HK1009874A - Digital vcr with non-standard speed playback - Google Patents

Description

Digital video cassette recorder for playback at non-standard speed

The present invention relates to the field of digital video recording, and in particular to the reproduction of high definition video signals at non-standard speeds.

Background of the invention

Digital video cassette recorders using the helical scan format have been proposed by the standardization committee. The proposed standard specifies digital recording of Standard Definition (SD) television signals (e.g., NTSC or PAL), or high definition television signals having an MPEG compatible structure (e.g., a proposed "Grand Alliance" signal). SD recorders use a format of compressed component video signals using intra/intra field DCT with adaptive quantization and variable length coding. The SD track format comprises an azimuthal recording of 10 μm tracks without guard bands, 10 or 12 tracks per NTSC or PAL frame, respectively. The tape cartridge used was a metal evaporation recording medium 1/4 inches wide. The SD digital VCR (or DVCR) is intended for consumer use and has data recording capabilities sufficient to record ntsc (pal) signals or advanced television signals.

Advanced Television (ATV) signals have been developed by the international consortium of the large alliance (GA). An illustrative document entitled "description of HDTV systems in the grand alliance" has been published in the 48 th annual meeting of the broadcast project proceedings of 1994. The GA signal uses an MPEG compatible coding method that employs intra-frame coded pictures called I-frames, forward predicted frames called P-frames, and bi-directional predicted frames called B-frames. These three frames occur in a group called GOP (group of pictures). The number of frames in a GOP may be specified by a user, and may include, for example, 15 frames. Each GOP includes an I-frame followed by B-frames, with P-frames interleaved between them.

In an analog consumer VCR, the "trick play" (TP) feature (e.g., fast search, fast motion or slow motion of the image in forward or reverse direction) can be easily implemented because each recorded track typically includes a field. Therefore, when reproducing at a non-standard tape speed, the reproducing head is caused to cross over a plurality of tracks to restore recognizable horizontal image segments. When recording GOPs of an ATV signal using I, P, B frames, multiple tracks may be occupied on the tape, for example, 10 tracks per frame and 150 tracks per GOP. In short, when the DVCR is operated at a non-standard reproduction speed, the playback head switches zones or segments from multiple tracks. Unfortunately, these track segments no longer represent regions from discrete recordings in sequential image frames. Instead, the segments include data formed primarily of predicted frames in the GOP. During operation at playback speed, the I frame data is recovered, which allows reconstruction of predicted B and P frames. It is clear that during "trick play" (TP) operation, the amount of I frame data recovered will gradually decrease as the TP speed increases. Therefore, the possibility of reconstructing B frames and P frames from the reproduced I frame data blocks is practically equal to zero. Thus, in order to provide the features of "trick play" or non-standard speed playback, it is necessary to record specific data that can be reconstructed into a picture without using information of adjacent frames when reproducing in the TP mode. Furthermore, since special data for "trick play" is recorded, the actual position of the track must allow it to be recovered in the TP mode.

Summary of The Invention

According to one arrangement of the present invention, a method for recording a digital video image representative signal on a magnetic tape having a spiral scan track format, the method comprising the steps of: determining a common area on each transducing track during a predetermined number of forward and reverse regeneration rates; processing the digital video image representative signals to have a data rate suitable for recording the layout in the common track area; and multiplexing the digital video image representative signals with processed digital video image representative signals for the recording layout on the magnetic tape, such that the processed digital video image representative signals are located within the common track area.

According to another arrangement of the present invention, a magnetic tape for a digital video cassette recorder, the magnetic tape recording digital signals representing digital image signals in an MPEG compatible format, the digital signals being recorded in sequential tracks on the magnetic tape, the magnetic tape comprising: a recording portion of the first data signal and the second data signal existing in each track; the first data signal represents a digital image signal represented in an MPEG compatible format and provides a first source for image reproduction; the second data signal also represents a digital image signal represented in an MPEG compatible format but having less data than the first data signal, the digital image signal represented in the MPEG compatible format being decoded to obtain a signal from which the second data signal is derived, the second data signal providing a second source for image reproduction.

According to yet another arrangement of the present invention, a method for determining a recording layout for a digital video image representative signal on a magnetic tape having a spiral scan track format, the method comprising the steps of: selecting a predetermined number of regeneration speeds in the forward direction; selecting a predetermined number of regeneration speeds in the reverse direction; determining a common area of tracks on each of the spiral scan transducing tracks during the periods of predetermined regeneration rates in the forward and reverse directions; determining a data capacity of the common track area; processing the digital video image representative signals to have a data rate suitable for the recording layout in the common track area; and multiplexing the digital video image representative signals with processed digital video image representative signals for the recording layout on the magnetic tape, such that the processed digital video image representative signals are located within the common track area.

Brief description of the drawings

FIG. 1 illustrates a track map showing the recording of the locations of data zones specified for standard definition DVCR;

FIG. 2 shows the path of playback of the head in the sync block recovery region at twice the playback speed;

FIG. 3 shows the path of playback of the head in the sync block recovery region at 4 times playback speed;

FIG. 4 shows the path of playback of the head in the sync block recovery region at 8 times playback speed;

FIG. 5 shows the path of playback of the head in the sync block recovery region at 16 times playback speed;

FIG. 6 includes a table showing recovered audio and video sync blocks at various trick mode playback speeds;

FIG. 7A shows the recovered sync blocks at 2, 4, 8, 16 times playback speed;

FIG. 7B shows the recovered sync blocks for common at 2, 4, 8, 16 times playback speed;

fig. 8 shows a first embodiment of a recorded track pattern representing advantageous sync block positions of the inventive trick play data layout;

FIG. 9 shows the path of the playback head at 3 times the play speed and the track area recovered by the sync block;

FIG. 10 shows the path of the playback head and the track area recovered by the sync block at 9 times the playback speed;

FIG. 11 shows the path of the playback head at 19 times the play speed and the track area recovered by the sync block;

FIG. 12 shows the path of the playback head at-1 times the play speed and the track area recovered by the sync block;

FIG. 13 shows the path of the playback head at-7 times the play speed and the track area recovered by the sync block;

FIG. 14 shows the path of the playback head and the track area recovered by the sync block at-17 times the playback speed;

fig. 15 shows the sync blocks restored at 3, 9, 19 times the play speed in the forward direction and at 1, 7, 17 times the play speed in the reverse direction;

FIG. 16 shows a second embodiment of a recorded track pattern representing the positions of the inventive sync blocks used for recording the inventive trick play data;

fig. 17 shows a video data area recorded with an ATV signal and a "trick play" signal of the present invention;

fig. 18A shows arrangement of data in an SD sync block; FIG. 18B shows an advantageously formatted sync block for recording an ATV data signal and a "trick play" data signal of the present invention;

fig. 19 is a system block diagram of an ATV digital video cassette recorder using the "trick play" recording and playback features of the present invention;

fig. 20 is a system block diagram of a "trick play" encoder and decoder of the present invention;

fig. 21 is a system block diagram showing an SD recorder and the control of "trick play" and high definition video playback of the present invention.

Detailed Description

Fig. 1 shows the recorded track format for a Standard Definition (SD), spiral scan digital video cassette recorder intended for consumer use. The effective data area shown in fig. 1 includes 4 areas on which various specific types of data are recorded. The ITI (insertion and tracking information) data area is used for tracking and editing, followed by an edit gap G1. The audio data area occupies 14 synchronous blocks, and the serial number is 0-13. The second edit gap G2 immediately follows the audio data area, followed by the video data area, numbered 0-148, that includes 149 sync blocks. The video data area is immediately followed by the third edit gap G3 and the sub code recording area is immediately followed by G3. The digital video recording rate of a digital video recorder (DVCR) is defined to be 24.948 Mb/s. This video bit rate may be used to record component video signals decoded from ntsc (pal) signals, or from processed advanced television signals, such as large association (GA) signals. Fig. 21 illustrates a simplified block diagram of the DVCR 350. The DVCR350 includes a head drum 510, the head drum 510 including a plurality of recording and reproducing heads coupled to a playback processor that generates 4 output signals 351, 352, 353, 354. The playback signal 354 represents the ATV data stream and blocks 359, 120, 130 depict the data processing path. The replay signal 353 represents "trick play" picture data, and the replay signal 353 is coupled to a subsequent "trick play" picture data processing section. The processing and selection between "trick play" pictures and ATV pictures will be described later. The cartridge 510 is shown inserted into the DVCR350 to introduce the tape 504 around the head drum 510.

SD track formats can be recorded using various head layouts on the drum or cylinder and various drum rotational speeds. The track diagrams immediately following fig. 1 show the path or track of the playback head for various "trick play" speeds. Furthermore, two possible head drum configurations are shown, namely: a pair of double-bit heads and two single heads facing each other at 180 DEG at both ends of the drum diameter.

Figures 2 to 5 show the playback head path for a selected one of the "trick play" playback speeds. The tape was recorded according to the SD digital recorder format, i.e. track width 10 μm, azimuth recording, without guard band, and the tape was shown to be reproduced with a reproduction head having a pole face width of 15 μm.

Fig. 2 shows the path or track of the playback head in double speed reproduction. The tracks shown are for a single pair of double-sided playback heads. It is assumed that the playback head can restore the sync block data from the recorded track up to half the width of the scanned recorded track. These figures show the track areas for the sync block data recovery with double hatching.

Fig. 3, 4, 5 show the playback trajectories at 4, 8, 16 times the play speed, respectively.

Fig. 6A is a table showing track numbers at the TP speed shown in fig. 2 to 5 and numbered sync blocks recovered from the audio data area. Fig. 6B shows the track and the numbered sync blocks recovered from the video data zone at the shown trick play speed.

The recovered video sync block data, shown with double hatching in fig. 2, 3, 4, 5, is combined with the numbered sync blocks in the table of fig. 6B for 2, 4, 8, 16 times TP speed, shown in fig. 7A. Fig. 7B shows the track area and the recovered numbered sync blocks common to those 4 speeds. Thus, fig. 7B indicates track positions identified by sync block numbers at which data can be recorded and recovered at playback speeds and 2, 4, 8, and 16 times playback speeds.

Fig. 8 shows a recording track comprising an ITI (insertion and tracking information) recording area, an edit gap G1, an audio data recording area occupying 14 sync blocks numbered 0-13. During ATV operation, audio and video data are transmitted within the ATV data transport stream so that an audio data field is not required for audio data applications and can be used for ATV and "trick play" data recording. A second edit gap G2 immediately follows the audio data area, and a video data recording area including 149 sync blocks numbered 1-149 immediately follows G2. The third edit gap G3 immediately follows the video data area, and immediately follows the sub code recording area after G3. The recorded tracks of fig. 8 show: for an advantageous first embodiment of the synchronization block allocation of the TP-data record of the invention, here 5 synchronization blocks are used for the audio zone and 40 synchronization blocks are used for the video zone. Thus, TP video data can be recorded with 45 sync blocks at each scan for recovery at both standard and non-standard playback speeds. These 45 TP sync blocks provide an effective playback data rate of about 1.06Mb/s at rated speed.

Fig. 9-11 show the path of the playback head at 3, 9, 19 "trick play" speeds and the head trajectory for both double-position heads and facing heads at 180 ° on either end of the drum diameter.

Fig. 9 shows the track area where the sync block is restored at 3 times the playback speed. Tracks T1 and T2 indicate reproduction with a pair of double-bit heads, and tracks T1 and T4 indicate reproduction with heads facing 180 °. Fig. 9 shows that, regardless of the playback head configuration, there are some track regions and, therefore, some sync blocks that cannot be recovered.

Fig. 10 shows the track area where the sync block is restored at 9 times the play-speed. Tracks T1 and T2 indicate reproduction with a pair of double-bit heads, and tracks T1 and T10 indicate reproduction with heads facing 180 °.

Fig. 11 shows the track area where the sync block is restored at 19 times the playback speed. Tracks T1 and T2 indicate reproduction with a pair of double-bit heads, and tracks T1 and T20 indicate reproduction with heads facing 180 °.

Fig. 12 shows the track area where the sync block is restored at-1 times the playback speed. Tracks T3 and T4 indicate reproduction with a pair of double-bit heads, and tracks T3 and T4 indicate reproduction with heads facing 180 °.

Fig. 13 shows the track area where the sync block is restored at-7 times the playback speed. Tracks T17 and T18 indicate reproduction with a pair of double-bit heads, and tracks T17 and T10 indicate reproduction with heads facing 180 °.

Fig. 14 shows the track area where the sync block is restored at-17 times the play-speed. Tracks T21 and T22 indicate reproduction with a pair of double-bit heads, and tracks T21 and T4 indicate reproduction with heads facing 180 °.

The sync blocks shown in fig. 9-14 that recover at various speeds in the forward and reverse directions are combined and illustrated as a single trace. FIG. 15A shows numbered sync blocks at 3 times tape speed; FIG. 15B shows the Sync Block (SB) recovered at 9 times the tape speed; FIG. 15C, for 19 times belt speed; FIG. 15D, for-1 times the belt speed; FIG. 15E, for-7 times belt speed; FIG. 15F, for-17 times belt speed. Fig. 15G shows an analysis of the common recovered sync block. Thus, FIG. 15G shows the numbered sync blocks recovered at 3, 9, 19 times the tape speed in the forward direction and 1, 7, 19 times the tape speed in the reverse direction.

Fig. 16 shows a second embodiment with advantageous track positions identified by sync block numbers, where 45 sync blocks of "trick play" video data of the invention can be recorded and recovered at play speed, 3 times, 9 times, 19 times in the forward direction and 1 times, 7 times, 17 times in the reverse direction.

The ATV bitstream can be recorded in a data capacity of 105 sync blocks, wherein 14 sync blocks are from an audio data area and 91 SBs (sync blocks) are from a video data area. The video data of the "trick play" of the present invention can be recorded using 45 SBs in the video data area. In fig. 17, a video data area showing an SB (sync block) structure for ATV data recording is shown.

Fig. 18A and B show the data structure of the sync block SB in the video data area. Fig. 18A shows a Standard Definition (SD) formatted sync block. The SD sync block includes 90 bytes, and 77 bytes include 6 sets of Discrete Cosine Transform (DCT) coefficient data. Each set of DCTs includes a DC coefficient value followed by an AC coefficient value with decreasing weight. Fig. 18B shows a sync block formatted with the "trick play" data of the present invention. The "trick play" data is compressed, discrete cosine transformed, variable length coded as will be described with respect to fig. 20. Two compressed TP macroblocks can be recorded into one sync block formatted as shown in fig. 18B.

Having identified the advantageous sync block locations for "trick play" reproduction at various speeds in both the forward and reverse directions, the "trick play" video data must be derived from the ATV data stream. As described above, during "trick play" playback, it is necessary to be able to decode the recorded TP sync blocks so that pictures can be generated without reference to or prediction from adjacent image frames. Obviously, "trick play" video data can be derived from intra-coded video (I-frames). However, due to the low repetition rate of I-frames in each GOP, when "trick play" video is derived exclusively from I-frames, the motion of the reproduction may be made to flicker or jump in the "trick play" mode. In this way, to avoid jerkiness in the "trick play" motion, the ATV or MPEG like data stream is decoded to obtain a video signal from which the video for the "trick play" recording process is advantageously obtained. Thus, each frame of decoded pictures resulting from I, P, B frames is processed to produce a corresponding "trick play" frame for recording. Thus, each frame recorded in the GOP includes a corresponding, processed "trick play" picture, which can be decoded during "trick play" reproduction to provide a picture in which motion has been smoothed.

The DVCR format allocates 10 tracks per ATV frame so that the same number of tracks are selected for "trick play" video data. In each track, ATV data may be allocated to 105 SBs, so that one recorded ATV frame corresponds to 1050 SBs. Since the video data for "trick play" can be allocated to 45 sync blocks in each video zone, a total of 450 SBs are available for data recording for "trick play". Therefore, each "trick play" video frame must be compressed to occupy the data capacity provided by the 450 sync blocks. The desired "trick play" video data compression ratio can be expressed as 450: 1050, which is approximately 2.3: 1.

Fig. 19 is a block diagram of an advanced television receiver using the trick play mode processing method of the present invention for recording MPEG like data streams on a Standard Definition (SD) digital video cassette recorder. The block diagram comprises: an ATV decoder 100, a trick play processor 200, and an SD DVCR 300. An exemplary modulated radio frequency advanced television signal is received by antenna 101 and coupled to an input of ATV decoder 100. The modulated rf signal may also be transmitted to decoder 100 via a cable television distribution system. The decoder 100 includes a channel demodulator 110, the channel demodulator 110 extracting a modulated ATV bitstream signal, such as MPEG, from a radio frequency carrier. The bit stream has a data rate of 19.3Mb/s and is coupled as output signals 111 and 112. The bitstream 111 is coupled to a transport packet decoder 120, which simply separates video data packets 121 from audio data packets 122 by the transport packet decoder 120. The video data packets 121 are coupled to a video codec 130 and the video codec 130 reconstructs the HD video image signal. Video signal 131 is coupled to video processor and sync generator 150. video processor and sync generator 150 generates high definition video signals, e.g., luminance and color difference signals Cr and Cb, having an original aspect ratio of 16: 9 at output 151. The video processor and sync generator 150 also receives a second input signal 132 from the pixel converter 280 of the trick play processor 200. The audio data packets 122 are coupled to an audio codec 140, and the audio codec 140 extracts and generates the original audio signal, which forms an audio output signal 141.

The MPEG like bitstream signal 112 is coupled to a bitstream rate converter 310. the bitstream rate converter 310 converts the 19.3Mb/s bitstream to a data rate of 24.945Mb/s as required for processing and recording by the SD recorder. The output of the rate converter 310 is coupled to an inner and outer parity generator 320. the inner and outer parity generator 320 generates a reed solomon error correction code which is included in the video data recorded in the video zone as depicted in fig. 1. After insertion of the RS error correction code, the data stream is coupled to the SD video data sync block construction means 330, and the sync block construction means 330 constructs the video data sync block structure required by the SD video recorder format.

Block 340 of fig. 19 constructs the audio and video zones according to the SD format, where the video data zone includes: the processed ATV data from block 330 plus the inventive "trick play" video data 251 from block 250 of the "trick play" video processor 200.

Fig. 17, 18A, 18B show the format or structure of the SD video area. Fig. 18A and 18B show that this zone includes: video front signal, video data and error correction code of 149 sync blocks, video rear signal. The sync blocks are numbered 1 to 149. Fig. 18A shows the format used during recording of an NTSC image source. Fig. 18B shows ATV video data that is advantageously recorded, occupying, for example, 105 sync blocks. The "trick play" video data of the present invention can be recorded with occupying, for example, 45 sync blocks, and the video auxiliary data can be recorded with two sync blocks. The outer parity error correction data is recorded using 11 sync blocks.

ATV video zone data, including "trick play" data and audio zone signals, is coupled from block 340 to a Standard Definition (SD) digital video cassette recorder 350. The SD recorder may also receive an analog ntsc (pal) input signal for recording. The analog signal is decoded into luminance and color difference components, which are 4: 1 for the NTSC input signal, sampled at 13.5MHz, and digitized to 8 bits. Compressing the digitized NTSC signal according to the SD recording format using intra-field/intra-frame DCT applied to 8 x 8 image blocks; followed by adaptive quantization and modified two-dimensional huffman coding. In order to prevent uncorrectable data errors due to damage to the recording medium, blocks are shuffled or redistributed within each frame. Since blocks have been shuffled prior to recording, any significant playback errors associated with the media will be distributed over the decoded frames due to the complementary de-shuffling used during playback. Thus, significant, possibly uncorrectable, and thus visible, errors are distributed and possibly corrected using both internal and external Reed-Solomon error correction codes. After compression, the data is encoded for recording using a 24: 25 transform, the 24: 25 transform allowing the frequency response to be shaped to provide automatic tracking capability for playback.

The SD recorder 350 reproduces four output signals: 351. 352, 353, 354. Output signals 351 and 352 are baseband analog signals that include video components Y, Cr and Cb, respectively; and an audio signal. Signal 351 includes a video component that is coupled to NTSC sync generator and encoder 360, sync generator and encoder 360 providing additional blanking and sync pulses for viewing on a video monitor. These components may be encoded to produce an NTSC signal for viewing on a standard definition television receiver.

The SD recorder 350 produces an ATV data bitstream output signal 354 and a "trick play" data bitstream output signal 353. The signal 353 is coupled through error correction block 359 to block 260 of the ATV and "trick play" processor 200 for decompression and subsequent upconversion to the ATV signal format. The operation of the trick play processor 200 will be described with reference to fig. 20.

The data bit stream 354 is coupled to the block 120 of the ATV decoder 100 through an error correction block 359 where the replayed transport packets are decoded. The decoded ATV signal 131 is coupled from the video compression decoder 130 to the line to line converter 210 of the ATV and "trick play" processor 200. The ATV signal comprises luminance, color difference signals Cr and Cb, which may comprise, for example, 1080 active horizontal scanning lines, each line comprising 1920 pixels or samples. The line-to-frequency converter 210 reduces the number of active scan lines to 1/3, i.e., 360 lines. In this way, the luminance and color difference signals are processed to form a "trick play" video signal having a vertical resolution of one third of the original ATV signal. The line number conversion is performed using a vertical low-pass filtering function. The line frequency reduced signal is coupled from the converter 210 to a pixel converter 220, and the pixel converter 220 reduces the number of pixels to 1/3 using low pass filtering. Thus, signal 221 comprises 360 horizontal lines, each line comprising 640 pixels, and the ATV signal has been converted (down-converted) to a signal having parameters like "NTSC". This is true for signal 131 because the aspect ratio of the ATV signal is 16: 9. However, the down converted signal 221 will display an image in the form of a 16: 9 letterbox.

The downconverted signal 221 is also coupled to NTSC encoder 360 for additional synchronization and blanking and is encoded for viewing of standard definition images on a receiver or video monitor. The signal 221 is also coupled to a signal compression processor, represented by block 230, the details of which signal compression processor 230 will be described with respect to fig. 20. Briefly, however, the purpose of the signal compression processor 230 is to generate the down-converted ATV signal in compressed form. For example, the signal compression processor 230 may compress the signal 221 by approximately 2.3 times.

The compressed and downconverted signal is used to provide "trick play" video data for recording on specific sync blocks in each track, for example as shown in fig. 8 and 16. The data for each TP video frame is recorded into the 10 tracks included in each ATV SD recording frame. Thus, it can be considered that TP video data is redundantly recorded in the video data track area including the ATV SD frame. The TP video data is reproduced together with the ATV data during normal speed playback, but the TP video data may not be used in the ATV image form. However, because "trick play" data frames occur once in every 10 recorded tracks, TP frames can be recovered during normal playback, which can be stored and utilized during playback mode transitions; for example, transition from forward normal speed playback to fast "trick play" or fast search video. In the worst case, when normal speed playback is started, about 140 recorded tracks may be reproduced before one I-frame is recovered. However, since TP data frames are advantageously recorded in I, P, B frames, a "trick play" processed image can be generated immediately after either type of frame is reproduced. Thus, during the initial period of normal speed playback before decoding an I frame, a picture that has been "trick play" processed can be output. When an I-frame is available, the output can be switched from "trick play" to ATV pictures.

The compressed TP signal is coupled from block 230 to an inner parity generator 240, and the inner parity generator 240 adds reed-solomon error correction data to the TP data stream. The RS internal parity added TP video data is coupled to a TP video data sync block formatter 250, and the TP video data sync block formatter 250 generates only sync blocks of a specific number required for "trick play" reproduction at a specific speed. For example, with the allocated sync blocks shown in the embodiments of fig. 8 or 10, "trick play" reproduction at various speeds is possible. These TP video data sync blocks are output as signal 251 which is coupled to the video and audio region composition means 340 of the SD DVCR 300.

During reproduction, the SD recorder 350 reproduces the "trick play" data signal 353, which is coupled to the error correction processor 259. After error correction, the TP data stream is coupled out for signal decompression in processing block 260 of the ATV and "trick play" processor 200. The detailed operation of block 260 will be described with respect to fig. 20. Briefly, however, the downconverted ATV image is regenerated from the compressed TP data recovered from the recording medium using decompressor 260.

The "trick play" signal compression processor used by the present invention to generate data signal 251 is shown in blocks 234-238 of fig. 20. The replayed TP data may be decompressed using blocks 262-266 of fig. 20. The reduced-rate ATV signal 221 is coupled to a formatter 234, which formatter 234 formats the scan lines of the signal 221 into a two-dimensional Macroblock (MB) structure comprising 4 DCT blocks. Thus, a macroblock has a size of 8 lines and 32 pixels in total. The formatted and reduced rate macroblock signal is coupled to block 235 for discrete cosine transform. The principle of discrete cosine transform is well known, and the data rate is reduced due to the control of the quantization of the coefficients. The DCT block 235 produces two output signals that represent amplitude values that comprise the frequency coefficients for each macroblock. An output signal is coupled to block 236. block 236 pre-analyzes the magnitude of the coefficients, controlling the degree of thickness quantized by quantizer block 237. The second output of the DCT block 235 is coupled to a quantizer block 237 for quantization where the value of the quantization step is dynamically controlled in response to block 236. The quantized DCT coefficients are coupled to block 238 for variable length coding. Various methods of Variable Length Coding (VLC) are known. Briefly, however, the corresponding short code words are assigned to the most frequently occurring quantized coefficient values, while those less frequently occurring coefficient values are assigned code words of increasing code length. In this way the overall data rate of the TP video data is further reduced so that data for "trick play" frames can be recorded into the 450 sync blocks provided in the 10 recording tracks.

The variable length coded TP data is coupled to block 240 for generating and appending a reed-solomon inner parity error correction code. The TP data with RS inner parity error correction is coupled to block 250 for formatting to have a particular SD sync block structure, such as that identified in fig. 8 and 16. TP data having the desired sync block structure is coupled to the SD recorder as described for "trick play" processor block 200.

During the replay mode, the regenerated TP data stream signal 353 is coupled to the decompression block 260 by error correction in block 259, the decompression block 260 performing the signal processing performed by block 230 in reverse. The VLC TP data signal 353 is input to block 266 where variable length decoding is performed. Various methods of decoding are known, for example, look-up tables may be used to convert VLC data words back into fixed-length quantized DCT coefficients. From block 266, the TP DCT coefficients are coupled to inverse quantizer 262, which inverse quantizer 262 may be considered to perform a digital-to-analog transformation of the TP DCT coefficients. The TP DCT coefficients are coupled to block 263 and block 263 performs an inverse discrete cosine transform that produces a formatted macroblock output signal representing the TP image. At block 264, the sampled macroblock TP signal is reformatted to produce a conventional line structured image. The output signal of the reformatter 264 is processed in block 265. block 265 may provide blanking insertions and add synchronization pulses, for example. Signal 261 is output from block 265 and may be coupled out for viewing a television program on a component video monitor or encoded for viewing a television program. The second output signal 271 is coupled from block 264 to blocks 270 and 280. blocks 270 and 280 provide up conversion from a nominal "NTSC" like line and pixel format to the line frequency and horizontal pixel count required for viewing by a high definition display.

The upconverted TP video signal 131 is coupled as a second input to a video processor and sync generator 150, which video processor and sync generator 150 generates a high definition output signal 151. Video processor and sync generator 150 generates an addition of video blanking and HDTV sync waveforms. In addition, however, video processor 150 also provides a selection function for switching between ATVs and "trick play" video images. Fig. 21 illustrates in block diagram form the playback datapath for ATV data stream 354 and "trick play" data stream 353, and the coupling between the two datapaths for output selection in video processor and sync generator 150. The selection of the output image source is ultimately responsive to a start control command transmitted by the user through the control system. For example, a play command will start the mechanical structure of the VCR and switch the electronic system from EE (electronic to electronic) mode to playback. However, the true instant of output signal switching may be determined by various other control factors. For example, the most important control event might be to take an I-frame from a recorded GOP and decode it. When this occurs, decoder 130 may signal and be coupled out to control a video output selector switch in video processor and sync generator 150.

As described above, a GOP of 15 frames will occupy 150 tracks of recording, so that when the play mode is started, the video picture played back may be delayed until an I-frame has been reproduced and decoded, i.e. up to 140 tracks may need to be reproduced until an I-frame is encountered. However, since the TP data is advantageously recorded and reproduced in a normal play manner within the range of each frame in the GOP, the output video signal can be generated using the TP data without waiting for the occurrence of an I frame. Thus, the redundant nature of the TP data record advantageously provides normal speed pictures derived from the TP data at the start of normal playback, and ATV pictures can be selected after the I-frame is available when available.

When a user initiates a command to start or end a "trick play" mode, the control system, and in particular the video processor and sync generator 150, may advantageously be controlled to provide the user with more aesthetically pleasing image transitions. For example, as described above, at the start of normal speed playback, a "trick play" picture may be output before an I frame is obtained and decoded. Another use of TP video data may be that TP video data that is recovered and stored during normal playback may be used with TP data that was converted during playback speed transitions during transitions to "trick play" playback speeds. Such an application of TP data provides another way to continue using the last ATV frame until TP video data is available at the selected TP speed.

When transitioning from "trick play" mode to normal play, the ATV signal 134 becomes available for display processing only after an I-frame has occurred in the GOP of the reproduced ATV signal. The occurrence of this I-frame depends on the rate of resynchronization of the SD recorder main guide axis servo and, more importantly, on the retrieval of the normal play speed in the recorded GOP sequence. Thus, various options can advantageously be provided for producing a favorable picture transition between "trick play" and normal playback. For example, when an end "trick play" command is issued, the last TP frame may be fixed and played back from memory until the ATV signal is reproduced. This method may indicate to the user that a control instruction has been received and executed. However, the fixed or still images, which are juxtaposed to the fast moving images generated in the TP, may be uncomfortable for the user. An alternative to the transition from "trick play" may be provided by continuing to reproduce the TP data and display the TP image for the duration of the servo resynchronization and the acquisition of the I-frame of the ATV signal. With this option, the redundancy characteristic of TP data can be utilized during the period when the tape speed changes due to servo resynchronization and during the period when the ATVI frame is waiting to appear. During the tape speed change period, some of the TP data may not be recovered despite its redundant nature, however, such errors may be concealed by repeating some TP image frames from memory. This advantageous method provides the user with a visual indication that the VCR has responded to its command because: the speed of the TP picture will change visibly as the capstan is resynchronized as it slowly enters the play speed. This feature may also allow for slower tape speed transitions to be used, thereby providing more smooth control without substantial damage to the tape, since acceleration or deceleration of the tape will be accompanied by acceleration or deceleration of the "trick play" image.

Claims (16)

1. A method for reproducing digital image signals in a helical scan digital tape recorder, comprising the steps of:

a) reproducing from each track recorded a first data signal representing said digital image signal and a second data signal representing a processed version of said digital image signal;

b) disregarding one of said first and second data signals; and

c) the other of the first and second data signals is coupled as a signal for additional processing.

2. A reproduction method according to claim 1, comprising the step of ignoring said second data signal during a play operation at a recording speed.

3. A reproducing method according to claim 1, comprising the step of ignoring said first data signal during a play operation at any speed other than a recorded speed.

4. A reproduction method according to claim 1, comprising the step of ignoring said first data signal during play operations in a trick play mode.

5. A reproducing method according to claim 1, comprising the step of selectively reproducing said first and second data signals from said recorded tape at a plurality of tape speeds.

6. The reproduction method according to claim 4, wherein said trick play mode includes a plurality of reproduction speeds.

7. A reproducing method according to claim 1, comprising the step of reproducing said second data signal from said recorded tape in a direction opposite to the tape motion.

8. A reproduction method according to claim 1, comprising disregarding said first data signal corresponding to trick play mode operation in a direction opposite to the tape motion.

9. A digital video cassette recorder for reproducing a digital image signal from a recorded magnetic tape at a plurality of reproduction speeds, said digital video cassette recorder comprising:

a playhead for sensing a combined digital signal from each track of the tape at one of a plurality of tape speeds of the plurality of playback speeds, the combined digital signal having first and second data signal components;

means for splitting said first and second data signal components; and

means for selecting said first data signal component as an output signal in a first play mode and said second data signal component as an output signal during a second play mode.

10. The digital video cassette recorder as claimed in claim 9, wherein said first and second data signal components represent MPEG compatible image signals.

11. The digital video cassette recorder as set forth in claim 9, wherein said first data signal component represents said digital image having a first data content.

12. The digital video cassette recorder as set forth in claim 11 wherein said second data signal component also represents said digital image but has less data content than said data signal component.

13. The digital video cassette recorder as set forth in claim 9, wherein said second data signal corresponds to said digital image having a reduced resolution.

14. The digital video cassette recorder as claimed in claim 10, wherein said second data signal component represents only MPEG compatible I-frame image signals.

15. A digital video recorder for recording encoded digital signals representing a digital image signal, said digital video recorder comprising:

means for receiving said encoded digital image signal;

means for decoding said encoded digital image signals and for coupling to said receiving means and generating image frames;

means, coupled to said receiving means, for reconstructing said digital image signals for audio recording;

trick play signal generating means coupled to said decoding means for compressing said image frames; and

a combiner having an output coupled to said reconstruction means and a plurality of inputs coupled to said trick play signal generation means, said combiner selecting between said inputs and generating an output signal for recording.

16. The digital video recorder of claim 15, wherein said combiner output signal includes said trick play signal at specific time intervals and said reconstructed digital image signal at other times.