JPH08111847A - Video recording / playback method - Google Patents

Abstract

(57) [Abstract] [Purpose] To realize an efficient file method in an optical disc device for compressing and recording digital video. [Structure] When a digital image is compressed, first temporary coding is performed, and the code amount for one disk is stored in units of 1 GOP. At this time, the motion speed and the motion correlation amount for each GOP picture are simultaneously stored. Next, the optimum value of the code amount and the optimum value of the number of frames are calculated for each GOP from the code amount, the motion vector amount, and the motion vector correlation amount obtained in the first time. Finally, the second encoding is performed, the code amount is controlled by feeding back to the adaptive quantizer 21, and the disk format encoder 5
At B, the B picture is deleted to optimize the number of frames.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a high density recording / reproducing method for an optical disc.

[0002]

2. Description of the Related Art FIG. 15 is a block diagram of a conventional optical disc recording / reproducing apparatus disclosed in Japanese Patent Laid-Open No. 4-114369. In the figure, 1 is a video A / D converter for converting a video signal into digital information, 2 is a video information compression means, and 35 is a frame sector conversion for converting the compressed video information into sector information equal to an integral multiple of a frame period. Means, 36 is an encoder, 37 is a modulator for converting into a predetermined modulation code to reduce inter-code interference in the recording medium, 38 is a laser drive circuit for modulating a laser according to the modulation code, 39 is a laser It is an output switch.

Reference numeral 40 is an optical disk, 41 is an optical head for emitting a laser beam, 42 is an actuator for tracking a light beam emitted from the optical head 41, 43 is a traverse motor for sending the optical head 41, and 44 is for rotating the disk 40. Disk motor, 47 is a motor drive circuit, 4
Reference numerals 5 and 46 are motor control circuits.

Reference numeral 48 is a reproduction amplifier for amplifying the reproduction signal from the optical head 43, 49 is a demodulator for obtaining data from the recorded modulated signal, 50 is a decoder, 51 is a frame sector inverse conversion means, and 52 is the compression described above. Information decompression means for decompressing information, and 53 is a D / A converter for converting the decompressed information into an analog video signal.

FIG. 16 is an MPEG standardized for compressing digital moving image information and transmitting / storing it.
In the figure, the data array structure (layer structure) of the method is simplified and shown. 54 is a GOP consisting of a plurality of frame information,
5 is a GOP composed of several pictures (screens)
A layer, 56 is a slice obtained by dividing one screen into several blocks, 57 is a slice layer composed of several macro blocks (MB), and 58 is a block layer composed of 8 pixels × 8 pixels.

FIG. 17 is a diagram showing a coding structure when 10 screens are set to 1 GOP. 61 is an I picture which is video information for performing intra-frame DCT, and 63 is DCT coding for performing forward motion compensation. P picture, which is the video information of the I picture 61
And the P picture 63 is a B picture on which DCT coding is performed with motion compensation using the P picture 63 as a reference screen.

FIGS. 18A and 18B show a case where the video data amount in one GOP has a variable structure in order to keep the image quality between GOPs constant, and a fixed rate in order to keep the recording time constant. It is the figure which compared what was done. In the figure,
64 is the innermost circumference of the disc, and 65 is the outermost circumference of the disc.

Further, FIG. 19 (a) is a diagram showing the data amount per 1 GOP when the image quality per 1 GOP is kept the same. Α represents the maximum value of the data rate, and β represents the average data rate. Further, FIG. 19B shows each image (e),
It is a figure which compared the image quality and data amount per 1 GOP in (d) and (c).

Next, the operation of the conventional example will be described. By recording the above compressed information on an optical disk as the compression technology of digital video information advances, an extremely easy-to-use video filing device with better searchability than a tape medium such as a conventional VTR is realized. It is possible to do. In addition, since such a disk file device handles digital information, there is no dubbing deterioration as compared with the case of recording an analog video signal, and since it is optical recording / reproducing, a contactless and highly reliable system is realized. it can.

Conventionally, when recording such compressed moving picture information on an optical disc, a method of recording digital compressed moving picture information such as the MPEG system shown in FIG. 16 on the optical disc 40 shown in the block circuit diagram of FIG. Is taken.
At this time, the video information digitized by the video A / D converter 1 is processed by the video information compression means 2 into, for example, MPE.
It is converted by a standard compressed moving image method such as G. The compressed video information is encoded and modulated to reduce the influence of intersymbol interference on the optical disc and recorded on the optical disc 40. At this time, it is apparent that, for example, the data amount in each GOP unit is set to be approximately the same amount, and the data is divided into sectors equal to an integral multiple of the frame period so that the GOP unit can be edited.

During reproduction, the optical disk 40
The video information recorded in (1) is reproduced by the optical head 41, amplified by the reproducing amplifier 48, restored to digital data by the demodulator 49 and the decoder 50, and thereafter, the frame sector inverse conversion means 51 reproduces the address, parity, etc. It is restored as pure video original data with the data removed. Further, the information decompression means 52 reproduces a video signal by performing MPEG decoding, for example, and the D / A converter 53 converts it into an analog video signal for display on a monitor or the like.

As described above, when the MPEG method is used as the digital moving picture compression method, as shown in FIG. 17, an I picture 61 for compression by intra-frame DCT and a DCT code for forward motion compensation are provided. The P picture 63, which is the video information by the encoding, and the B picture 62, which is DCT-encoded with motion compensation using the I picture 61 and the P picture 63 located temporally before and after as a reference screen, are combined. The coding structure is recorded in the optical disc 40 as it is.

Among these pieces of information, the I picture 61 performs the intra-frame DCT, and therefore it is possible to reproduce the image by this information alone, but the P picture 63 performs the forward motion compensation. Therefore, the image can be reproduced only after the I picture 61 is reproduced.
Since the B picture 62 is a prediction screen from both directions,
It cannot be played back until after the I picture 61 or the P picture 63 located before and after it is played back. Of these pieces of information, the B picture 62 for which bidirectional prediction is naturally performed has the smallest data amount and the coding efficiency is good.

However, since the B picture 62 cannot be reproduced independently, the I picture 61 and the P picture 63 are required. However, if the number of B pictures 62 is increased by that much, the buffer memory amount in the processing circuit increases and There is a problem that the delay time from data input to video reproduction increases. However, in storage media typified by optical disks and the like, an encoding method with good compression efficiency is desired for long-term recording, while the delay time of video reproduction does not pose a problem. A coding method as shown in 17 is suitable.

Next, if video data is recorded so that the image quality is constant in any part of one optical disk,
The variable rate structure is as shown in FIG. This is shown in FIG. 19 when the image quality per GOP is constant.
This is because the amount of video data required for 1 GOP fluctuates as shown in (a). This is because, for example, when the amount of data required for an I picture increases in the case of a fine image, or when video data that moves rapidly is continuous, the compression efficiency of a P picture or a B picture does not increase so much. . Further, as is apparent, as shown in FIG. 19B, when the data amount per 1 GOP is increased, the S / N ratio of the image is improved although it varies depending on the pattern.

On the other hand, in order to keep the recording time for one optical disk constant, the format for recording at a fixed rate shown in FIG. 18B is suitable. However, unlike a magnetic tape medium, in the case of a video recording / reproducing apparatus using an optical disc medium, the total amount of data per package is small, and therefore the compression efficiency must be increased as much as possible while maintaining high image quality. To that end, it goes without saying that the variable rate system shown in FIG. 18A has a better file efficiency of video data per optical disc.

Therefore, for example, in a read-only optical disk device, by encoding in advance,
Since it is possible to know the data amount distribution of the entire optical disc at the variable rate, the entire data distribution is adjusted at the time of the second encoding, and as a result, the reproduction time per disc is also changed at the variable rate. It is possible to adjust it to a constant level.

[0018]

As described above, in the conventional optical disc video recording system, the file efficiency of the video data recorded on one optical disc medium is improved by making the data rate per GOP variable. Although improved, a file method having a higher compression efficiency than before has been desired for an optical disk medium having a much smaller total data amount than a magnetic tape medium.

Further, not only a read-only optical disk medium but also a recordable optical disk medium, and a file having a higher compression efficiency than a normal compression method even when the remaining recording area runs out during recording. The method was desired.

The present invention has been made to solve the above problems, and an object thereof is to obtain a video recording / reproducing method with high compression efficiency.

[0021]

According to the invention of claim 1, at the time of recording, the digital video information is compressed on the basis of the motion estimation information between the frames temporally preceding and succeeding, and the frames included in the compressed video information. The amount of motion vector data that increases or decreases according to the motion of each image (hereinafter,
If the "motion vector amount" is smaller than a predetermined absolute amount, the number of frames per second is reduced and recorded on the recording medium, and at the time of reproduction, the frames before and after the reduced frame are displayed in duplicate. In this way, a predetermined required number of frames is secured.

According to a second aspect of the present invention, at the time of recording, the digital video information is compressed based on the motion estimation information between the frames temporally preceding and following, and the motion of the video for each frame included in the compressed video information is calculated. If the amount of motion vector that increases or decreases in accordance with the amount is smaller than the range between the first absolute amount and the second absolute amount that are set in advance, the number of frames per second is reduced in proportion to the amount of the above motion vector. If it is larger than the above range, the number of frames per second is reduced so as to be inversely proportional to the above-mentioned dynamic vector amount, and then recorded on the recording medium.
The frames before and after the reduced frame are overlapped and displayed to ensure the required number of frames.

According to the invention of claim 3, in the invention of claim 1 or 2, the video information recorded on the recording medium is:
An I picture, which is video information for performing intra-frame DCT,
A P picture, which is video information by DCT coding for motion compensation in the forward direction, and a B picture for which DCT coding has been performed with motion compensation using the I and P pictures that are temporally preceding and following as a reference screen. Is a series of video information blocks in units of several frames to several tens of frames, and the motion vector amount is stored in the video information block unit, and the B picture of the image data is deleted according to the motion vector amount. It is the one.

According to a fourth aspect of the present invention, at the time of recording, the digital video information is compressed based on the motion estimation information between the frames temporally preceding and following, and the motion of the video for each frame included in the compressed video information is recorded. 1 if the total amount of motion vector amount that fluctuates depending on the magnitude is smaller than the first absolute amount that is determined in advance and the amount of change in the motion vector amount for each screen is greater than the second absolute amount that is determined in advance. The number of frames per second is reduced and recorded on a recording medium, and at the time of reproduction, the frames before and after the reduced frame are overlapped and displayed to secure the required number of frames.

According to a fifth aspect of the invention, in the fourth aspect of the invention, the video information recorded on the recording medium is an intra-frame DC signal.
Motion compensation is performed using an I picture, which is video information for T, a P picture, which is video information by DCT encoding for motion compensation in the forward direction, and the I and P pictures that are temporally preceding and following, as reference screens. The number of consecutive video information blocks in units of several tens of frames from the number of mixed B pictures for which DCT encoding has been performed is performed. And detecting the motion vector correlation amount in the video information block by comparing each P picture of
The conversion table for determining the number of dropped frames of a picture determines the amount of dropped frames in the video information block.

According to a sixth aspect of the present invention, at the time of recording, the code amount after encoding and the motion vector amount of the video information block are stored for each unit of the video information block composed of a plurality of pictures at the time of the first encoding. From the stored code amount and motion vector amount, the optimum code amount and the number of frames for each video information block unit for recording on the recording medium for a predetermined time are calculated, and at the time of the second encoding, the optimum The B-picture data is deleted by the number of frames described above while controlling so that the number of codes becomes large.

According to a seventh aspect of the present invention, in the sixth aspect of the invention, the video information recorded on the recording medium is DC in the frame.
Motion compensation is performed using an I picture, which is video information for T, a P picture, which is video information by DCT encoding for motion compensation in the forward direction, and the I and P pictures that are temporally preceding and following, as reference screens. A series of video information blocks in units of several frames to several tens of frames in which a B picture for which DCT encoding has been performed is mixed is stored, and the total amount of motion vectors is stored in units of the video information blocks. Based on the above motion vector correlation amount,
The conversion table for determining the number of dropped frames of the B picture determines the amount of dropped frames in the video information block.

[0028]

According to the first aspect of the present invention, when the motion vector amount that increases or decreases according to the motion of the image in each frame is smaller than a predetermined absolute amount, the number of frames per second is reduced and recorded. At the same time, during reproduction, the frames before and after the reduced frame are displayed in an overlapping manner to secure the number of frames required for display.

According to the second aspect of the invention, when the motion vector amount that increases or decreases according to the motion of the image for each frame is smaller than the predetermined range between the two absolute amounts, it is proportional to the motion vector amount. The number of frames per second is reduced, and when the number is larger than the above predetermined range, the number of frames per second is reduced and recorded in inverse proportion to the above motion vector amount, and at the time of reproduction, it is reduced. The number of frames required for display is secured by displaying the frames before and after the frame in duplicate.

According to the third aspect of the present invention, the motion vector amount in the video information block unit is stored, and the B picture of the image data is deleted according to the motion vector amount.

According to the fourth aspect of the present invention, the total amount of motion vector included in the compressed video information and fluctuating according to the magnitude of the motion of the video for each frame is smaller than the predetermined first absolute amount, and When the amount of change in the motion vector amount for each screen is larger than the second predetermined absolute amount, the number of frames per second is reduced and the frames before and after the reduced frame are displayed in duplicate during playback. Secure the number of frames required for display with.

According to the fifth aspect of the present invention, the total amount of motion vectors in video information block units is stored, and the motion vector amount is compared for each preceding and following P picture to calculate the motion vector correlation amount in the video information block. Based on the total motion vector amount and the motion vector correlation amount, the conversion table determines the number of dropped frames of the B picture to determine the amount of dropped frames in the video information block.

According to the invention of claim 6, the code amount after encoding is stored for each unit of the video information block consisting of a plurality of pictures in the first encoding, and the motion vector amount of the video information block is stored. Remember
From the stored code amount and motion vector amount, the optimum code amount and the number of frames for recording on a single optical disc for a predetermined time are calculated in video information block units, and the optimum code is determined at the time of the second encoding. The quantizing means is controlled so that the amount becomes the amount, and the B picture data is deleted by a predetermined number of frames.

According to the invention of claim 7, the total amount of motion vectors in video information block units is stored, and the total amount of motion vectors is compared for each preceding and following P picture, and the motion vector in the video information block is compared. Detect the correlation amount, based on the total amount of the motion vector amount and the motion vector correlation amount,
The number of dropped frames of the B picture is determined by the conversion table to determine the amount of dropped frames in the video information block.

[0035]

【Example】

Example 1. FIG. 1 is a block circuit diagram of a first embodiment of the present invention. By creating a recorded video file in which the number of frames per second of digital video information is reduced when creating an optical disc on which digital moving picture information is recorded. , Is designed to improve the compression efficiency of digital video. In the figure, 1 is a video A / D converter, 2 is video information compression means, 3 is motion vector amount detection means for detecting a motion vector amount, 4 is a frame drop amount determination means, 5 is a disk format encoder, and 6 is modulation. Means, 7 is a recording data file, 8 is a ROM disc mastering device, and 9 is a creation R
It is an OM disc.

FIG. 2 is a diagram showing how frames are dropped in accordance with the motion vector amount in the first embodiment. FIG. 2 (a) shows a case of moving 24 frames / second and 30 frames / second. FIG. 2B shows a case where switching is performed according to the vector amount, and a case where 30 frames / second, 27 frames / second, and 24 frames / second are switched, respectively.

The number of frames per second on the TV screen is N
Although it varies depending on the TSC area or the PAL area, in the case of Japan or the United States, it is 30 frames / second. The number of frames displayed on the TV screen must correspond to the format of the TV system, but the video data recorded on the optical disk does not necessarily have to be a file of all the frames, and even if some frames are dropped. It is possible to delete data in picture units in a blind area. In this case, when the screen is displayed, a flag for repeatedly reproducing the same screen of the previous time is set in the header part provided for each GOP composed of a plurality of pictures or the header part provided at the beginning part of the picture data. It is possible to deal with this.

However, in the case of a motion picture film whose data is originally composed of 24 frames / sec.
Keeping the number of pictures reduced in GOP units constant may be noticeable depending on the pattern when frames are dropped. Therefore, in the first embodiment shown in FIG. 1, the number of dropped frames is adaptively changed according to the speed of movement of the screen.

Next, the operation of the first embodiment will be described. The motion vector amount of the video report compressed by the video information compression means 2 is
It is extracted by the motion vector amount detecting means. In general, a motion vector code is assigned a smaller number of bits for less motion and a larger number of bits for greater motion. Therefore, just by counting the amount of motion vector, the speed of motion for each screen can be calculated. Can be grasped quantitatively.

Depending on the design, even if most of the screen is close to a still image, a part of the screen may move significantly. In such a case, it is not the average level of the entire picture, but the macroblock (MB). It is more suitable to extract the maximum value of the motion vector data in units to obtain the motion vector amount.

Therefore, the motion vector amount detection circuit 3 sets 1
It is possible to count the motion vector amount per GOP, and determine the number of dropped frames per GOP by the dropped frame amount determination means 4 depending on whether or not the counted value exceeds or exceeds a predetermined value. Further, in this case, among the compressed video data once stored in the memory by the disk format encoder 5,
The operation of deleting the data of the B picture and rewriting the header information allocated in units of 1 GOP or the header information allocated in units of 1 picture is performed.

This is because if the I picture or P picture is deleted, the B pictures before and after cannot be decoded, and the information of the deleted picture is written in the header to freeze the screens before and after the reproduction. This is because it is possible to match the number of frames per second with the required number of TV systems.

In the case of the first embodiment, the amount of data to be recorded on the optical disk is reduced by reducing the frame drop when the motion vector amount is large, and by increasing the frame drop when the motion vector amount is small near the still image. Can be reduced. In such a system, the amount of dropped frames is variable according to the movement of the screen, so that even if dropped frames are not noticeable to the human eye. The detection of the motion vector amount in this case is 1G.
It is desirable to start from P pictures that are evenly allocated in OP, and it is possible to start from B pictures, but the continuity of compressed images and the presence of bidirectional data may complicate the system.

In addition, as shown in FIG. 2A, the frame drop amount determining means 4 sets the number of frame drops to zero according to the magnitude of the motion vector amount, and 6 frames per second (24 frames / second for pictures). 2), it is possible to set it as shown in FIG.
As shown in (b), it is possible to set from 0 to 6 frames per second in multiple stages. In the case of film movies, etc., since it is 24 frames / second, MPE
In the standard such as G, the reproduction system from 24 frames / sec. Therefore, in particular, if the dropped frames are specified in the two stages of 24 frames / second and 30 frames / second described above, the system configuration becomes simple, which is extremely practical.

Example 2. In the first embodiment, the number of dropped frames is determined stepwise in inverse proportion to the motion between pictures. However, considering the characteristics of the actual human eye, the range in which the motion is too fast and the human eye can follow is determined. If it exceeds, it may not be noticeable even if the number of dropped frames is increased. this is,
This is because it is expected that human eyes have the detection limit characteristic of frame drop as shown in FIG. In this case, it is naturally difficult to detect dropped frames in a video close to a still image. On the other hand, even when the motion of the image is so intense that it is difficult for the human eye to follow the frame drop, it is naturally difficult to detect the dropped frame. The second embodiment of the present invention utilizes such a characteristic to correct the extracted value of the motion vector amount.

FIG. 3 is a block circuit diagram of the second embodiment. The same reference numerals as those in FIG. 1 indicate the same parts, and 10 is a data conversion table. In the second embodiment, when the motion vector amount is equal to or less than a predetermined absolute amount, the number of frames dropped is in inverse proportion to the motion vector amount, and when the motion vector amount is equal to or more than the absolute amount, the number of frames proportional to the motion vector amount is increased. It is designed to drop frames.

FIG. 4 is a diagram showing an arrangement of digital compressed video data recorded on the disc before frame dropping and in the second embodiment, and FIG. 5 shows video information data recorded on the optical disc with a reduced number of frames. FIG. 5A is a diagram showing what the display on the screen looks like at the time of reproduction. FIG. 5A is a display screen of video data in which no frames are dropped, and FIG.
Display screen of video data with one screen dropped,
FIG. 5C shows a display screen of video data in which two screens are dropped on three screens, and FIG. 5D shows a display screen of video data in which one screen is dropped on five screens.

FIG. 6 is a diagram showing the human visual characteristics, and shows how many dropped frames are detected with respect to the speed of image movement. Further, FIG. 7 shows the second embodiment.
FIG. 7 is a diagram showing an example of how much frame dropping is performed in accordance with the motion vector amount.

In the second embodiment, the number of dropped frames is determined in inverse proportion to the motion vector amount until the motion vector amount reaches a predetermined absolute amount, and when the motion vector amount is equal to or more than the above absolute amount, it is proportional to the motion vector amount. It is possible to drop frames. Other operations are the same as those in the first embodiment, and the B picture is deleted and the header information is rewritten to complete the frame dropping processing of the compressed video data.

As described above, according to the method of the second embodiment, it is possible to drop frames even if the human eye cannot catch up with the motion of the method of the first embodiment. The compression efficiency of the video data is higher.

Further, it is also possible to realize the characteristic of the frame drop amount judging means 4 as shown in FIG. 7 without using the conversion table 10 described above. In this case, the allowable value of the number of dropped frames with respect to the movement of the screen has a non-linear characteristic. Even if the motion is too intense and the motion vector amount cannot be detected, a certain amount of frame drop is allowed. Also, when the image is close to a still image, the number of dropped frames can be increased.

The video information actually recorded on the optical disk has a form as shown in FIG. 4 (c). Original digital compressed video data has a data array as shown in FIG. This is because although an I picture can be reproduced independently, a P picture needs a prediction screen from a temporally previous I picture or a P picture, and a B picture from a preceding or subsequent I picture or P picture. This is because a prediction screen is needed. Therefore, unlike the playback order of the screen, the data of the P picture is arranged next to the I picture,
Next, B picture data is arranged.

However, regarding the data arrangement on the optical disk, it is very convenient that the I picture and the P picture are arranged consecutively in consideration of the case where only the I picture and the P picture are reproduced during special reproduction. Well, Figure 4
It is rearranged as shown in (c). This is because the amount of data after compression is sufficiently smaller than the amount of original video signal data, so the memory of the optical disk reproducing device can be easily rearranged and sufficient for the above-mentioned data array conversion. This is because it can handle In the case of this method, the B picture is partially deleted in the case of a GOP in which video data with further rapid movement is continuous, so that the data amount is reduced and the file efficiency is improved as shown in FIG. 4C. ing.

The rearrangement of data and the reduction of picture data are performed by the disc format encoder 5 shown in FIGS. 1 and 3, for example, but removal of intersymbol interference at the time of further high density recording, etc. For the purpose of
For example, modulation such as EFM modulation or 1-7 modulation is applied,
It is once recorded on a recordable file device (for example, a magnetic disk, a magnetic tape, a magneto-optical disk, or the like). Thus, for example, a master is created by the ROM disk mastering device 8 from the data once stored,
The ROM disks 9 are mass-produced by the stamper.
As a matter of course, when the optical disk is a recording / reproducing apparatus, the above operation is performed without the recording data file 7 and the ROM disk mastering apparatus, and the data is directly recorded on the optical disk.

Next, when the above-mentioned compressed video data is read from the optical disk, reproduced, and displayed on the screen, FIG.
It becomes as shown in. FIG. 5A shows a video in which there is some motion between pictures that does not allow frame dropping, and corresponds to the video before encoding in units of picture. On the other hand, if one frame is allowed to be dropped from three screens, the result is as shown in FIG. 5B, in which the B2, B4, and B6 pictures are dropped and the B1, B3, and B5 pictures are frozen ( Repeatedly display). Also,
Further, when dropping of two screens is allowed in three screens, a screen without B picture is reproduced. In the case of FIG. 5C, the state of dropped frames is easily detected by human eyes. However, when the frame is extremely close to a still image, it is not noticeable to allow dropped frames to this extent.

In addition, as shown in FIG.
One frame may be allowed to be dropped for every 5 screens in the middle of (a) and FIG. 5 (b). Even in this case, the B picture is dropped and it is possible to freeze the previous and next pictures. Is. In the case of FIG. 5 (d), it is desirable to fix the freeze position on the five screens in order to prevent the eyes from being noticeable to the human eye. In order to perform the frame dropping of only the B picture, not only the previous image is frozen as shown in FIG. 5D, but also the later image is also frozen. Such freeze control is performed by the header section provided at the beginning of the GOP or the header section provided at the beginning of the picture data.
This is done by providing a flag or the like.

Example 3. 8 is a block circuit diagram of a third embodiment of the present invention, the same reference numerals as those in FIG. 3 denote the same parts, 11 is a memory for storing the motion vector amount for one screen, and 12 is the current motion vector. A subtracter for subtracting the data and the motion vector amount of one screen before, 13 is a subtractor 11
An absolute value detector 18 for determining the absolute value of is a reference numeral 18 is a frame drop amount determination table for determining the number of frame drops from the motion vector conversion amount detector 4 and the absolute value detector 13.

FIG. 9 is a characteristic diagram showing the detection limit (whether or not frame drop is noticeable) of a frame drop person to the motion vector correlation amount corresponding to the smoothness of motion for each picture. However, in the image with jerky movements, dropped frames are not noticeable. On the other hand, in a smoothly moving image represented by, for example, panning and zooming at the time of shooting (a moving image is close to a constant speed), dropouts are easily noticeable.
Even if PAL, which is a European TV system, is converted to NTSC, which is a Japanese American TV system, it is easy to notice dropped frames and screen freezes due to TV format conversion adjustments when panning and zooming. In addition, even when converting a motion picture film to the NTSC format, a slight freeze may be noticeable in smooth motion.

The third embodiment is an optical disk data file system utilizing the characteristics of FIG. 9 and is stored in the memory 11 in order to detect the motion vector correlation amount corresponding to the smoothness of the motion of the screen. The subtractor 12 calculates the difference between the motion vector amount for one screen and the current motion vector amount, and the absolute value detection circuit 13 calculates the absolute value to detect the motion vector correlation amount. Based on the correlation amount, the frame drop amount determination table 18 determines the number of frame drops.

Further, in the system of the third embodiment, the correlation amount and the absolute amount of the motion vector are two parameters for dropping frames, but by setting the dropping table 18 as shown in FIG. It will be possible. This Figure 1
0 corresponds to the motion vector amount and the motion vector correlation amount, and d ~
When a table like this is used, even if it is too large or too small, it is possible to use a non-linearity such that four frames are used up to a. In contrast to the characteristics, it is possible to give a linear characteristic to the motion vector correlation amount such that frame drop is allowed when the correlation is weak, and to make a complicated decision as a whole.

Further, as shown in FIG. 11, when frame dropping is performed from only two types of moving image data of 30 frames / second and moving image data of 24 frames / second, if the correlation is extremely weak, 24 frames / second. A method in which the second is fixed and 30 frames / second is fixed when the correlation is strong, and switching is performed according to the motion vector data amount only when the correlation amount is medium is considered.

Example 4. FIG. 12 is a block circuit diagram of a fourth embodiment of the present invention, in which the same reference numerals as those in FIG. 1 denote the same parts respectively, 15 is a memory, 16 is a motion vector amount detecting means within 1 GOP, and 17 is a motion vector correlation. It is a quantity detecting means.

In the fourth embodiment, after the image data is temporarily recorded in the memory 15, the absolute amount of motion is detected by the motion vector amount detecting means 16 in 1 GOP, and the motion vector correlation amount before and after the picture in 1 GOP is detected. The motion vector correlation amount detection means 17 detects it, and the frame drop amount determination table 18 determines the frame drop amount within 1 GOP based on these two pieces of information.

In the fourth embodiment, not only the motion vector amount as in the first to third embodiments but also the moving image information itself is used. As shown in FIG. 12, a memory for storing image data is used. 15 is used to detect the motion vector amount for each picture by taking the difference from the data of the previous picture, and the difference information obtained by taking the difference for each picture is compared with the previous difference information. Motion vector correlation (whether motion is smooth or jerky)
Is detected, and the detection means 16 and 17 add up the GOPs in units of 1 GOP to obtain the input information of the frame drop amount determination table 18.

According to the fourth embodiment, it is possible to determine the number of dropped frames by directly storing the image data without detecting the motion vector amount.

Example 5. 13 is a block circuit diagram of a fifth embodiment of the present invention. The same reference numerals as those in FIG. 11 denote the same parts, 19 is a motion compensation predictor, and 20 is a DCT.
Encoder, 21 is a quantizer, 22 is a variable length encoder, 2
3 is a code amount counter that counts the code amount per 1 GOP, 24 is a code amount memory that stores the code amount count value,
25 is a GOP rate setting device for setting an ideal GOP rate from the output of the code amount memory, 26 is a code amount allocating device for allocating the code amount based on the output from the GOP rate setting device 25, and 27 is the current code. Amount and the ideal code amount are subtracted, 28 is a switch for separating the tentative encoding and the main encoding, and 34 is the ideal number of frames from the frame drop amount determination table 18. It is a frame number allocator.

In the fifth embodiment, in order to make the image quality on the disc uniform and further improve the file efficiency, the data rate per GOP as shown in FIG. It is intended to be filed in.

Generally, in a video compression method typified by MPEG or the like, as shown in FIG. 13, a motion compensation predictor 19 which uses a predicted image from the preceding or following I picture or P picture as a reference screen, It is composed of a DCT encoder 20, a quantizer 21 for adjusting the code amount, and a variable length encoder 22. First, the image quality is made constant, and the switch 28 is turned down to perform the first temporary encoding. Further, the code amount per GOP at the time of the tentative encoding is counted by the code amount counter 23, and the code amount corresponding to one disk is stored in the code amount memory 24. In addition, the GOP rate setting 25 is set again so that the code amount for one disk stored in the code amount memory 24 becomes an ideal code amount in consideration of the total data storage amount of the entire disk. The target rate is stored in the allocator 26.

Next, the switch 28 is tilted in the direction of the main encoding and the second encoding operation is performed.
At this time, in the quantizer 21, the code amount assigner 26
The feedback control is performed so that the output of the code length and the output of the code amount counter 23 in the variable-length encoder 22 at the time of the second encoding are matched, whereby the code amount assigner 26 at the time of the first encoding. It is controlled so as to substantially match the stored ideal code amount.
At the same time, at the time of the first encoding operation, 1G
The motion vector amount and motion vector correlation amount in the OP are detected by the detection means 16 and 17, and the frame drop amount determination table 1
The frame drop amount is determined at 8 and stored in the frame number assigner 34. The actual frame dropping operation is performed by the disc format encoder 5 at the time of the second main encoding.
This is done by deleting the picture data and additionally writing the freeze information of the screen, which is required at the time of reproduction, in the header portion.

As described above, in the system of the fifth embodiment, the image quality of each GOP unit is made constant, and the ideal code amount for variable rate is allocated, and at the same time, the ideal number of frames is allocated to 1 GOP unit for the second time. Since the main encoding is performed, the file efficiency can be further improved.
In particular, in the area of the original image where the compression ratio cannot be improved in terms of image quality, it is possible to allocate a larger number of bits as much as the amount of dropped frames is allowed, resulting in higher image quality.

Example 6. FIG. 14 is a block circuit diagram of the sixth embodiment of the present invention, and the same reference numerals as those in FIG. 13 denote the same parts respectively, 29 is a frame memory, 30 and 31 are quantizers, and 32 is a differential element. The subtractor 33 is a subtractor that constitutes an addition element.

The frame memory 29 stores a motion vector amount for one frame, and using this motion vector amount, the motion vector absolute amount and the motion vector correlation amount are detected, and the frame drop amount determination table 18 detects the number of frame drops. Is determined and stored in the frame number assigner 34 to allow a variable data rate and dropped frames.

In the sixth embodiment, in addition to the method using the motion vector as in the fifth embodiment, the number of frames is determined from the output of the video A / D converter 1. Also in the sixth embodiment, the first encoding operation and the second encoding operation are performed.
The fifth embodiment is that the feedback control is performed so that the ideal code amount stored in the code amount assigner 26 is achieved by performing the encoding operation for the second time.
Is the same as in. However, in the case of the sixth embodiment, the output of the video A / D converter 1 is directly stored in the frame memory 29, and the difference between the preceding and following pictures is obtained by the subtractor 32 to detect the motion amount for each picture. The signal after the above subtraction is stored in the frame memory 29 again, and the output of the subtractor 32 which is temporally preceding and following is further taken by another subtractor 33 to detect the correlation value (smoothness) of the motion, The drop amount determination table 18 determines the frame drop amount.

The frame number allocation amount thus obtained is
It is stored in the frame number assigner 34, and at the time of main encoding, the B picture is deleted in the disc format encoder 5. As described above, also in the case of the sixth embodiment, the same effect as in the case of the fifth embodiment can be obtained, and in the area portion of the original image in which the compression rate cannot be increased in terms of image quality, the number of dropped frames is increased by the amount allowed. It has become possible to allocate the number of bits of, and high image quality has become possible.

[0075]

According to the first aspect of the present invention, the number of pictures is reduced when the total amount of motion vectors is below a certain level. Therefore, the optimum number of pictures per second is set according to the speed of the screen movement. Is possible, and by reducing the number of pictures above,
The file efficiency of a digital compressed moving image signal using motion compensation to an optical disc is improved.

According to the second aspect of the present invention, in addition to the case where the total amount of the motion vectors is equal to or less than the constant level shown in the first aspect of the invention, the second level or more having a set value larger than the first level Since dropping of frames is allowed in, the speed of movement in picture units is high, and it is possible to reduce the number of frames even for images that show movements that the human eye cannot follow, for example, and file efficiency for optical discs is improved. To do.

According to the third aspect of the present invention, the digital compressed image is divided into an image for which intra-frame DCT is partially present and an image with motion compensation prediction other than that. In addition to enabling special playback using images, it is possible to increase the compression rate using motion-compensated prediction images, and further, of the images with motion-compensated prediction, only B-pictures that require prediction from preceding and following pictures. By allowing frame dropping, it has become possible to avoid the unpredictable state of pictures before and after the frame dropping.

According to the fourth aspect of the present invention, the number of dropped frames can be determined in consideration of not only the absolute amount of motion speed for each picture but also a certain degree of motion (smoothness). However, it is possible to prevent unreasonable frame dropping when frame dropping is easy to notice due to smooth movement, and since frame dropping is allowed in images with jerky movement, further file Efficiency is improved.

According to the invention of claim 5, the digital compressed video is divided into an image for which the intra-frame DCT partially exists and an image accompanied by motion compensation prediction other than that, and the motion compensation prediction is performed. By using P-pictures that regularly exist in the accompanying images and that use a prediction screen from one direction, it is possible to accurately extract the absolute amount and smoothness of the motion vector amount, and to move back and forth. Since dropping of frames is allowed only for B-pictures that require prediction from pictures, it is possible to avoid unpredictable states of preceding and following pictures due to dropping of frames.

According to the invention of claim 6, the encoding operation at the time of manufacturing the disk is performed twice, and first, each G
Since the optimal code amount and the optimal number of frames are set for each OP and the compressed video bitstream is realized so that the second encoding operation achieves the optimal value, not only the amount of code but also the number of frames are made variable. Greatly improve efficiency.

According to the seventh aspect of the invention, the digital compressed video is divided into an image for which intra-frame DCT is partially present and an image with motion compensation prediction other than that. Of the accompanying images, by using P-pictures that are regularly present and use the prediction screen from one direction, it is possible to accurately extract the absolute amount of motion and smoothness, and it is also possible to extract from the preceding and following pictures. Since dropping of frames is allowed only for B pictures that require prediction, it is possible to avoid the unpredictable state of the preceding and following pictures due to dropping of frames.

[Brief description of drawings]

FIG. 1 is a block circuit diagram of a first embodiment of the present invention.

FIG. 2 is a diagram showing the number of dropped frames with respect to a motion vector amount according to the first embodiment.

FIG. 3 is a block circuit diagram of a second embodiment of the present invention.

FIG. 4 is a diagram showing a bitstream structure according to a second embodiment.

FIG. 5 is a diagram showing a video reproduction screen according to the second embodiment.

FIG. 6 is a characteristic diagram showing a detection limit of dropped frames with respect to the speed of movement.

FIG. 7 is a diagram illustrating a number of dropped frames with respect to a motion vector amount according to the second embodiment.

FIG. 8 is a block circuit diagram of a third embodiment of the present invention.

FIG. 9 is a characteristic diagram showing a detection limit of dropped frames with respect to smoothness of movement.

FIG. 10 is a diagram showing an example of a frame drop amount determination table according to the third embodiment.

FIG. 11 is a diagram illustrating another example of the frame drop amount determination table according to the third embodiment.

FIG. 12 is a block circuit diagram of a fourth embodiment of the present invention.

FIG. 13 is a block circuit diagram of a fifth embodiment of the present invention.

FIG. 14 is a block circuit diagram of a sixth embodiment of the present invention.

FIG. 15 is a block circuit diagram of a conventional optical disc recording / reproducing apparatus.

FIG. 16 is a diagram showing a picture structure of digital moving image information.

FIG. 17 is a diagram showing a picture structure of digital moving image information.

FIG. 18 is a diagram showing a picture structure of digital moving image information.

FIG. 19 is a diagram showing a data amount at a variable rate.

[Explanation of symbols]

1 video A / D converter, 2 video information compression means, 3
Motion vector amount detecting means, 4 frame drop amount determining means, 5
Disk format encoder, 6 Modulating means, 7
Recording data file, 8 ROM disc mastering device, 9 creation ROM disc, 10 data conversion table, 11, 15 memory, 12, 27, 32, 33
Subtractor, 13 Absolute value detector, 14 Motion vector conversion amount detector, 16 Motion vector amount detecting means, 17 Motion vector correlation amount detecting means, 18 Frame drop amount determination table, 1
9 motion compensation predictor, 20 DCT encoder, 21, 3
0, 31, quantizer, 22 variable length encoder, 23 code amount counter, 24 code amount memory, 25 GOP rate setting device, 26 code amount assigner, 28 switch, 29
Frame memory, 34 frame number assigner, 35 frame sector converting means, 36 encoder, 37 modulator,
38 laser drive circuit, 39 laser output switch, 4
0 optical disk, 41 optical head, 42 actuator, 43 traverse motor, 44 disk motor,
45,46 motor control circuit, 47 motor drive circuit, 4
8 reproduction amplifier, 49 demodulator, 50 decoder, 51
Frame sector inverse conversion means, 52 information expansion means, 5
3 D / A converter.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical indication location H04N 5/93 7/30 7/32 H04N 7/133 Z 7/137 Z (72) Inventor Mishima Hidetoshi Nagaokakyo Baba Institute 1 Video System Development Laboratory, Mitsubishi Electric Corporation (72) Inventor Yoshinori Asamura Nagaokakyo Baba Institute 1 Video System Development Laboratory

Claims (7)

[Claims] 1. When recording, digital video information is compressed based on motion estimation information between frames that are temporally preceding and succeeding, and a motion that increases or decreases according to the motion of the video for each frame included in the compressed video information. When the vector amount is smaller than the predetermined absolute amount, the number of frames per second is reduced and recorded on the recording medium, and at the time of reproduction, the frames before and after the reduced frame are displayed in duplicate and a predetermined required frame is displayed. Video recording / playback method to secure the number. 2. When recording, digital video information is compressed based on motion estimation information between frames that are temporally preceding and following, and a motion that increases or decreases according to the motion of the video for each frame included in the compressed video information. When the vector amount is smaller than the range between the first absolute amount and the second absolute amount, which is determined in advance, the number of frames per second is reduced in proportion to the above-mentioned motion vector amount, and the vector amount is smaller than the above range. If it is larger, the number of frames per second is reduced so as to be inversely proportional to the above-mentioned dynamic vector amount, and the result is recorded on the recording medium. During playback, the frames before and after the reduced frame are displayed in duplicate to secure the required number of frames Video recording / playback method. 3. The video information recorded on a recording medium includes an I picture, which is video information for performing intra-frame DCT, and a P picture, which is video information by DCT coding for performing forward motion compensation. Motion compensated D using the above I and P pictures located before and after as a reference screen
It is a series of video information blocks in units of several frames to several tens of frames in which a B picture to be CT-encoded is mixed, and a motion vector amount is stored in the video information block unit, and according to the motion vector amount. The video recording / reproducing method according to claim 1, wherein the B picture of the image data is deleted. 4. When recording, the digital video information is compressed based on motion estimation information between frames that are temporally preceding and following, and the digital video information is changed according to the magnitude of the motion of each frame included in the compressed video information. The number of frames per second when the total amount of moving vector amounts to be set is smaller than the first predetermined absolute amount and the amount of change of the moving vector amount for each screen is larger than the second predetermined absolute amount. Is recorded on a recording medium with reduced number of frames, and at the time of reproduction, the frames before and after the reduced frame are overlapped and displayed to ensure a required number of frames. 5. The video information recorded on a recording medium includes an I picture, which is video information for performing intra-frame DCT, and a P picture, which is video information by DCT encoding for performing forward motion compensation, in terms of time. Motion compensated D using the above I and P pictures located before and after as a reference screen
It is a sequence of video information blocks in units of several tens of frames from the number of mixed B pictures to be CT-encoded, and the motion vector amount is stored in the video information block unit. A motion vector correlation amount in the video information block is detected by comparing each picture, and a conversion table that determines the number of dropped frames of the B picture based on the total amount and the motion vector correlation amount is used in the video information block. 5. The video recording / reproducing method according to claim 4, wherein the amount of dropped frames is determined. 6. At the time of recording, a code amount after encoding and a motion vector amount of the video information block are stored for each unit of a video information block composed of a plurality of pictures at the time of the first encoding, and the stored code is stored. From the amount and the motion vector amount, the optimum code amount and frame number for each video information block unit for recording on the recording medium for a predetermined time is calculated, and the optimum code amount is obtained at the time of the second encoding. A video recording / reproducing method in which B picture data is deleted by the number of frames described above. 7. The video information recorded on a recording medium includes an I picture, which is video information for performing intra-frame DCT, and a P picture, which is video information by DCT coding for performing forward motion compensation, in terms of time. Motion compensated D using the above I and P pictures located before and after as a reference screen
A series of video information blocks in units of several frames to several tens of frames in which a B picture to be CT-encoded is mixed, and the total amount of motion vectors is stored in units of the video information blocks. 7. The video recording / reproducing method according to claim 6, wherein a conversion table for determining the number of dropped frames of the B picture is used to determine the amount of dropped frames in the video information block based on a vector correlation amount.