terminal 65 is input in line units and sent to the buffer memory 51 . The operation from the buffer memory 51 to the variable length coder 55 is the same as the above example, and the description thereof will be omitted.

For the I picture in the bit stream of the video signal output from the variable length encoder 55 in units of GOPs, the area rearranger 62 rearranges the data so that the area located at the center portion of the picture is arranged at the head of the bit stream. Here, the I picture is divided into three areas as shown in Fig.48. Data corresponding to regions 1-3 in the I picture are defined as I(1), I(2) and I(3). However, each area shown in Fig. 48 is a collection of a plurality of MPEG slice layers. In FIG. 48, areas 1 and 3 are composed of 6 tiles, and area 2 is composed of 18 tiles.

Actually, the area rearranger 62 detects the tile header of the I picture of the bitstream, and divides each tile into three areas, as shown in FIG. 48, thereby forming a bitstream for each area, rearranging bitstream. In other words, as shown in FIG. 49, the bit stream is rearranged in this sector unit so that the bit stream is arranged at the head of the GOB in the order of I(2), I(3) and I(1). The rearranged bit stream is output to the format encoder 61 in GOP units.

Next, the operation of the format encoder 61 is explained based on FIG.50. Fig. 50 is a flowchart illustrating an algorithm for formatting video data into PED packets in GOP units. In the case of the screen center portion priority mode, the bit stream image header and image information to be input are detected. In the case of an I picture, the center portions of the planes I(2), I(3) and I(1) shown in FIG. Address information is made into a 24-bit wide binary number. On the other hand, the data length is detected in picture units with respect to P and B pictures, so that the data length is converted into a 24-bit wide (3-byte) binary number to make address information.

The formatting unit collects input address information and video data bit streams, which are divided into two types of PES packets. In other words, a PES packet having only address information and a PES packet having only audio are constituted.

Therefore, in the case where one GOP is constituted by 15 frames, as shown in FIG. 6, there are 17 kinds of pictures as address information, for example, 3 kinds of I pictures, 4 kinds of P pictures, and 10 kinds of B pictures. Also, there are two types of address information (absolute addresses on the disc) of the leading GOP and the following GOP on the disc as address information at a particular playback time. The address information for these entries is collected in a packet and formatted as a PES packet. Actually, these address information items are collected in one pack and formatted as a dedicated I pack of the PES pack shown in FIG. 51 . In Fig. 51, the absolute addresses of the leading GOP and the following GOP are arranged in front of the pack data on the disc. Then, the address information of each image is sequentially arranged. However, address information with a length of 3 bits (24 bits) is specified for each address information, so the length of the packet is 57 bits.

During this period, as far as a GOP part of the bitstream other than address data is concerned, the bitstream is formatted into PES packets (video packets) by dividing the bitstream into packets and adding headers such as synchronization signals information.

Furthermore, the format encoder 61 divides the input bit stream of audio data into a plurality of PES packets, constituting an MPEG2-PS system stream having PES packets of video data. Actually, as shown in FIG. 52, the bit stream of one GOP portion of video data and the bit stream of audio data are divided and arranged in a plurality of packets of the same group. In this case, a packet representing the aforementioned address information is arranged in the preceding packet of the system flow of FIG. 52 . Then, the apparatus is constructed in such a manner that a packet containing a bit stream at the center portion of the screen of the I picture is constructed.

Next, the operation at the time of playback is explained with reference to FIG.47. In FIG. 47, since the operations of the program stream header detector 71, PES packet header detector 72, video bitstream generator 73 and mode converter 76 are the same as those of the conventional method, explanations are omitted.

In the decoded video bit stream, the data at the center portion of the screen of the I picture is placed at the front of the bit stream. The area rearranger 85 is based on the data lengths of the bit streams I(2), I(3) and I(1) output from the address memory 75, and in the order of I(1), I(2) and I(3) Reassembles I image data. The recombined bit stream is input to a variable length decoder 77, and decoded into block data, motion vectors, and the like. Since the operation after variable length decoding is the same as the conventional method at the time of normal playback, its explanation is omitted.

In high-speed playback, as described above, a GOP data portion is allocated to a group of system streams, so it is considered that there is a way that when the optical head reads data from the disc, it jumps to the front address of each GOP to Only the I-picture data placed in front of the system stream is read, so that the optical head jumps to the front of the succeeding GOP. In this case, a PES packet having an address information record arranged in front of the system stream is detected, and the disk drive is controlled by decoding the address on the disk of the succeeding GOP and the address information of the I picture.

In the case of Fig. 6, when all I pictures in each GOP can be read in one frame, 15 times high-speed playback can be realized. When the I picture in each GOP is read in two frames, 7.5 times high-speed playback can be realized. In this way, it takes less time to read data on the disc during high-speed playback.

In the case of reading data from a recording medium such as an optical disc, even if the previous address is known, there is a waiting time for the rotation of the disc. At this time, the optical head will jump to the position where the data is actually recorded. When the video signal is coded at a variable rate, the amount of information of the I picture is variable, and the time to read the I picture is also variable. Therefore, the higher the speed of high-speed playback, the shorter the time to read data on the disc. Since the waiting time of the disc rotation is indeterminate, it is impossible to stably read the entire data of the I picture.

In Embodiment 10, at the time of high-speed playback, for data recorded in GOP units on a recording medium (optical disc, etc.), the optical head jumps to the data recorded in the GOP in a limited time unit. In this way, the data portion of the I picture is read from the disc. In this case, it seems that the entire data of the I picture cannot be read, and the optical head jumps to the front of the subsequent GOP. In other words, the optical head jumps to the front address of each GOP in a certain time unit to read as much data as possible from the front of the system stream, and then jumps to the front of the subsequent GOP.

In this case, PES packets containing addresses of disc data and addresses of subsequent GOPs and PES packets containing data of the central portion of I pictures are arranged at the front of the system stream. Therefore, even when the entire data of the I picture cannot be read at the time of exclusive playback, at least the address on the disc of the subsequent GOP and the data of the central part of the I picture can be decoded as an address and data for controlling the disc drive.

At the time of exclusive playback, only the central portion of the screen can be decoded, only the data that can be decoded by the area rearranger 85 is output to the variable length decoder 77, and like this, the video data of the variable length decoding is quantized by reverse quantization. tor and inverse DCT are input to frame memory 83. Meanwhile, the area rearranger 85 inputs area information which cannot be decoded to the frame memory 83 . Regarding the area that cannot be decoded, the output data of the previous frame is left as it is and output.

Fig. 53 shows an example of reproduced pictures, in which high-speed reproduction is performed by reproducing only I pictures in nth GOP to n+4th GOP. Fig. 53A shows a case where the entire I picture can be decoded. In the case of Fig. 53B, areas 2 and 3 can be decoded. In area 1 which cannot be decoded, the value of the previous frame is saved and output as it is. In addition, in the case of Fig. 53C, only precinct 2 can be decoded. In areas 1 and 3, the value of the previous frame is preserved as it is.

In a general video signal recording and reproducing apparatus, a format is adopted in which I pictures are recorded in frame units at the time of recording. On the contrary, in Fig. 52, the area located outside the center portion of the screen, I picture data (divided into three parts) is arranged at the front of the GOP because it has priority. Therefore, even in the case where only the area of the I picture portion can be read from the disc at a certain time at the dedicated playback timing, at least the playback screen at the center portion of the screen is output.

As described above, in Embodiment 10 (FIG. 52), regarding the I picture for exclusive playback, the data located in the center area of the screen is arranged in front of a GOP, and this area is recorded on the recording medium with priority, By making the area 2 located at the central portion of the screen have playback priority even when the high-speed playback speed is high, it is easy to see the content of the high-speed playback image. When the dedicated playback is completed, the optical head jumps to the front of the GOP in a certain time unit, and as a result, the output picture is updated at a predetermined high speed.

Incidentally, the aforementioned embodiment may be constructed such that data of areas that can be decoded at the time of exclusive playback are all output, and for areas where data cannot be decoded, data of the previous frame is left intact. However, only the center portion of the screen can be played back at a dedicated playback moment.

In this case, the area rearranger 85 decodes only the data of the I picture area 2 read from the disc. As for areas 1 and 3 whose data cannot be decoded, for example, gray data is masked to output a high-speed playback image of the frame memory 83 .

Fig. 54 also shows a playback picture in which only the area 2 of the I picture of the nth GOP to n+4th GOP is used for high-speed playback. In Fig. 54, areas 1 and 3 at both ends of the screen are masked with gray data. Even when the information amount of the I picture is small and the waiting time of the disk rotation is short, there is enough time to read the data of the areas 1 and 3. Data in areas 1 and 3 are not decoded either.

This is because, if the data of areas 1 and 3 can be read and output on the screen, if areas 1 and 3 are not updated for a certain period of time, the high-speed playback image becomes unnatural. Therefore, only When the center portion of the screen of the I picture is played back at the dedicated playback time, the area to be updated does not change, and the played back image becomes natural.

In the above-described embodiment, only the areas of the central portion of the I picture (these areas can be decoded at the time of exclusive playback) are displayed to cover both ends of the screen. However, the center portion of the screen can be expanded to a screen size and output.

In this case, in the frame memory 83, the decoded data of area 2 is expanded to the size of one screen, as shown in FIG. 55 . However, in the case of FIG. 55, the central portion of the area 2 surrounded by the dotted line is expanded to double the size by means of linear interpolation in the horizontal and vertical directions. In other words, in the case of Fig. 55, the size of the portion surrounded by the dotted line is 360 pixels horizontally x 240 lines vertically. This dotted portion is linearly interpolated to expand to the size of a screen, 720 pixels horizontally by 480 lines vertically.

Therefore, only when the data at the central portion of the screen is decoded to expand it to the size of one screen at the time of special repetition, the area of the output data becomes smaller. In this way, it is possible to eliminate the portion masked at both ends of the screen, which is only noticeable when the central portion of the screen is output.

In the above embodiment, only the center portion of the screen of the I picture is prioritized in the bit stream. However, other configurations are also possible, for example, both the center of the screen of the P picture and the center of the screen of the I picture have priority. In this case, the data of the center portion of the screen plane of the P picture is arranged after the bit stream of the I picture.

In the above-described embodiments, image data is converted into a bit stream and then rearranged in units of areas. However, image data may not necessarily be rearranged after conversion into a bit stream but before conversion into a bit stream.

Fig.56 shows a flow chart of the playback section of Embodiment 10. The process of the flow chart has been described before, and will be omitted here.

The recording section of Embodiment 10 is described similarly to the playback section. There is also a case where recording and playback constitute a pair of hard disk-like things. It is also possible to consider a case where recorded data is assumed to be recorded when a part is played back like the case of a compact disc (CD) now. Apparently, screen data rearrangement of precinct units can be realized in predictive data decoding circuit 82 and frame memory 83 by using data of the lower 8-bit long sector vertical position of the sector start code of the sector header.

Example 11

Next, Embodiment 11 of the present invention will be described. Fig. 57 shows a digital video signal encoding processing unit in a digital video signal recording and reproducing apparatus in which frame DCT data is layered into a low frequency area and a high frequency area. Fig. 57 is a block diagram at the time of recording. At this time, the screen is divided into a plurality of areas, and the center portion of the screen in the low frequency area has priority to be arranged in front of the GOP. In Fig. 57, reference numeral 62 denotes an area rearranger. Similar parts and corresponding parts in Fig. 57 are given the same reference numerals and will not be described here.

Next, the operation of the device is explained. The video data to be input includes an effective screen size of 740 pixels horizontally by 480 pixels vertically. Motion compensation and DCT are used to add high-efficiency coding to image data. Here, the operations before data division and layering are the same as those in Embodiment 9, and will not be repeated here.

Embodiment 11 is the same as Embodiment 9 in that the data can be divided by restrictive conditions, spatial resolution, and frequency domains related to division layers. In Embodiment 11, the important data that is further divided and layered by the data rearranger 62 is divided into each area of the screen as in Embodiment 10, so that the central part of the screen has priority to be arranged in front of the GOP. In other words, data is divided into important data and unimportant data so that it can be recorded on the disc in a predetermined order of priority in the zone.

In this method, the low frequency components of the I picture and the p picture are divided so that the central part of the screen has priority in arrangement. Only when the center portion of the screen of these low-frequency components is read and played back at the special playback time, the amount of data read at the special playback time can be greatly reduced. Therefore, the speed of reading the recording medium can be more than ten times or even tens of times to realize extremely fast skipping retrieval.

Here, the screen central portion of the low frequency of the screen is arranged in front of the GOP, and the data of the P picture follows the data of the peripheral portion of the screen in the low frequency area of the I picture. As a result, since only the low frequency of the I picture is reproduced The central part of the screen, so it can achieve ten times or even dozens of times the high-speed playback. Also, the screen central portion of the low frequency components of the I picture and P picture has a small amount of data for special playback, so the data at the screen central portion can be easily read from the disc and decoded. Thus, high-speed playback can be realized at several times the speed. In other words, since the data amount of the center portion of the screen is smaller than that of the entire low frequency component for the I picture and the low frequency component of the P picture, playback can be performed at a faster speed than that of Embodiment 9.

Next, the operations of the precinct rearranger 62 and the format encoder 61 are described, and Fig. 58 is a flowchart thereof. At the beginning, that is, the encoding starts, the slice header of the I picture of the low-frequency component portion is detected, and each slice is divided into three areas, as shown in FIG.48. Then, a bitstream for each region is prepared to recombine the bitstream for each region. In other words, the data is reorganized for each area, and the bit stream is arranged in front of the GOP related to the low frequency area I picture in the order of low frequency area I(2), low frequency area I(3) and low frequency area I(1), as Figure 49.

Then, in the case of the screen center portion priority mode, the image header of the bit stream to be input is detected to detect image information. Here, in the case of the low-frequency region I image, the low-frequency region screen center portion I(2), the low-frequency region I(3), and the low-frequency region I(1) are extracted to detect the data length, thereby preparing from the data length of each region Address information. Meanwhile, in the case of P pictures and B pictures, the data length is detected in picture units, thereby preparing address information. In the non-picture central part priority mode, the operation is as in Embodiment 9.

Next, judge the classification mode. In non-hierarchical mode, information indicating that the mode is non-hierarchical is inserted into the system stream, following the traditional stream structure. In the case of layered mode, confirm the setting of the sequence header. That is, the acknowledgment sequence may measure extended data. In the case that the data is correctly defined, the front of the picture is recognized by the picture header, thus, the low-frequency area data of the I picture and the P picture reconstructed in the screen area are extracted, and the length of the data is checked. At the same time, the data length of the B picture is checked for each picture frame.

Next, prepare a pack in which only address information is recorded in the case where the screen central part of the low frequency area of the I picture and the P picture is arranged in front of the GOP, this pack includes the low frequency of the I picture and the P picture The address information of the central part of the screen, the peripheral part of the screen, the high frequency area part of the I picture and the P picture, and the B picture of the area part is used to record the data length of each data. Therefore, the front position of each data stream is obtained as an address relative to the front of the GOP header.

In Fig. 59, a bitstream prepared in this way is shown. As shown in Fig. 59c, the I-picture and P-picture low-frequency areas rearranged in sector units are rearranged at the front of the GOP. The case shown in FIG. 59D is that the recombined low-frequency region data is arranged in packets, so address information is arranged in special 2 packets, as shown in the flow chart of FIG. 58 . In this case, the address information can be replaced by the aforementioned relative address associated with the front of the GOP header. In addition, address information can also be expressed in such a way that which byte in which packet falls in front of each picture.

Needless to say, address information can also be represented by sector addresses added to the disc.

Fig. 60 shows a case where address information is included in a dedicated 2-packet. In this case, a PES packet is set as a dedicated 2-packet stream ID, so the layered mode, layered category, used at the time of dedicated playback are described. The image category and the number of the start address, wherein the start address refers to: the start address of the screen center part of the low frequency area of the I picture, the start address of the screen peripheral part of the low frequency area of the I picture and the reserved B The starting address of the image.

The sector addresses of the leading and subsequent GOPs are added to the disc to cause the optical head to skip at the specific playback time. In this case, considering the high-speed times of the special-purpose playback time, in addition to the leading and subsequent GOP addresses, the section addresses of several previous and subsequent GOPs are also added, and the change of the high-speed times of the special-purpose playback will be added big. It is also explained that the address information is specified in the dedicated 2 pack in the PES pack. Apparently, the address information can be specified in special defined parts such as program flow diagram and other user areas.

The playback section of the apparatus in Embodiment 11 will now be described based on FIG. 61. FIG. Fig. 61 is a block diagram of a digital video signal decoding unit. Similar parts and corresponding parts in the drawings are marked with the same reference numerals, and will not be repeated here.

Next, FIG. 61 is explained based on FIG. 62 . Fig. 62 is a flow chart showing a flow chart of the format decoder at the time of playback. The program stream header in the bitstream output by ECC is detected, and the bitstream is separated for each PES packet. The headers of the PES packets of the bitstream are inspected to distinguish specialized packets containing address information from video packets.

In the case of a special package, the address information contained in the package is extracted and stored. Meanwhile, in the case of a video packet, the bit stream of the video packet is extracted. In the case of special packs and normal playback, or in the case of video packs, low-frequency components and data of low-frequency components are extracted from the bit stream of video data of I pictures and P pictures, thereby recombining the data for output playback image.

Meanwhile, in the case of exclusive pack and normal playback, it is judged at the beginning whether the time is suitable for playback of the entire low-frequency I picture. In the case that the time is suitable for playback, it is further judged whether the time is suitable for playback of the P picture. Complete two or one of the above judgments. Therefore, in the case where the time is suitable for reproducing the low-frequency I picture and P picture, I picture and I picture are played back. In the case where the time is suitable for playback of the entire low-frequency I picture but not for playback of the low-frequency P picture, only the low-frequency P picture is played back. Furthermore, in the case where the time is not suitable for reproducing the entire I picture, the center portion of the screen of the low frequency I picture is reproduced. In the above three cases, let the bald head jump to the head of the next GOP.

In the case where these addresses are specified in the bit stream, address information is extracted and stored after the bit stream is formed. In the case where these addresses are specified in a specially defined part of the program flow graph, the address information is extracted and stored in the stage where the program flow header is inspected. Obviously, the address information can be either a relative address of the program or an absolute address of the program.

Actually, as shown in FIG. 61, mode signals of skip scan, normal continuous playback, etc. are input to the mode converter 76 from the microcomputer. Simultaneously, the reproduction signal from the disc is amplified by the amplifier, and the reproduction signal is reproduced with the clock output from the PLL, so that the phases are synchronized. Then, the signal is digitally modulated, error corrected, and the program flow resumed. Then, the program stream header detector 71 which detects each program stream header obtains information of the data following the header.

The address information of each picture specified in the special 2 pack of the PES pack and the dedicated playback data (low frequency data and screen area arrangement data) are detected by the PES pack header detector 72, and the information is stored in the address memory 75 . Here, it is judged whether the PES packet is an audio PES packet, a character PES packet, or a video PES packet. The video bitstream generator 73 erases the headers removed from the PES packets, and outputs a bitstream. Thereafter, based on the address information obtained from the address memory 75, the data rearranger 74 reorganizes the bit stream output from the mode converter 76, and outputs the bit stream in normal playback.

The output signal (control signal) of the mode converter 76 is supplied to a data rearranger 74 and a decodable determiner 84 . Here, the data rearranger 74 synchronizes with the low-frequency components and high-frequency components that are layered and rearranged for each region, and outputs the synchronized components. At the same time, only the low frequency component data or only the low frequency component data of the center portion of the screen is output to the variable length decoder 77 at the exclusive playback timing. In other words, at the time of normal playback, the low frequency components of the I picture and the P picture are recombined in the order of the areas on the screen. Then, the low-frequency components are synchronized with the high-frequency components, and the device operates to reconstruct the data in the original order of the image. At the exclusive playback timing, the low frequency component areas of the I picture and the P picture at the center portion of the screen are switched and output. At dedicated playback moments using only low frequency components, the PTS and DTS time stamps are not used.

On the contrary, the variable length decoder 77 and the decodable determiner 84 extract the boundaries of events in low frequency component regions marked by priority discontinuities, so that the data up to the boundaries are decoded and output to the switch 78. The effect of switch 78 is not to insert 0 at normal playback moment. At the exclusive playback time, the decodable determiner controls the switch 78 so that O is inserted into the high-frequency component after the priority interruption point.

The decoding working principle of the low frequency component is shown in Figure 44. Its explanation is omitted here. At this point, the reorganization of the screen area is the same as explained in Embodiment 10. Its explanation is omitted here.

The encoding region is defined at the boundary of the event, and it goes without saying that the event boundary can also be defined by other methods. For example, the coded region may also be separated by the end points of a predetermined number of events, and the data may be partitioned by subjecting the data to a coarse data conversion at quantization 54 to obtain the difference between the coarse and fine digital conversions. Data segmentation can also be accomplished by encoding images whose spatial resolution has been reduced to half due to thinning, images whose resolution has been restored from half the level to the original level, and images that differ from images with the original resolution. In other words, dividing highly-coded image data can be accomplished by dividing quantization and spatial resolution (in addition to dividing frequency regions).

At this point, in the case of frequency division, the more important data of an image is the data of the low frequency region. In the case of quantized partitioning, the more important data is the coarsely quantized coded data. In the case of spatial resolution partitioned data, the more important data is the data obtained by encoding the thinned image. In this case, the image is reproduced using only the data of these items, constituting more important data for areas more acceptable to humans. In other words, an efficiently encoded data is divided into more essential, important data and less important data, and at the time of playback from the disc, the essential and important data are reproduced.

Embodiment 11 is described in the recording section corresponding to the playback section. There may be a case where recording and playback constitute a pair of hard disk-like devices. Consider a case where it is assumed that data is recorded like a CD disc. Regarding the component reorganization for each area of the screen, it is obvious that the method of outputting the screen shown in Figs. 54, 55 of Embodiment 10 is available. Obviously, if the data of the vertical position of the tile in the header of the tile is used, the reorganization in the area unit on the screen can also be realized with the predictive data decoding circuit 82 and the frame memory 83. In Embodiment 11, only the basic I picture Data is divided by screen area. The data can be segmented by the low frequency of the P-picture or otherwise.

Example 12

Embodiment 12 of the present invention is explained with reference to FIG. 63 . Fig. 63 is a block diagram showing a digital video signal encoding processing unit in the digital video signal recording and reproducing apparatus. In Fig. 63, reference numerals 101 and 104 are preprocessors, 102 and 105 are motion vector detectors, 103 are resolution converters, 106 and 107 are subtractors, 108 and 109 are DCT circuits, and 110 and 111 are quantization 112 and 113 are variable length encoders, 114 and 115 are inverse quantizers, 116 and 117 are inverse DCT circuits, 118 and 119 are adders, 120 and 121 are image memories, and 122 and 123 are rate A controller, 124 is a resolution reverse converter, and 125 is a data reconstructor as a data recombination device. Fig. 63 shows the first encoding means and the second encoding means as an example. Specifically, the subtractor 106 outputs difference components between the first encoding means and the second encoding means in the two encoding processes.

Next, the operation of Embodiment 12 is explained. Video data is input to the resolution converter 103 in the order of interlaced raster scanning. The resolution converter 103 filters and thins the input video data to remove repetitive noise in the high-frequency region. Figure 64 is an illustration of the principle of resolution conversion for images. For example, in the case of data with horizontal 704 pixels and vertical 480 pixels, filter first, and then thin out to horizontal 352 pixels and vertical 240 pixels, that is, have half the resolution of the original respectively, thus transforming to low Resolution screen data.

This low resolution screen data is input to pre-processor 104 for conversion from raster scan to block scan. The order fed into the block scanner is the order of the DCT blocks. The coding of the I-picture does not use the output signal of the frame memory for intra-coding to perform inter-frame calculations.

In the case of an I picture, there is no output from the picture memory 121 which is an input terminal of the subtractor 107, so the video signal passes through the subtractor 107. This data is orthogonally transformed into frequency components by the DCT circuit 109 . This orthogonally transformed data is input to the digitizer 111 and digitized sequentially from the low frequency region in a zigzag manner. The digitally converted image data is converted into entropy coding by the variable length coder 113 and output to the data recombination device 125 .

Simultaneously, the data digitized by the digitizer 111 is inversely digitized by the inverse quantization means 115 . The image data is inversely converted by the inverse DCT circuit 117 from frequency component data into spatial components. The decoding of the I-picture does not perform inter-frame calculations using the intra-coded output signal from the frame memory. Therefore, in the case of an I picture, the data passes through the adder 119 because there is no input from the picture memory 121 of the adder 119 . Another output of the adder 119 is used as data stored in the image memory 121 . At least, I picture data or I picture and P picture data are stored in the picture memory. This is because I-picture and P-picture data are normally decoded as reference data in MPEG1 and MPEG2.

The image memory 120 inputs the output signal of the decoded data from the adder 118, and the result of the number of restored pixels completed by pixel interpolation by the resolution inverse converter, to store the decoded image data averaged with a weight. Regarding the weights, for the sake of brevity, there is a case where weight 1 is used as the output of the resolution inverse converter and weight 0 is used as the output of the adder 118 .

The output video signal is buffered in the pre-processor 101 for raster-scan-to-block-scan conversion. Then, at the subtracter 106, the video signal is subtracted from the signal data in the image memory 120 which has been subjected to the aforementioned low-resolution processing (referred to as resolution residual components). The resolution residual component is orthogonally transformed by the DCT circuit 108 from the low frequency region to a frequency region that can be properly converted by the quantizer 110 . This data is encoded into an entropy code by the variable length encoder 112 and output to the data rearranger 125 .

During this period, the data converted by the digitizer 110 is inversely converted in the inverse quantizer and inversely converted into data of the spatial region in the inverse DCT circuit 116 . The adder 118 adds the reverse conversion data from the image memory 120 through the low-resolution processing to the conversion data from the reverse DCT circuit 116, and obtains the decoding result of the data as a data (rather than the low-resolution ), which is formed by two layers, namely low-resolution data and residual component data. This layer is determined by the frequency of resolution transitions. It is possible to form this layer into three layers with two resolution conversions. In the same way, it is possible to prepare data at any number of levels in a similar manner.

With normal MPEG encoding, I pictures and P pictures are decoded and stored as decoded data, and B pictures are encoded by bi-directional prediction of I pictures and P pictures. In this method, I-pictures and P-pictures are coded, followed by B-picture processing.

The aforementioned encoding process for I-pictures, P-pictures and B-pictures is performed on both low-resolution components and high-resolution components. In this way, a sequence can be constituted in which the low resolution component R (hereinafter referred to as R component) and the resolution residual difference component S are arranged side by side. The operation is performed by the data rearranger 125, so that the data is arranged at the head of the GOP, and its position makes it easy for the optical head to seek. For example, data can be arranged as the sequence of FIG. 65 . When the data is arranged as shown in Fig. 65, the L component occupies half of the area, and the low-resolution component can be reproduced. The data amount of the resolution residual difference component is smaller than that of the non-resolution residual component, and the data can be effectively layered. In other words, here, the first coding device and the second coding device that code according to predetermined conditions are set to perform effective layering, and the second coding device codes the remaining part of the first coding device, for example, encoding the first coding device Encode video information other than video data.

Figure 65 shows an example of the data structure result. In FIG. 65, sequence a is a sequence generated by the encoding process of the twelfth embodiment. Sequence b is the sequence generated by the encoding process of another embodiment. Sequence c is a sequence generated by the encoding process of yet another embodiment. In sequence b, the symbols L are low frequency components and H are high frequency components. In sequence C, symbol C is the coded component of the coarse quantization and A is the residual component of the coarse quantization. As shown in sequence a in Fig. 70, the above operation is performed only for I pictures and P pictures. Only these components can be fixedly arranged at the head of the GOP.

In this method, when only the low-frequency components are fixedly arranged at the head of the GOP, the proportion of the L component to the whole is greatly reduced. In this way, the reading speed from the medium is shared, and the skip retrieval is easily realized. In addition, such as the sequence a, when only the R components of the I picture and the P picture are fixedly arranged in front of the GOP, the operation is performed so that only the low-resolution data of the I picture and the P picture are decoded. In the foregoing embodiments, in the case of the explanation, the thinning ratio is 1/2 times for the horizontal and 1/2 times for the vertical. Apparently, the ratio can be set to different values, but artificially setting the ratio can be added to the embodiment.

Encoding modes include MPG1, MPG2 and JPEG etc. In resolution layering, it is not necessary to use common coding techniques. This is because, in the case of encoding data at reduced resolution, it is possible to encode sufficiently corresponding to the MPEG1 mode. Furthermore, in JPEG mode, the superimposition of one frame on another constitutes a moving image. Therefore, it is possible to decode correctly even in the case where the data occupies a special position of the GOP. Furthermore, an explanation for two levels of resolution is given, but, obviously, a large number of layers can be utilized. The differential component can be encoded in the following manner: the data of the low-resolution component is encoded with the first encoding means of FIG. 63; the output signal of the first encoding means is interpolated; The differential components of the image are obtained from the subtractor 106; the differential components are encoded by differential component encoding means.

Frames read from the image memory are derived from predicted reference frames. Due to the presence of low-resolution frames, data is stored in memory (including memory addresses) by adjusting the time axis. Obviously, information means may be provided to add audio signals, headers etc. and error correction signal additional information to the differential components.

Example 13

Embodiment 13 of the present invention is explained with FIG. 66 . In Embodiment 13, the DCT block is divided into a low-frequency area layer and a high-frequency area layer, and only the low-frequency area is arranged at the head of the GOP. Fig. 66 is a block diagram of a digital video signal encoding processing unit. In Fig. 66, reference numerals 126 and 127 denote a first variable length encoder and a second variable length encoder, respectively. Parts in FIG. 66 that are similar or corresponding to those in FIG. 63 are marked with the same reference numerals and will not be described here.

Next, the working process is explained. This interpolation video data is a data item having an effective screen size of, for example, 704 pixels horizontally and 480 pixels vertically. The video is passed through and output because the decoding of the I-picture does not use the already intra-coded output of the frame memory for inter-frame calculations. Video data is orthogonally transformed into frequency components by the DCT circuit 108, and transformed from a low frequency region to block scanning. Then, the video data is converted from the low-frequency region to block scanning, and quantized appropriately by the quantizer 110 .

The data reorganization of the DCT coefficients within the DCT is shown in FIG. 67 . In FIG. 67, low-frequency components are located at the upper left, and high-frequency components are located at the lower right. Among the DCT coefficient data arranged in this DCT block, DCT coefficient data of a low-frequency region preceding the DCT coefficient data of a specific position (for example, the hatched portion in FIG. 67 ) is processed by the first variable-length coder 126 (low-frequency region extracting means). entropy coding, and output to the data rearranger 125. The second variable length encoder 127 performs variable length encoding on the DCT coefficient data following the above-mentioned special position DCT coefficient data. That is, in this manner, data is divided and encoded.

Regarding the encoding of motion vectors and DC components, encoding can be done with only the first variable length encoder. A second variable length coder is not required. This is because, at the time of normal playback, the output data of the first and second variable length encoders 126, 127 can be synchronized and encoded.

The judgment of the coding area is done at the fixed position of the DCT coefficient, and it can also be done in other ways. For example, encoding can also be judged by a fixed number of events. In other words, a unit that provides a variable-length code-Huffman code is an event. A coding region can also be set with a predetermined number of events (eg, three as a unit). In one example, at the output bit stream of the data rearranger 125, with the reorganization of the sequence b of FIG. 65, only when the first half of the low frequency region is read, the low frequency region image can be reproduced. In a sequence such as the sequence b in Fig. 70, the coding region can be judged in a variable length manner.

Meanwhile, the data quantized by the quantizer 110 is inversely quantized. Then, the data is inversely transformed into spatial region data by an inverse DCT circuit 116 . The I picture is coded without inter-frame calculations using the output signal of the frame memory subjected to intra-frame coding. Thus, there is no input to picture memory 120 from adder 118 in the case of an I picture. Therefore, data is allowed to pass through adder 118 . Adder 118 is used for data stored in image memory 120 .

At least, I pictures and P pictures are stored in the picture memory because I pictures and P pictures are normally used as reference data for describing B pictures in MPEG1 and 2.

When constructed in this way, the proportion of the L component is greatly reduced, allowing data to be read from the medium, enabling jump retrieval. As will be shown later, only when I pictures and P pictures are fixedly arranged, the apparatus of the present invention works so that only data of low frequency components can be easily decoded. Since the amount of data in the high-frequency area is smaller than that in other areas, it is possible to realize efficient data formation compared to extracting data in the low-frequency area and storing the data before storing the data in all areas.

When the coding of the I picture is completed by allowing the I picture to pass through the subtractor 106, the B picture is coded with bi-directional prediction with respect to time using the last P picture of the GOP. The output of the pre-processor 101 is compared with the data of the reference frame memory (the arrow is omitted in the figure), the motion vector is detected, and the predicted mode and frame structure are judged. Based on the judgment result, the output of the pre-processor 101 and the data of the reference frame memory are most advantageously consistent with each other, and the data of the reference frame memory is read as data from the forward direction part and the backward direction part of the frame memory 120. The data read in this way is then subtracted in the subtractor 106 from the output of the preprocessor 101 for the B picture (this result is the time residual component of both the P picture and the I picture). A DCT calculation is performed on the temporal residual component, and the result is quantized and variable-length coded.

Example 14

Embodiment 14 of the present invention is explained on the basis of FIG. 68 . In Embodiment 14, data is divided into layers of DCT coefficient coarse quantized components and rough residual differential components (example of coarse quantized components), thereby arranging the coarse digitized components at the front of the GOP. Fig. 68 is a block diagram showing a digital video signal encoding processing unit. In FIG. 68, reference numeral 128 is a subtracter, and 129 is an adder. Parts in FIG. 68 that are similar or corresponding to those in FIG. 63 are marked with the same reference numerals and will not be described here.

Next, the operation of Embodiment 14 is explained. For example, the effective screen size of its interpolated input image is 704 pixels horizontally and 480 pixels vertically. I-pictures are decoded without inter-frame calculations using the output signals of the intra-coded frame memory. Therefore, in the case of an I picture, there is no input to the picture memory, and thus no input to the subtractor 106, and the video signal passes through the subtractor 106 as a result. This data is orthogonally converted into frequency components by the DCT circuit 108, and converted from the low frequency region to block scanning. Then, quantizer 110 performs appropriate coarse quantization, reducing the amount of encoded data to less than half. The quantized data is encoded into an entropy code by the variable length encoder 112 and output to the data reorganization device 125 .

At the same time, inverse quantization is performed on the data converted by the quantizer 110 (the result of which is taken as a rough quantization result). The inversely quantized data is sent to another encoding processing unit (indicated by a dotted frame in FIG. 68). Meanwhile, the data is back-transformed into spatial zone data. Here, coding of I pictures is described. Although the image memory 121 does not output, under normal circumstances, the encoding result of this encoding processing unit is stored in the image memory 121, like this, the motion vector detection is carried out to the data in the motion vector detector 102, and the predetermined mode and the DCT block are determined model. The position data suitable for the judgment mode is sent to the subtraction input side of the subtracter 107 .

DCT is performed on the output of the subtracter 107, and the residual difference (which is the coarse quantization residual difference) is determined using the coarse quantization and the result of the subtractor. The residual difference of the coarse quantization is finely quantized (considering the control of the coding amount, the fine quantization is the same level as the normal coding). Variable length coding is performed while inverse quantization, inverse DCT, and decoding are stored in the image memory 121. This encoding result and the roughly quantized encoding result determine the position of the necessary data, the header to be added, and the like.

As an example of this output data, sequence C shown in Fig. 65 is used. The decoding result of the coarsely quantized picture can be obtained by reading only the first half of the GOP. Furthermore, since the coarse quantization residual difference data is relatively small compared with the fine quantization data, a more data-efficient configuration can be obtained than storing the extracted coarse quantization data before the coarse quantization data.

In another example, variable processing of the permutation of sequence C shown in FIG. 70 can be accomplished. With such a configuration, the proportion of the C component (encoded component subjected to rough quantization) to the whole is greatly reduced, thereby satisfying the reading speed from the medium, enabling skip search, and the like. As will be described later, only when I pictures and P pictures are fixedly arranged, the apparatus of the present invention operates such that only roughly quantized data of I pictures and P pictures are decoded.

When the coding of the I picture is completed by passing the I picture through the subtractor 106, the B picture is coded in a bidirectional predictive manner with the last P picture of the leading GOP according to time. The output signal of the preprocessor 101 is compared with the data of the reference frame memory (the arrow in Fig. 70 is omitted), so that the motion vector is detected, and the predictive mode and frame structure are judged. On the basis of the judgment result, the output signal of the pre-processor 101 is most consistent with the data in the reference frame memory, and the data in the reference frame memory read out is used as the forward part and the reverse part of the image memory 120, The thus read data is subtracted from the output result of the preprocessor 101 of the B picture in the subtractor 106 (this result is the time residual difference component of both the P picture and the B picture). The data of the forward part and the reverse part are read from the image memory 121, and the data and the output signal of the preprocessor 101 are subtracted in the subtractor 107 for orthogonal transform entropy coding. The same process is also done for the encoding of P-pictures.

Fig. 69 shows encoded data statistics. The figure shows the number of codes when the number of frames of GOP is as follows: N=15, and the period of I picture and P picture: M=3. Also shown in FIG. 69, I pictures and P pictures account for about 50% of the whole. When layering with resolution, frequency, at least about the quantization of this part of the I picture, the amount of codes to be reproduced is further reduced, and the travel time of the optical head can be shortened, thereby facilitating functions such as jumping and searching.

Fig. 70 shows the processing sequence in the above case. In Fig. 70, the arrangement of the I picture, the P picture and the B picture of the original picture is coded, and only the I picture and the P picture of the aforementioned picture are processed according to Embodiment 12, 13, 14, and B-pictures are coded without layering. The sequence for processing I-pictures and P-pictures according to the processing procedure of Embodiment 12 is sequence b, while the sequence for processing I-pictures and P-pictures according to the processing procedure of Embodiment 14 is sequence C.

In each sequence, the data is formed by fixing the I picture component and the P picture component of each low resolution component (R), high frequency component (L) and coarse quantization component (C) by the data rearranger 125. And arranged to be completed in front of the GOP. Regarding the sequence a, the low-resolution images of the I picture and the P picture can only use the low-resolution components of the I picture and the P picture (in the sequence a of Figure 70, the core area part of the data reconstruction) Decoding, as a result, the device of the present invention can easily implement skip scanning. In fact, the data in the non-core area is not required to be arranged according to Figure 70. Obviously, data can be arranged in the order of frame numbers at the time of encoding.

About sequence b, can use low-frequency component (in the sequence b of Fig. 70, the core area part after data recombination) to form the low-frequency component of I picture and P picture, so, the device of the present invention can finish skip scanning easily. Regarding the sequence C, only the roughly quantized components of the I picture and the P picture (the core area part after the sequence C data recombination in FIG. 70 ) can decode the roughly quantized images of the I picture and the P picture, so this The inventive device can easily realize skip scanning.

For example, in the structure of FIG. 63, the encoding loop including the preprocessor 104 is not used for the B picture. Therefore, the operation mode of the apparatus of the present invention is to use only the encoding loop including the preprocessor 101 for data encoding. In the structure shown in FIG. 66, all frequency components are coded by the first variable length coder 126. In the structure shown in FIG. 68, fine quantization is performed in the digitizer 110 to encode data.

Optimally, basic data such as low-frequency portion data can be collected at the head of the GOP. Obviously, these data can be slightly shifted. It overlaps with the head of the unit constituting the error correction code. Arrangement of basic data and error correction code units in this way can also be realized in the same way in other embodiments.

Example 15

Embodiment 15 of the present invention will be explained with reference to FIGS. 71 and 72. FIG. Fig. 71 shows the structural arrangement of the DCT block, and an example of the structural outline of the frequency components of the bit stream in one block. FIG. 71A shows a macroblock of the entire DCT block structure, consisting of a macroblock header, DCT blocks Y1-Y4 of luminance signals, DCT block U1 of color difference signals (B-Y), and DCT block V1 of color difference signals (R-Y). Fig. 71B shows a low-frequency component of a macroblock with respect to the structure of a low-frequency component DCT block. It consists of macroblock header, luminance signal and DCT blocks Y1-Y4, DCT block UIL of color difference signal (B-Y) and DCT block VIL of color difference signal (R-Y).

The high-frequency macroblock shown in FIG. 71C is a structure related to high-frequency component DCT blocks, and consists of luminance signal DCT blocks Y1H-Y4H, color difference signal (B-Y) and DCT block U1H, and color difference signal (R-Y) and DCT block V1H. Fig. 71D shows the principle of frequency component data arrangement in the bit stream of one block. 72A and 72B are block diagrams showing a digital video signal decoding processing unit and its working principle. In Fig. 72A, reference numeral 130 is a mode converter as a mode conversion device, 131 is a data rearrangement device as a data reorganization device, 132 is a decodable determiner, 133 is a variable length decoder, and 134 is a a switch. The decodable determiner 132 and the switch 134 constitute data manipulation means. Reference numeral 135 is an inverse quantizer, 136 is an inverse DCT circuit, 137 is an image memory, 138 is an adder, and 139 is an inverse scan converter.

Next, the operation is explained. The data in Fig. 71 is a code structure combination of, for example, 8 bits (one byte) in the vertical direction. In each macroblock, information is specified with a macroblock header. This information refers to incremental address, quantization scale code, motion vector, flag bit, macroblock model, etc.

The coded data for each DCT block follows the macroblock header. The way to encapsulate this data is to make bytes out of the bit stream, arranging each byte in sequence. Because each DCT block has a variable code length, block boundaries and boundaries between headers and data are not completed in byte units. It often happens that the boundary is in the middle of a unit byte. The data of each block has a variable length, and the low frequency area is set on the side closer to the header of the macroblock.

This data is divided into low-frequency components (L) and high-frequency components (H), and the coded data shown in FIG. 71B and FIG. 71C are constituted by setting a fixed-length code amount irrespective of an event as a maximum value. (The event is a unit that provides a variable-length code. In the case of a DC component, the DC component constitutes an event, while in the case of an AC component, the combination of non-zero DCT coefficients and the operation length constitutes the execution of the operation-length code. An event. The completion of an event is indicated by a code EOB at the end of the block).

Next, the operation of Fig. 72 will be described. Initially, a mode signal from a microcomputer or the like is input to the mode switch 130, a signal indicating skip scanning is being performed, or normal continuous playback is being performed. Simultaneously, the amplifier amplifies the reproduction signal from the disk, and performs error correction by digitally demodulating an output data which is reproduced by a synchronized clock signal and output from a PLL etc. And obtained, then, the audio signal is separated from a certain system layer composed of the video signal data and the audio signal. Thereafter, the bit stream of the video signal is extracted and input to rearranger 131 .

The output (control signal) of the mode converter 130 is supplied to a data rearranger 131 and a decodable determiner 132 . The data rearranger 131 receives the control signal, reconnects the data, divides the data into the L component and the H component shown in FIG. 71, and outputs only the L component to the variable length decoder 133. The variable length decoder 133 and the decodable determiner 132 together extract event boundaries in the L component area. The portion before the boundary is decoded and output to switch 134 . Switch 134 is connected so that, during normal playback, no O's are inserted. The switch 134 is controlled by the decodability determiner 132, and inputs the decoded low-frequency component to the DCT block. At the same time, the entire DCT block is constructed such that O is inserted into the high-frequency side of the DCT block.

When decoding, the DCT block data formed by the aforementioned method is subjected to inverse DCT processing. Then, the reading of the image memory 137 is controlled according to the condition of each image to be added by the adder 138 . In the case of an I picture, the output of adder 138 is passed through. In the case of a P picture, the P picture is corrected only by the motion vectors of the I picture and the P picture to be added. In the case of a B picture, the B picture is corrected only by the motion vectors of the I picture and the B picture to be added.

The DCT mode and predictive mode motion vector at this time are controlled based on the information obtained by decoding the header code. The motion-compensated predicted data is decoded and stored in the image memory 137 according to the foregoing processing. The pictures are restored to the order of the original composition of the GOP. Inverse scan converter 139 converts buffering and block search to raster scan in image output order.

Embodiment 15 proposes a fixed length, even though this embodiment is shorter than macroblock jumps or data of a predetermined length. However, when the length is shorter than the fixed length, by detecting the EOB each time, the L component can definitely be taken out. Obviously, even in the case where the L component data is connected to subsequent blocks, no problem arises. When the length exceeds a predetermined length, EOB can be added to the event boundary as L component data. Apparently, means for adding additional information such as audio signal, header, etc., and connection code are also provided to add it to the data in the high-frequency region, which is not specifically described in the figures of the foregoing embodiments.

Example 16

Next, Embodiment 16 of the present invention will be described with reference to FIGS. 73 and 74. FIG. 73A and 73B are block diagrams of a digital video signal encoding processing unit. Explain how it works. In Fig. 73A, reference numeral 140 is a rate controller. Here, the first variable length encoder 126 and the second variable length encoder 127 serve as encoding means. Parts similar to or corresponding to those in Fig. 63 are given the same reference numerals.

Next, its working process is explained. The interpolated input image data is buffered by preprocessor 101, which converts raster scan to block scan. The decoding of the I-picture does not require inter-frame calculations using the output signal of the intra-coded frame memory. Therefore, in the case of an I picture, there is no input signal to the picture memory 120 as an input to the subtractor 106, and therefore, the video signal passes through the subtractor 106.

This data is quadrature converted to frequency components by the DCT circuit 108 and converted to block scanning in the low frequency region, and quantized appropriately by the quantizer 110 . The low-frequency region data before the DCT coefficient data at a specific position in the quantized data is entropy encoded and output to the data rearranger 125 via the first variable length encoder 126 .

The second variable length encoder 127 performs variable length encoding on the DCT coefficients after the above-mentioned special position. As for motion vector and DC component coding, at least only the first variable length coder can be used. The EOB is required to be added to both the L and H components even within a block, so the boundaries of the L components change at a rate that is not code limited. By temporarily arranging the EOB code at the boundary portion of the L component and the H component, the boundary of the L component can be changed at a certain rate.

Meanwhile, the data quantized by the quantizer 110 is inversely quantized by the inverse quantizer 114 and then inversely transformed into spatial components by the inverse DCT circuit 116 .

The decoding of the I-picture does not require inter-frame calculations using the output signal of the intra-coded frame memory. Thus, in the case of an I picture, data passes through adder 118 since no signal is input from picture memory 120 to adder 118 . The output of adder 118 stores data in image memory 120 . At least the data of the I picture, or the data of the I picture and the P picture are stored in the picture memory 120 . This is because both I picture and P picture data are used for normal decoding of B pictures as reference data in MPEG1 and MPEG2.

At the end of the encoding of the I picture, the B picture is bidirectionally predictively encoded using the last P picture preceding the GOP. Then, the output of the pre-processor 101 is compared with the data of the reference frame memory (arrows are omitted in the figure) to retrieve the motion vector and judge the predicted mode and frame structure. According to the judgment result, the data in the reference frame memory most suitable for the output signal of the pre-processor 101 is read from the image memory 120, and the data is divided into a forward part and a reverse part. The data is subtracted in the subtractor 106 from the output result of the preprocessor 101 (the result refers to the time residual difference between the P picture and the B picture). The time residual difference component is subjected to DCT processing, quantization and variable length coding processing.

In the case where the data is divided into low-frequency and high-frequency regions, the aspect of rate frequency components becomes uncertain. Therefore, the controllable range of the optical head driver cannot be fully compensated due to the uncertain data rate in the low frequency region. Here, the rate controller 140 makes the low frequency component region variable. The speed controller 140 controls the speed, so the size of the low frequency region is variable with respect to the target speed, as shown in FIG. 73B.

In other words, while monitoring the output of the first variable length encoder 126, the rate controller 140 reduces the size of the area occupied by the low frequency region data when the monitored output is greater than the target rate set by the application. When the code amount of the first variable length coder 126 is small, the rate controller 140 expands the area of the low frequency region. In this way, when monitoring the code amount, the rate controller 140 changes the first variable length coder appropriately. 126 and the area occupied by the low frequency region set in the second variable length coder 127.

In addition, for example, instantaneous encoding can be performed to determine a standard to set the area occupied by the low-frequency area according to which area has a larger code number and which area has a smaller code number, so as to set a target rate.

Fig. 74 is a block diagram of a digital video signal decoding processing unit. A processing procedure for decoding data encoded as above is shown. In FIG. 74, reference numeral 141 is an EOB retrieval unit. Wherein, the parts similar or identical to those in Fig. 72A use the same reference numerals. A mode signal indicating that a skip search is being performed or a normal continuous playback is being performed is input to the mode converter 130 from a microcomputer or the like. Simultaneously, the playback signal from the disc is amplified by an amplifier, and the playback signal is digitally demodulated by differentially using the clock of the PLL. Error correction is performed, and the video bit stream to be input to the data reassembler 131 is extracted, thereby separating the audio signal from a system layer. The output of the mode converter 130 serving as mode converting means is supplied to a data rearrangement unit 131 and a decodable determiner 132 . The data rearranger 131 gets this control signal and connects the data before the separation of the L component and the H component. Also, only the L component is output to the variable length decoder 133 (decoding means), and is not connected to the H component.

Theoretically, the L component is never split in the middle of an event. It can be considered that when signals such as skip searches are not good enough, the variable length decoder 133 and the variable decision unit 132 determine the boundary of the event, so that the part before the boundary is decoded and output to the switch 134 . The switch 134 operates in such a way that it is always "on" when the signal quality of the reproduced data is as good as during normal reproduction. Here, the decodability determiner 132 and the switch 134 constitute data manipulation means.

The switch 134 is controlled by the decodable determiner 132 so that an O is inserted on the high frequency side of the block, leaving the successfully decoded low frequency components, thereby forming a DCT block. The data is then subjected to an inverse DCT, and in the case of an I-picture, the output of adder 138 is passed. In contrast, in the case of a P picture, the data is corrected and added to the motion vector portion of the reference I picture. As for the B pictures, the reading from the picture memory 137 is controlled and added in the adder 138, so that the B pictures are partially corrected by the motion vectors of the I pictures and P pictures to be added. The DCT mode and the predictive mode motion vector are controlled by a code solved for the optical head. The motion-compensated predicted data is decoded in this way and stored in the image memory 137 . Then, the images are reconstructed into the order of the original composition. The inverse scan converter 139 buffers data and converts the data in output order of blocks to raster scan.

In the above explanation, control of the size of the DCT coefficient is taken as an example for illustration. It can also be replaced by a control of the number of events. In this case, there may be an event that the L component does not reach a predetermined number, and EOB is added. However, since the EOB retrieval unit 141 monitors the occurrence of EOB, the L component can be retrieved with certainty. Specifically, the data is constituted from the data of the low-frequency region, the data of the high-frequency region, and the EOB, respectively. That is, the data reorganizer 131 and the EOB retrieval unit 141 constitute a data reorganization means.

Since the energy after DCT is very small, it is obvious that the L and H components are best coded in the same way as non-coded blocks. Regarding the H component, data other than the L component is desirably operation-length encoded. By setting the L component to 0, the H component can be encoded. Since it is possible to complete the structure with the same variable length decoder of normal MPEG, it can be simplified in circuit.

Example 17

Embodiment 17 of the present invention is explained with reference to FIG. 75 . Fig. 75 shows a block diagram of a digital video signal decoding processing unit. In FIG. 75, reference numeral 142 is a multiplexer, 143 is a switch, 144 is a first variable length decoder, 145 is a second variable length decoder, 146 is a first inverse quantizer, 147 is The second inverse quantizer, 148 and 149 are adders, 150 and 151 are image memories, and 152 is a resolution inverse converter. Fig. 75 also shows a low-resolution decoding unit as the decoding means. Similar or identical parts to those in FIG. 72A are designated by the same reference numerals, and explanations thereof are omitted.

Next, the operation of Embodiment 17 is explained. The content shown in FIG. 75 can be regarded as the process of processing the reproduction signal video data block of the disc in the case where the coded data described in FIG. 68 is recorded on the optical disc or the like.

A mode signal indicating, for example, a state where skip search or normal continuous playback is being performed is input from a microcomputer or the like to a mode switch 130, which is mode switching means. Simultaneously, the playback signal from the disc is amplified by an amplifier, and the playback signal is differentiated with a clock of the PLL for digital demodulation. Then, the audio signal is separated from the system layer to extract the video bitstream.

This extracted video bitstream is input to the multiplexer 142 . The multiplexer 142 sends the low-resolution component data to the second variable-length decoder 145 , while sending other data to the first variable-length decoder 144 via the switch 143 .

The switch 143 is controlled by the mode converter 130 . As a mode, although only low-resolution reproduced image output is required at the time of jump search, etc., the switch 143 is operated to suspend the transmission of redundant data during the reproduction of the residual difference in resolution.

The second variable length decoder 145 decodes the Huffman code and the operational length code, and then inversely quantizes for the second inverse quantizer 147 , and thereafter is converted from the frequency component to the space component by the inverse DCT circuit 136 .

For an I picture, the converted data is stored in the picture memory through the adder 149 . In the case of a P picture, the first frame P picture is read from the I picture in the picture memory, and the second frame P picture or subsequent pictures refer to those stored in the picture memory and partially corrected by the passive vector Subsequent images for motion compensated prediction in adder 149. In the case of B pictures, the same operation is performed for I pictures and P pictures.

In FIG. 75, the variable length decoder outputs the motion vector, the quantization parameter of inverse quantization, and the predictive pattern. These motion vectors, quantization parameters and prediction modes are the same as those in FIG. 74 . A loop shown by a dotted line frame in Fig. 75 is a constituent unit for decoding a low-resolution component. Since the decoded result is interpolated between pixels by the resolution inverse converter 152 (interpolation video generating means), the decoded result is compensated as a resolution residual difference component, and the decoded result is input to the image memory 150 .

The decoding result of the resolution residual component during normal playback is combined with the result of the low frequency component (or according to the time-division processing result in the case of time-division operation for decoding the resolution residual component) as an image output by the reverse scan converter. The data can be decoded into frequency components by the first variable length decoder via the switch 143 . The first inverse quantizer 146 inversely quantizes the data, and the inverse DCT circuit 136 decodes the data into resolution residual component data of the spatial region.

The image memory 150 is associated with the pixel interpolation data of the low-resolution components, the P-picture is associated with the I-picture, and the B-picture is associated with the I-picture and the P-picture, so that the data passive vector part is corrected in place, As a result, data is read from image memory 150 and adder 148 performs decoding of the motion compensated prediction.

In the case of jump search, in order to prevent the residual resolution difference from being reproduced midway, the switch 143 prohibits output of redundant data from the inverse DCT circuit 136, that is, interrupts the output of the residual resolution difference. Therefore, only the pixel-interpolated low-resolution component data is output through the image memory 150 and the inverse scan converter 139 (the operation is the same as setting a switch at the input portion of the image memory 150).

Example 18

Embodiment 18 is explained with reference to Figs. 76 and 77 . The block diagram of Fig. 76 is a block diagram of a GOP address generation unit and a disk control unit, which particularly shows a processing block diagram in the case where the aforementioned rate information is recorded on successive headers. In FIG. 76, 153 is a register, 154 is a GOP address calculator, and 155 is an optical head/disk rotation control switch as optical head position switching means and disk rotation control switching means. Fig. 77 is a block diagram showing a GOP address generation unit and a disc control unit including playback processing, and particularly shows a structure for performing GOP playback of a disc, in several positions on such a disc The aforementioned rate information is collected on . Referring to Fig. 77, 156 is a playback amplifier, 157 is a digital demodulator, 158 is an error corrector, 159 is a system layer processor, and 160 is a rate data memory. The system layer processor 159 and the rate data memory 160 constitute data rate information extraction means. 161 is a GOP number counter. The GOP address calculator 154 and the GOP number counter 161 constitute position information calculating means.

Next, the operation of Embodiment 18 is explained. By making the rate variable for each GOP, as usual, the overall rate of a program can be optimized, and as a result, the quality of the image can be considered to be improved. However, it does not mean that the data falls at the head of the GOP until the content of the data is observed. Also, in the case where it is desired to play back from that point again halfway through software that has been matched, the only way is to detect the start position by constantly retrieving the data on the disc.

In this case, in the initial stage, set the rate control of variable rate as discrete rate target, such as 1M bit, 1.5M bit, 2M bit, 2.5M bit, 3M bit, etc., thus, in all GOP Each rate information is recorded on the disc. Specifically, it should be most effective when the rate information of each GOP is recorded on the TOC (Table of Contents: recording area is assigned to the very beginning of the disc so that title, recording time, etc. are recorded), half TOC, etc.

The rate information about the GOP can be combined with the sequence header of the video bitstream. For example, two hour software 14.4K pieces of GOP (two hour software 14.4 kpieces of GOP). When the rate information can be divided into 5 rates, the rate information at this time can be represented by 3 bits. Therefore, all GOP rates can be recorded on the disc with 5.4K (14.4K x 3 bits ÷ 8 bits/byte).

High-speed access can be realized by storing the rate information of each GOP in the rate data memory 160 of Fig. 77, and adding an information length corresponding to the value.

The apparatus of the present invention is explained with reference to FIG. 76 . Decode the Huffman code and the operation length code, and decode the header, so that the motion vector and the image type are judged.

Simultaneously, the sequence header is decoded so that the rate information is input to the GOP address calculator 154 . In addition, the address information of the GOP being accessed at that time is stored in the register 153, so that the address before the next GOP to be accessed is calculated and stored in the register 153. At the same time, the position of the optical head is determined based on the address by turning the optical head/disk control switch 155 to the head of the next GOP to be accessed. Then, the control signal for the next access is calculated from the difference between the GOP being accessed and the previous address to be accessed. According to this control signal, the position control of the optical head driver and the control of the disk rotation are realized.

Referring to Fig. 77, the playback process will be described. Control the optical head and the rotation of the optical head, directly or indirectly read the address from the TOC area or the area of the corresponding TOC (after specifying the speed information to describe the address, access this address part to read the speed information). Then, the reproduced signal from the optical head is amplified by the reproduced amplifier 156, and this signal wave is checked by the digital demodulator 157, and the signal is differentiated into a digital signal for digital demodulation.

The reproduced signal digitally demodulated into a digital signal is input to an error corrector 158 to correct errors in the reproduced signal. The error-corrected data is separated in the system layer processor 159 into audio bitstream, video bitstream and other data items.

For example, determine which type of data this signal belongs to (binary data such as AV (video and audio) data, file data, and programs), and cut and divide stream channels accordingly. In such processing, the aforementioned rate information is stored in the rate data memory 160 .

Instead, the GOP number counter 161 is used to generate information on the number of GOPs to be processed. According to the address calculated by the GOP address calculator 154, the optical head driver and the disk rotational speed are controlled.

In the above description, an example was given in which the GOP number counter receives a signal from the system layer processor 159 . In the case of using an I/F (such as a microcomputer) instead of processing or operating from playback to skip scanning, it is more efficient to process video bit streams with input data from a variable length decoder or the like.

Example 19:

Embodiment 19 of the present invention is described with reference to FIGS. 78 , 79 and 80 . Its working will be explained as follows. Fig. 78 shows a signal processing unit in which division according to the frequency of the digital signal playback unit and division according to quantization are performed, which is a block diagram, in which case data rate information is collected and recorded on an optical disk several places on the Parts that are the same as or corresponding to those in the drawings in the foregoing embodiments are denoted by the same symbols.

The optical head or head rotation is controlled so that data can be read directly or indirectly from the TOC area or an area corresponding to the TOC area. The playback signal is amplified by playback amplifier 156 . The signal is then detected by digital demodulator 157 . As a result, the reproduced signal which has become digital data is input to the error corrector 158 to correct errors contained in the reproduced signal. Error-free data is divided into audio bitstream and video bitstream by the system layer processor 159, and other data are processed in the same way.

For example, it is determined which of AV data, text data, or binary data such as a program the signal is to separate data stream channels. The aforementioned data rate information is stored in the data rate data memory 160 . Instead, information corresponding to the GOP number that needs to be processed is generated by using the GOP number counter 161 . Subsequently, the address is calculated by the GOP address calculator 154 . On the basis of the address calculated by the GOP address calculator 154, the rotational speed of the optical head actuator and the optical disk is controlled.

In this way, during skip playback, skips are made on the disc to find addresses to be accessed. When skipping is performed to access a desired GOP, low frequency area data obtained with the structure described in Embodiment 16 is reproduced and the next address is calculated in the same manner as described on the screen.

A mode signal representing a state that data is skipped or that normal playback is performed is input into the mode switcher 130 from the microcomputer. As described above, the video bitstream to be input into the data rearranger 131 is extracted. The output of the mode switcher 130 is supplied to a data rearranger 131 and a decodable determiner 132 . The data rearrangement unit 131 operates by obtaining such a control signal, so that the data before being divided in the L component and the H component in FIG. 71 are reconnected. Otherwise, the data rearranger 131 only outputs the L component to the variable length decoder 133 without linking the L component to the H component.

In Embodiment 19, theoretically, the L component does not appear to be cut off midway. However, considering that the semaphore with undesired skip search etc. is decoded, the boundary of the event is confirmed by the variable length decoder 133 and the decodable determiner 132, so that the part up to the boundary can be decoded and is output to switch 134.

The switch 134 is controlled by the output of the decodable determiner 132 so that zeros are input to the high frequency side of the block from the low frequency components that are fully decoded constituting the DCT block. Subsequently, the read-out of the picture memory 137 is controlled, and the addition is provided by the adder 138, so that the data will be inversely DCT-transformed as follows, and the output of the adder 138 is passed in the case of an I picture, At the time of the P picture, the data is corrected by the motion vector portion to be added, and at the time of the B picture, the data is corrected by the motion vector portion from the I picture and the P picture and added.

Also, at this point in time, the DCT mode and prediction mode motion vector are controlled by decoding the header code. In this way, data for motion compensation prediction is decoded and stored in the image memory 137 to form images in the original composition order. In the reverse scan converter 139, data is buffered to convert the block scan data into raster scan data in the output order of the image. In addition, the switch 134 is not connected so that zeros are inserted during normal playback, but is controlled to work and only playback data is played back.

After the data is divided and encoded into the low frequency area and the high frequency area, the following will appear: a quantization table that places emphasis on the low frequency side, a quantization table that places emphasis on the high frequency side, and a quantization table that places emphasis on the high frequency side regardless of the frequency area. Pretty average refinement. This can be achieved when two sets of variable length decoders and inverse quantizers are used as the native decoder shown in FIG. 68 . At this time, the data rearranger 131 will be a multiplexer.

Next, the operation of the nineteenth embodiment will be explained on the basis of FIG. 79. FIG. Figure 79 shows a signal processing unit in which division is carried out according to the bit length of the digital signal playback unit, which is a block diagram for a few specific places where the data rate information is collected and recorded on the disc with respect to the aforementioned data rate The playback processing of the case explains the embodiment. For example, at the very beginning of optical disc reproduction, the optical head or head rotation is controlled, so that data can be read directly or indirectly from the TOC area or an area corresponding to the TOC area. The reproduction signal from the optical head is amplified by the reproduction amplifier 156, so that it can be detected by the digital decoder 157 and differentiated into a digital signal for digital demodulation.

As a result, the reproduced signal which has become digital data is input to the error corrector 158 for error correction of the reproduced signal. Error-free data is divided into audio bitstream and video bitstream, and other data are processed in the same way.

For example, this signal determines whether the data is binary data such as video and audio data, text data, or programs to separate data flow channels. From this data, the aforementioned data rate information is stored in the data rate data memory 160 . At the same time, the GOP number counter 161 generates information as the number of GOPs to be processed, and the GOP address calculator 154 calculates the address to control the rotation speed of the actuator and the optical disc.

In this way, the address of the GOP to be accessed at the time of skip playback is searched by skipping on the disc. When the desired GOP is accessed, the next address will be calculated in the same manner, and at the same time, the low-frequency region data obtained with the structure described in Embodiment 15 is reproduced to reproduce the GOP on one screen. data.

A mode signal representing a state that data is skipped or that normal continuous playback is performed is input to the mode switcher 130 from a microcomputer or the like. The video bitstream is extracted and input into the data rearranger 131 . The output of the mode switcher 130 is supplied to a data rearranger 131 and a decodable determiner 132 . The data rearrangement unit 131 operates by obtaining such a control signal that the data before being divided in the L component and the H component in FIG. 71 are reconnected. Otherwise, the data rearranger 131 only outputs the L component to the variable length decoder 133 without linking the L component to the H component.

The boundary of an event among the L components is extracted by the variable length decoder 133 and the decodable determiner 132 , so that the part up to the boundary can be decoded and output to the switcher 134 . The switch 134 is controlled by the output of the decodable determiner 132 so that zero values from the low frequency components that are fully decoded to form the DCT block are inserted at the high frequency end of the block. The data is subjected to an inverse DCT transform. In the case of an I picture, the output of adder 138 is passed. In the case of a P picture, the image is corrected by adding the portion of the motion vector in the reference I picture. The readout of the image memory 137 is controlled and added by the adder 138 so that the data can be corrected by the motion vector portion.

Also, at this point in time, the DCT mode and prediction mode motion vectors are controlled by the decoding of the header code. In this way, data for motion compensation prediction is decoded and stored in the picture memory 137 to form pictures in the order in which GOPs were originally formed. In the inverse scan converter 139, data is buffered to convert block scan data to raster scan data. In addition, the switch 134 is turned off so that zeros are inserted during normal playback, and when turned on, only the playback data is played back.

Next, the operation of Fig. 80 is explained. Figure 80 shows a signal processing block in which data is divided according to the resolution of the digital signal playback portion, which is a block diagram for several specific places where information is collected and recorded on an optical disc with respect to the aforementioned data rate information Example of replay processing in case of explanation. At the very beginning of disc playback, the head or head rotation is controlled so that data can be read directly or indirectly from the TOC area or an area corresponding to the TOC area. Thus, the digital decoder 157 can be used to detect the difference into a digital signal for digital demodulation.

As a result, the reproduced signal which has become digital data is input to the error corrector 158 for error correction of the reproduced signal. Error-free data is divided into audio bitstream and video bitstream, and other data are processed in the same way.

For example, this signal determines whether the data is binary data such as video and audio data, text data, or programs to separate data flow channels. From this data, the aforementioned data rate information is stored in the data rate data memory 160 . At the same time, the GOP number counter 161 generates information as the number of GOPs to be processed, and the GOP address calculator 154 calculates the address to control the rotation speed of the actuator and the optical disc.

In this way, the address of the GOP to be accessed at the time of skip playback is searched by skipping on the disc. When the desired GOP is accessed, the next address will be calculated in the same manner, and at the same time, the low-frequency region data obtained with the structure described in Embodiment 15 is reproduced to reproduce the GOP on one screen. data.

A mode signal representing a state that data is skipped or that normal continuous playback is performed is input to the mode switcher 130 from a microcomputer or the like. The video bitstream is extracted and input to multiplexer 142 . The multiplexer 142 supplies the low-resolution component data to the second variable-length decoder 145 , while supplying other data to the first variable-length decoder 144 via the switcher 143 . The switcher 143 is controlled by the mode switcher 130 . Although the playback image output of only the low-resolution component is required as a mode in a skip search or the like, the switch 143 is turned off during playback of the rest of the resolution components. In addition, when the playback operation is in progress, the switch 143 is turned on, wherein good signal transmission quality can be obtained for normal playback.

The second variable length decoder 145 decodes Huffman codes and run length codes. The data inversely quantized by the second inverse quantizer 147 will be converted from the frequency domain to the space domain by the inverse DCT circuit 136 . When the data is an I picture, the data passed through the adder 149 will be stored in the picture memory. When the data is a P picture, the P picture is fetched from the picture memory and then corrected in the correct position by the motion vector section to decode the motion compensation prediction by the adder 149 . When the data are B pictures, the same operation is performed with respect to I pictures and P pictures.

In Fig. 80, motion vectors, quantization parameters of inverse quantization, and predictive patterns are output from the variable length decoder. Since the information flow is the same as that of Fig. 74, no explanation is given. A loop in the lower part of Fig. 75 is the loop for low-resolution component decoding. The decoded result is subjected to pixel interpolation by the resolution inverse converter 152 to compensate the decoded result as a residual difference in resolution, and the data is input to the image memory 150 .

Example 20:

Embodiment 20 of the present invention is explained with reference to FIGS. 81 , 82 and 83 . 81 and 82 are block diagrams of the decoding process. In FIGS. 81 and 82, reference numeral 162 denotes a video signal encoder as encoding means. 163 is an audio signal encoder, 164 and 167 are memories, 165 and 168 are memory controllers. The memory 164 and the memory controller 168 constitute data providing means. Furthermore, the video signal encoder 162 and the memory controller 165 constitute a code amount comparing means. Reference numeral 166 denotes a system layer bit stream generator. 169 represents a variable length coder, and 170 represents a signal processor for decoding after the variable length coder. Encoders 169 and 170 function as data decoding means. The data rearranger 131 of Fig. 82 is used as data reconstruction means.

First, the structure shown in Fig. 81 is explained. A memory 164 is disposed between the video signal encoder 162 and the system bit stream generator 166 . After the data is embedded between each GOP of the encoded video signal, each GOP is input to the system layer bit stream generator 166, and the audio signal encoded by the audio signal encoder 163 is subsequently combined with the video for adding headers, etc. The signals are input into the system bit stream generator 166 together.

Here, the data embedding operation in the memory 164 is described. The memory controller 165 functions as a control circuit of the memory 164 to control a video signal for controlling video signal data encoding so as to bury a space between each GOP. Signal processing is explained below with reference to FIG. 83 . Fig. 83 is a diagram showing the concept of processing on a digital video signal recording and reproducing apparatus. For example, when nGOP generates redundant data with respect to the access address of the optical head and the control unit of error control, (n+1) GOP is cut off halfway with respect to the access position of the optical head or the control unit of error correction to generate the space of the data area In this case, the device of the present invention makes nGOP redundant as shown in FIG. 83A and embeds (n+1) GOP in the space after the data, and in the same way makes (n+1) which cannot be embedded due to nGOP embed. 1) A small amount of remaining data of GOPs and (n+2) GOPs are embedded in the space portion of (n+3) GOPs (the embedding direction is from left to right as viewed on paper).

In addition, as another control method, redundant data cannot be transmitted in the reverse direction as described above. As shown in FIG. 83B, when the (n+2) GOP exceeds the access position a little, the (n+3) GOP ends halfway with respect to the access position of the optical head and the control unit of the wrong control, and the device of the present invention is controlled, In this way, the redundant (n+3) GOP data is buried in the space part after the (n+2) GOP data, in this way, the (n+2) GOP which cannot be buried because the (n+3) GOP is buried The remaining data of the GOP is embedded in the space part of the (n+1)GOP, and the part of the (n+1)GOP that cannot be embedded is buried in the space part of the nGOP (the embedding direction is from right to left on paper) .

The operation of the structure shown in Fig. 82 is explained below. The memory 167 is controlled by the memory controller 168, and the data rearranged according to the rules described above in FIGS. 83A and 83B is restored to its original state. For example, after the data shown in FIG. 83A is restored, the device of the present invention works to restore the GOP data to the original state, and the nGOP part with (n+1) GOP is connected to the left part of (n+1) GOP, and then, The (n+1) GOP data is linked to the (n+2) GOP data.

This rearrangement rule will be defined along with the formatting rules of the medium, so that the rule is recorded as marking information following a rearranged area such as the TOC area, where the format is not defined, the rule Must be clearly described somewhere on the medium.

Example 21

Embodiment 21 is explained below with reference to FIGS. 84 , 85 and 86 . Fig. 84 is a block diagram representing a signal processing unit in which division by frequency or division by quantization is performed in the digital signal reproduction section. Fig. 85 is a block diagram representing a signal processing unit in which division by bit length or division by quantization is performed in the digital signal reproduction section. Fig. 86 is a block diagram representing a signal processing unit in which division by resolution or division by quantization is performed in the digital signal reproduction section. In FIGS. 84, 85 and 86, reference numeral 171 denotes an IP selection indicator, 172 a decider, and 173 a switcher. In FIG. 86, corresponding parts are shown as first decoding means, second decoding means and third decoding means. In Figs. 84, 85 and 86, similar parts are denoted by the same reference numerals, and descriptions thereof are omitted.

The operation of Embodiment 21 is explained below. In FIGS. 84 and 85, the optical head or the rotation of the optical head is controlled so that data is read directly or indirectly from the TOC area or the area corresponding to the TOC area. The playback signal from the optical head is amplified by playback amplifier 156 . This signal is retrieved and differentiated by digital demodulator 157 into a digital signal for digital demodulation. As a result, the reproduced signal, which has been digital data, is input to the error corrector 158 to correct errors contained in the reproduced signal. Error-free data is separated into audio bitstream and video bitstream by system layer processor 159, as are other items of data. In FIGS. 84, 85 and 86, a control signal is input into the IP selection indicator 171 from the mode switcher 130 as the mode switching means to switch the decoding means on the basis of the individual playback speed. Embodiment 21 is controlled so that Embodiment 21 can switch between a mode of displaying only I pictures or a mode of displaying I pictures and P pictures using the control signal and the skip scan speed.

When the skip search speed is 100 times, if the I picture and the P picture are output on the screen, the GOP must be output with a considerable amount of thinning. As a result, the image appears rather unnatural relative to the movement of the playback screen. In this case, in order to eliminate this unnaturalness, it is necessary to switch the mode to a mode for reproducing only I pictures. The decodable decider 132 (172 in Fig. 86) is used to decode not only the B picture but also the P picture (the switch 173 is used for this function in Fig. 86). Meanwhile, the picture memory 137 (150 and 151 in Fig. 86) is controlled to display only I pictures.

Screen displays of I pictures and P pictures generally favor speeds at 15x speed, but screen displays of only I pictures favor speeds of 15x and above. This is because when the entire I picture and the entire P picture are displayed at 15x speed, since the replayable GOP is located at the fifth GOP position from the currently displayed GOP for each frame update of the screen in subsequent processing, consecutive Extremely poor movement. Furthermore, when the number of frames N=15 in a GOP and the period of I pictures and P pictures is 3, all P pictures are decoded and only the second frame of I pictures and P pictures (the third frame in the GOP) and the ninth frame) are output, and a finer skip search can be performed.

As described above, the data state is divided and recorded by dividing the data state on the basis of a predetermined condition in the case of recording data at a predetermined position every GOP read from the recording medium by dividing the frequency area. , for the case of recording data, the case of dividing data by resolution so that data is recorded, and the case of dividing data by the quantization level to be recorded. When the first data which is the basic data in the playback data and the second data other than the basic first data are reproduced from the data arranged collectively, a decoding device is used to obtain one of the playback images from the following cases : A case where all the first and second data are decoded, and a case where the obtained data is corresponding to the low frequency area of the I picture and the P picture or the number of thinned pixels. Subsequently, the decoding means used at the time of the exclusive playback can be switched on the basis of the individual playback speed.

Needless to say, the manner of displaying I pictures and P pictures can be changed to positive and negative directions during playback. Since the P direction can be decoded only in the forward direction, the screen existing before the decoding at the time of reverse playback of the P picture is stored. As a result, redundant memory needs to be used on that part. In order to facilitate reverse playback without using such redundant memory, I pictures and P pictures can be played back during forward skip search, and only I pictures can be played back during reverse skip search.

Example 22:

Embodiment 22 is explained with reference to FIGS. 87 , 88 , 89 and 90 . Fig. 87 is a block diagram representing a signal processing unit in which division by frequency or division by quantization is performed in the digital signal reproduction section. Fig. 88 is a block diagram representing a signal processing unit in which division by bit length or division by quantization is performed in the digital signal reproduction section. Fig. 89 is a diagram for explaining the concept of processing at the time of skip search. In Figs. 87 and 88, reference numeral 174 denotes a field display controller. The system layer processor 159 functions as video data extraction means. In addition, Figs. 87 and 88 show examples of video data decoding and reproducing apparatuses of corresponding parts. Reference numeral 130 denotes a mode switcher as mode switching means. The same or similar parts are denoted by the same or similar symbols as in the previous embodiments.

The operation of Embodiment 22 is explained below. In FIGS. 87 and 88, the optical head or the rotation of the optical head is controlled so that data is read directly or indirectly from the TOC area or the area corresponding to the TOC area. The playback signal from the optical head is amplified by playback amplifier 156 . This signal is retrieved and differentiated by digital demodulator 157 into a digital signal for digital demodulation. As a result, the reproduced signal, which has been digital data, is input to the error corrector 158 to correct errors contained in the reproduced signal. Error-free data is separated into audio bitstream and video bitstream by system layer processor 159, as are other items of data.

When skipping searches, such as I pictures and P pictures are continuously output on the screen, and the screens are output in the order represented by the arrows in FIG. 89 . At this time, at the time of the skip search, the even fields of the I picture and the odd fields of the P picture are continuous, and there are four empty fields between them in the coded data. In other words, the playback speed in the coded data is five times the playback speed in the space between the odd and even fields of the I picture. Therefore, since the playback speed changes from 1x speed to 5x speed for one field, the motion of the I picture changes unnaturally.

This can be replaced by odd or even fields of I pictures on the same screen. Otherwise, the screen can be prepared by embedding the average value of the upper and lower scan lines to allow for the intersection of the outputs. The picture memory 137 shown in Figs. 87 and 88 is controlled by the field display controller 174 to form the same screen in subsequent P pictures. As a result, the amount of jumping between fields that is the space between fields encoded when recorded on each field of the playback image can be uniformly obtained as the result from jumping (unnatural motion) becomes inconspicuous.

In addition, FIG. 90 is a diagram for explaining the concept of processing at the time of reverse playback, in which the order of fields at the time of reverse playback is particularly shown. Next, the field order at the time of reverse playback will be explained on the basis of FIG. 90. FIG. In reverse playback, reverse playback is performed in frame units constituting a pair of odd and even fields. Specifically, when the operation is changed from an odd field to an even field, playback is performed in the same direction as on the picture (the playback operation procedure a is described in Fig. 90). When the operation changes from an even field to an odd field, the two field portions are sent in reverse in the direction opposite to that when on the picture (Fig. 90A shows process b in skip).

However, when the aforementioned playback process is performed, the result of the triple speed playback makes the playback appear to be moving in an unsuitable manner, so that the motion sequence does not feel smooth. After the image memory 137 is controlled by the field display controller 174 of FIGS. 87 and 88, the playback image is displayed on a screen in the opposite direction to the order of odd, even, odd, and even shown in FIG. 90B. Since the amount of jumping between fields can be obtained almost uniformly, the jumping becomes inconspicuous. However, with respect to the synchronization signal at this time, it is necessary to maintain the normal odd-numbered and even-numbered field relationship without inverting the field synchronization signal.

The image memory 137 does not receive the output of the adder 138 because it uses a display method different from each field, but the output of the adder 138 can be received from the image memory 137 alone. To perform this operation, the order is changed by providing buffers as separate means. It is possible to use a memory with 3 ports in which address control can be set up individually. The aforementioned operations can be realized even when data is multiplexed with a memory from which data can be read out at a very fast speed. Furthermore, since the reverse scan converter 139 provides at least one memory at least one field plus one splice, it goes without saying that this buffering function can be introduced into the reverse scan converter 139 .

Also, with regard to the slow playback output for dedicated playback, when the same frame is repeatedly output, skipping becomes inconspicuous. As a result, frames are reconstructed and output, and this playback interval is the same. For example, during slow playback at 1/3x speed, decoded B frames are not tripled after output of decoded I frames. Instead, the first frame is made up of the odd fields of the I-frame. In even fields, the average value of the upper and lower lines can be used.

In this constitution, since no image shift occurs in the vertical direction of the screen due to the line shift of the image at the boundary, a stable image can be obtained. The next frame outputs the original I frame, and the next frame forms a frame with the odd field of the I frame (at the end of the even field, the average value of the uplink and downlink can be taken). Subsequently, the following frame outputs the original B frame, and the next frame forms a frame with the even field of the B frame (at the end of the odd field, the average value of the uplink and downlink can be taken). As a result, slow playback can be achieved at equal intervals within this time interval.

Claims (4)

1. a compact disk recording method is used for the digital video signal of record by utilizing motion compensated prediction and DCT to encode on CD, and described method comprises the following steps:

System flow is arranged, this system flow comprises the video information packets that formed by block of video data and has the specific information bag of the stream ID of dedicated stream, and each described block of video data comprises by the image as the I image of intraframe coding image, the video data that constitutes as the P image of single directional prediction coding image with as the B image of bi-directional predictive coding image;

The specific information bag is arranged, so that follow in described specific information bag back corresponding to the block of video data of described specific information bag; And

Recording address information in described specific information bag, this address information with before corresponding to the block of video data of described specific information bag and the sector of block of video data afterwards corresponding.

2. an optical disc recording apparatus is used for the digital video signal of record by utilizing motion compensated prediction and DCT to encode on CD, and described device comprises:

Be used to arrange the equipment of system flow of the specific information bag of the stream ID that comprises the video information packets that formed by block of video data and have dedicated stream, each described block of video data comprises by the image as the I image of intraframe coding image, the video data that constitutes as the P image of single directional prediction coding image with as the B image of bi-directional predictive coding image; With

Be used for the specific information bag is arranged, so that follow in described specific information bag back corresponding to the described block of video data of described specific information bag, and in described specific information bag the equipment of recording address information, this address information with before corresponding to the block of video data of described specific information bag and the sector of block of video data afterwards corresponding.

3. an optical disc reproducing method is used to play the CD that comprises by the digital video signal that utilizes motion compensated prediction and DCT coding, wherein:

A kind of system flow of specific information bag that comprises the video information packets that formed by block of video data and have a stream ID of dedicated stream is arranged in the described CD, and each described block of video data comprises by the image as the I image of intraframe coding image, the video data that constitutes as the P image of single directional prediction coding image with as the B image of bi-directional predictive coding image;

Arrangement is corresponding to the specific information bag of block of video data, so that follow in described specific information bag back corresponding to the block of video data of described specific information bag;

Record is with before corresponding to the block of video data of described specific information bag and the corresponding address information in sector of block of video data afterwards in described specific information bag; And

Described playback method comprises the steps: when resetting described video data, reference record in described specific information bag corresponding to before described block of video data and the address information of the sector of described block of video data afterwards.

4. a disc reproducing apparatus is used to play the CD that comprises by the digital video signal that utilizes motion compensated prediction and DCT coding, wherein:

A kind of system flow of specific information bag that comprises the video information packets that formed by block of video data and have a stream ID of dedicated stream is arranged in the described CD, and each described block of video data comprises by the image as the I image of intraframe coding image, the video data that constitutes as the P image of single directional prediction coding image with as the B image of bi-directional predictive coding image;

Arrangement is corresponding to the specific information bag of block of video data, so that follow in described specific information bag back corresponding to the block of video data of described specific information bag;

Record is with before corresponding to the block of video data of described specific information bag and the corresponding address information in sector of block of video data afterwards in described specific information bag; And

Described replay device comprises: be used for obtaining being recorded in described specific information bag corresponding to before described block of video data and the address information of the sector of described block of video data afterwards, with the address information that is obtained by reference, the equipment of the described video data of resetting.

Images