JPH03224380A - Recording and reproducing system for dct compressed moving picture data - Google Patents

Abstract

PURPOSE:To realize data compression effectively by providing a freezing special reproduction means outputting a zero prediction error signal in place of a prediction error signal according to a frame command commanding a still picture of a display pattern inputted from an interactive device and outputting a pseudo control code displaying the implementation of an inter-frame prediction coding in place of a control code to a reproduction means. CONSTITUTION:Each frame forming a moving picture data fed from an input terminal IN to a prediction coding section 11 is divided equally into plural blocks. That is, each frame as shown in figure is divided equally into 1/m and 1/n respectively in row and column directions to generate mn-set of blocks from B11 to Bmn. Each block consists of 8X8 element data arranged in row and column directions. A recording data composing section 16 reads a variable length quantization conversion coefficient from a buffer memory 15 and applies it to an optical disk recording section 17 as a DCT compressed moving picture data in a prescribed format while a prescribed index and a header or the like are added to the coefficient.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a DCT compressed video data recording/playback device used as an interactive video recording/playback system for educational or entertainment purposes.

(Prior Art) Interactive video recording and playback systems for educational and entertainment purposes are currently being developed. In this system, as in the case of television signals, a series of moving images is processed as a collection of screens called frames that are discretely arranged at predetermined intervals on the time axis.

In such a moving image recording/playback system, highly efficient data compression is required in order to reduce the recording capacity.

In addition, in order to increase the convenience of an interactive playback system, a data format suitable for various special playbacks such as static playback and reverse playback is required.

"Image compression recording system" related to the applicant's earlier application
(Japanese Patent Application No. 62-108350), between frames/
A compression recording method is disclosed that uses a combination of intraframe predictive coding and quantization to improve the efficiency of data compression and facilitate reverse playback.

(Issue 8 to be solved by the invention) In the image compression recording system related to the above-mentioned prior application, a considerable degree of data compression is achieved by a combination of interframe/intraframe predictive coding and quantization. In terms of rate, it is still not sufficient.

Furthermore, although special playback such as reverse playback has been realized in the compression recording system for images related to forward order, it is necessary to further improve the usability as an interactive system by realizing special playback in a wide variety of layers.

(Means for Solving the Problems) According to the DCT compressed video data recording/playback method according to the present invention, a predetermined number of pixel data arranged in rows and columns for each of a plurality of continuous frames constituting a video. A prediction error signal within a block is created by dividing the block equally into multiple blocks including a group, performing interframe predictive coding or intraframe predictive coding for each block, and applying a discrete cosine to the prediction error signal within each block. Conversion, quantization, and variable-length encoding are performed to create variable-length quantized transform coefficients, and a control code indicating the compression method applied during this compression is variable-length encoded and added to the optical disk as DCT compressed video data. and a compression/recording means for recording on the optical disk, and performing inverse variable length encoding, inverse quantization, inverse discrete cosine transform, and inverse predictive encoding on variable length quantized transform coefficients included in the DCT compressed video data read from the optical disk. By including a reproducing means for simultaneously reproducing data and a display means, the present invention is configured to achieve further data compression than a conventional method using only predictive encoding, quantization, and variable length encoding.

Further, the reproduction means reproduces all zero transform coefficients instead of the actual transform coefficients, or reproduces a pseudo control code instead of the actual control code such as type information indicating the type of predictive encoding or motion vector. By providing special playback means for performing various functions, the system is configured to realize a wide variety of special playback functions in response to commands such as freeze, fade-out, mosaic playback, and scroll-out from the interactive device.

Hereinafter, the operation of the present invention will be explained in detail together with examples.

(Embodiment) FIG. 1 is a block diagram showing the configuration of a compression/recording system to which a DCT compressed video data recording/playback method according to an embodiment of the present invention is applied, where IN is a video to be compressed/recorded. Data input terminal, 11 is predictive coding section, 12 is discrete cosine transform (DCT) section, 13 is scalar quantization (S Q
) section, 14 is a variable length coding (VLC) section, 15 is a buffer memory, 16 is a recording data assembling section, and 17 is an optical disk recording section.

Each frame constituting the video data supplied from the input terminal IN to the predictive encoding unit 11 is equally divided into a plurality of blocks. That is, as shown in FIG. 2, mn blocks from Bll to Bmn are generated by equally dividing each frame into 1/m and 1/n in the row and column directions, respectively.

Each P7 is composed of 8×8 pixel data arranged in rows and columns.

Next, data compression by interframe predictive coding or intraframe predictive coding is performed for each block. Inter-frame predictive coding is performed by calculating the difference with the corresponding pixel data in the previous frame to generate an inter-frame prediction error signal, and intra-frame predictive coding is performed by calculating the difference with the corresponding pixel data in the frame. This is done by calculating and generating an intra-frame prediction error signal. This predictive coding is mainly based on interframe predictive coding, and in order to prevent deterioration of image quality due to the accumulation of error signals between the compression/recording system and the playback system, intraframe predictive coding is used at an appropriate frequency. A change to the current version (refresh) is performed. Further, when the correlation between frames is extremely reduced due to scene switching, a change to intra-frame predictive coding is adaptively performed.

As shown in FIG. 3, the prediction error signal e (i, j) generated by the above interframe/intraframe predictive coding is arranged in 8 rows and 8 columns within each block, similar to the pixel data before encoding. Arranged. This prediction error signal e (i, j
) is subjected to a discrete cosine transform in the next stage discrete cosine transform unit <DCT) 12, and as shown in FIG. 4, a group of transform coefficients E (uV ) is generated. The next stage of scalar quantization (
The SQ) unit 13 performs scalar quantization on this transform coefficient group E (u, V). The term scalar quantization (SQ) refers to a group of transform coefficients whose (u).

■) This is to distinguish it from vector quantization, which is two-dimensional quantization based on the distribution pattern in space.In this scalar quantization, as in the case of normal quantization, a predetermined value is used for each transform coefficient. Quantization is performed in the quantization step.

This quantized transform coefficient group is then subjected to variable length coding (
In the VLC unit 14, variable length coding is performed in which a shorter code is assigned to a code that appears more frequently, and is stored in the Halafah memory 15 as a variable length quantization conversion coefficient.

The recording data assembling unit 16 reads the variable-length quantized conversion coefficients from the buffer memory 15, adds a predetermined index, header, etc. to them, and converts them into DC data in a predetermined format.
The data is supplied to the optical disk recording section 17 as T-compressed moving image data. This header contains various information regarding the compression method applied during compression (type information indicating whether interframe or intraframe predictive coding was applied, quantization characteristics including quantization step width, etc.). A variable-length encoded code is included as a control code.

The DCT compressed video data recorded on this optical disc is
As illustrated in the data format in Figure 5, it consists of multiple scenes with indexes added to the beginning.
Each scene consists of multiple shots, and each shot consists of multiple frames and a null code. This shot is set to be an integral multiple of a 2-byte CD frame (named this way to distinguish it from a frame that represents one screen), which is the physical recording unit of an optical disc.

Due to variable length encoding, the data length of each frame varies, and as a result, an empty portion whose length is less than the data length of one subsequent frame occurs at the end of the shot. A null code, which is treated as invalid data, is inserted into this empty space.

One frame of data consists of multiple block data, each block data includes a block header added to the beginning, compressed pixel data following this, and an end of block (end of block) which is an end code indicating the end of the block. EOB
).

As shown in FIG. 6, the index added to the beginning of the DCT compressed video data consists of the data length, title, number of scenes of the entire DCT compressed video data, and a set of scene information regarding each scene. Each scene information consists of a scene number, scene title, number of shots, etc., and a set of shot information regarding each shot. Each shot information consists of a shot number, a storage start address, an image file name, an image title name, the number of frames, and a spare area for expansion. In addition, the picture header added to the beginning of each frame data includes the frame data length, frame type information indicating the predictive coding format for each frame, quantization characteristics, global motion vector indicating the motion vector for each frame, parity, Consists of a reserve area. Furthermore, the block header added to the beginning of each block data is based on block attributes such as the block address, block type information indicating the predictive coding format for each block, scan class, and differential motion vector indicating the motion vector of the block terminal. configured.

FIG. 7 shows recording and recording of DCT compressed video data in the above embodiment.
It is a block diagram showing the configuration of a playback device to which the playback method is applied, in which 21 is an optical disk playback section, 22 is a buffer memory, 23 is an inverse variable length coding section (VLD), and 24 is an inverse quantization (Q-'). part, 25 is the inverse discrete cosine transform (D CT'
-') section, 26 is an inverse predictive encoding section, 27 is a display section, 28
2 is a main control unit, and 29 is an interaction device.

When the user issues a command to start playback from the dialogue device 29, the main control unit 28 receives this command and controls the optical disc playback unit 21.
starts the playback operation. The compressed video data played back by the optical disk playback unit 21 is transferred to the SC3I interface unit 3.
5. The data is written to the buffer memory IJ22 via the bus 36.5C3I interface section 37 and the common bus 38 under the control of the write control section 31. The compressed image data read out from the buffer memory 22 under the control of the readout control unit 32 is supplied to the inverse variable length encoding unit 23 via the switch 39, where it is converted into fixed length quantized transform coefficients and control codes. is decrypted.

The quantized transform coefficients decoded by the inverse variable length encoder 23 are supplied to the inverse quantizer 24. On the other hand, control codes such as quantization characteristics, frame type information, block type information, and motion vectors included in the picture header and block attributes decoded to a fixed length are sent to the subsequent inverse quantization unit 24 and inverse predictive encoding unit 26. Supplied.

The inverse quantization section 24 also converts the fixed length quantized transform coefficients supplied from the inverse variable length encoding section 23 into the inverse variable length encoding section 2.
3 to transform coefficients according to the quantization characteristics supplied as a control code. , is supplied to the inverse discrete cosine transform unit 25. The inverse discrete cosine transform section 25 restores the transform coefficients supplied from the inverse quantization section 24 into a prediction error signal in real space, and supplies it to the inverse predictive coding section 26 . Inverse predictive encoding unit 2
6 performs inter-frame or intra-frame processing on the pixel data supplied from the inverse discrete cosine transform unit 25 according to the type information indicating the predictive encoding format supplied from the inverse variable length encoding unit 23 as a part of the control code. By performing inverse predictive encoding, the data is restored to a group of pixel data before compression and is supplied to the display unit 27.

As shown in FIG. 8, the above-mentioned inverse predictive encoding unit 26 has an input terminal 11 for a prediction error signal and an input terminal 1 for a control code.
2.13, an adder 51, a frame memory 52, a delay device 53, and a data output terminal O.

The prediction error signal output from the inverse discrete cosine transform section 25 at the previous stage is supplied from the data input terminal 11 to one input terminal of the adder 51 . On the other hand, control codes such as frame type information and block type information indicating whether the predictive coding is interframe predictive coding or intraframe predictive coding are supplied to the input terminal I2, and according to this control code, the switch 54 is switched. If the prediction error signal supplied to one input terminal of the adder 51 is based on interframe predictive coding, the pixel data of the preceding frame output from the frame memo I 752 is sent to the other input terminal of the adder 51 via the switch 54. The pixel data is supplied to the input terminal, restored to the pixel data of the current frame by addition of both, and supplied to the display unit via the data output terminal 0, as well as to the frame memory 52 and the delay unit 53. On the other hand, if the prediction error signal supplied to one input terminal of the adder 51 is based on intraframe predictive coding, the preceding pixel data in the current frame output from the delay device 53 is is supplied to the other input terminal of the adder 51 through the addition of both, restored to the pixel data in the current frame, and supplied to the display section through the data output terminal 0, and is also supplied to the frame memory 52 and the delay device 53. is supplied to

As shown in the block diagram of FIG. 9, the inverse variable length encoding unit 23 in FIG.
3, decoding ROM 62a to 62n, sequencer 64, data output terminal 01, control code output terminal 02,
03. Since different variable length encoding is applied for each quantization transform coefficient and control code during compression and recording, decoding ROMs 62a to 62n are installed for each quantization transform coefficient and control code, and the data input terminal 4
The corresponding decoding ROM is selected by the selectors 61 and 63, which are switched in synchronization with the appearance of .

That is, variable length control codes such as variable length quantization conversion coefficients and motion vectors read from the hash sofa memory 22 in FIG. is supplied to one of the corresponding decoding ROMs 62a to 62n, decoded into fixed length data, and then passed through the selector 63, which is switched in synchronization with the selector 61, to the output terminal 01 of the quantized transform coefficients.
and one of the control code output terminals 02 and 03.

The operation of the sequencer 64 in the inverse variable length encoder 23 in FIG. 9 will be explained with reference to the flowchart in FIG. 10.

When the sequencer 64 starts operating, it waits for the appearance of a frame synchronization signal added to the beginning of the frame (step 71), and when it detects the appearance, it determines whether the next block to appear is the last block in the frame. If the block is determined to be the final block (step 72), the block address (BA) attached to the block is decoded and set as the jump block number (JB). This block address (BA) indicates the number of blocks that are skipped because the prediction error signal with the previous frame is zero in interframe predictive coding during compression and recording, and this is the number of blocks that are skipped because the prediction error signal with respect to the previous frame is zero. Appended to the block immediately following it. In the next step 74, it is determined whether or not JB is zero, and if it is zero, decoding to one block worth of quantized transform coefficients is performed in step 75, and step 7
A return to 2 is performed.

On the other hand, if it is determined in step 74 that JB is not zero, one block of W is decoded into quantized transform coefficients.
A timer for transitioning to the AIT state is started, and at the same time, insertion of zero data to replace the quantized transform coefficients is started. W regarding decoding to this quantized transform coefficient
Even under the AITa' condition, the presence or absence of a frame synchronization signal and the presence or absence of a timer timeout are determined in steps 77 and 78.
Detected in If a timeout occurs before the appearance of the frame synchronization signal, JB is decremented by 1 in step 79, a return is made to step 74, and the above-described operations are repeated.

Decoding ROMs 62a to 6 in the above-mentioned inverse variable length encoding unit 23
The quantized transform coefficient decoding ROM, which is one of the 2n
As shown in the figure, the data input terminal 6 and the main control unit 28
(Fig. 7), an input terminal I7 for receiving commands, a decoding ROM 81, 82, and 83 for normal playback, fade-out playback, and mosaic playback, and an in-block counter 84.
, a zero data insertion register 85, a selector 86, and a data output terminal O.

As shown in FIG.
Input terminals 19 and 11 for address and switching signals from (Fig.)
0, a decoding ROM 91 for normal playback, a decoding ROM 92 for scrolling out, and an in-block counter 93.
It is composed of a zero data insertion register 94, a selector 95, and a data output terminal 05. This motion vector C is a control code that indicates from which point in the previous frame each point in the current frame has moved during interframe predictive coding during compression in the recording system.

The operation of special playback will be explained below.

■, "Freeze" A. Freeze Part 1 When the main control unit 28 receives a freeze command from the dialogue device 29, it instructs the data insertion unit 30 to insert a pseudo control code and a pseudo block address for the freeze operation. At the same time, it instructs the inverse variable length encoding unit 23 to start a freeze operation. Upon receiving the above instruction, the data insertion unit 30 uniformly replaces the control code indicating whether interframe or intraframe predictive coding is included in the frame type information of the picture header in each frame by switching the switch 39. At the same time, a pseudo control code indicating that interframe predictive coding is used is output, and a value greater than the upper limit value is inserted in place of the block address (BA) of the first block in each frame. In addition, the inverse variable length encoding unit 23 (9th
When the sequencer 64 in Fig. 1 receives a freeze command from the input terminal I5, the motion vector decoding ROM (Fig. 12)
By switching the selector 95, the zero data being held in the zero data insertion register 94 is outputted instead of the output of the decoding ROM 91 for normal reproduction.

As a result, as shown in the flowchart of Figure 10,
The sequencer 64 in the inverse variable length encoding unit 23 outputs zero data for each frame instead of the quantized transform coefficients in all blocks, and also outputs a pseudo control code and zero data indicating that interframe predictive encoding is performed. The motion vector replaced with is output. As a result, the contents of the frame memory 52 shown in FIG. 8 are repeatedly output, and the display screen is frozen (still).

B. Freeze Part 2 When the main control unit 28 receives a freeze command from the dialogue device 29, it instructs the data insertion unit 30 to insert a pseudo control code indicating interframe predictive coding, and also The long encoder 23 is instructed to perform a freeze operation. Upon receiving this command, the data insertion unit 30 changes the picture in each frame by switching the switch 39, and inserts a control code indicating whether predictive coding is between frames or within a frame included in the frame type information of the frame. Instead, a pseudo control code uniformly indicating interframe predictive coding is output. On the other hand, the sequencer 6 in the inverse variable length encoder 23
4 receives the freeze command from the terminal I5, and by switching the selector 86 in the quantization conversion coefficient decoding ROM, outputs the output of the zero data insertion register 85 to the output terminal 04 instead of the output of the decoding ROM 81 for normal reproduction. Output. At the same time, the sequencer 64 switches the selector 95 of the motion vector decoding ROM (FIG. 12) to output the zero data held in the zero data insertion register 94 instead of the output of the normal reproduction decoding ROM 91.

As a result, the contents of the frame memory 52 shown in FIG. 8 are repeatedly output from the output terminal 0 to the subsequent display section, and the display screen is frozen (still).

(2) “Fade out” When the main control unit 28 receives a fade out command from the dialogue device 29, it instructs the data insertion unit 30 to insert a pseudo control code indicating interframe predictive coding, and also instructs the data insertion unit 30 to insert a pseudo control code indicating interframe predictive coding. The encoder 23 is instructed to perform a fade-out operation. Upon receiving this command, the data insertion unit 30 outputs a pseudo control code indicating interframe predictive coding by switching the switch 39 in place of the control code included in the frame type information of the picture header in each frame. . On the other hand, the sequencer 64 in the inverse variable length encoding unit 23
When receiving the freeze command from the terminal I5, the output terminal outputs the output of the decoding ROM 82 for fade-out instead of the output of the decoding ROM 81 for normal reproduction by switching the selector 86 in the quantization conversion coefficient decoding ROM (Fig. 11). 04 to output.

As shown in FIG. 13, this fade-out decoding ROM 82 decodes only the DC component (u = v = 0) to "-1" among the transform coefficients by discrete cosine transform, and decodes all other frequency components. A value is set to decode it to "0". As a result, the frame memory 52 in FIG.
Only the DC value of each block is decreased by 1 in each frame period, and a fade-out is performed in which the brightness of the display screen monotonically decreases.

(2) "Mosaic Reproduction" When the main control section 28 receives a mosaic reproduction instruction from the dialog device 29, it instructs the inverse variable length encoding section 23 to perform a mosaic reproduction operation. Sequencer 64 in the inverse variable length encoder 23
receives a mosaic play command from input terminal I5,
By switching the selector 86 in the quantization transform coefficient decoding ROM (Fig. 11), the normal reproduction decoding ROM 8
The output of the mosaic reproduction decoding ROM 83 is passed through the output terminal instead of the output of the mosaic reproduction decoding ROM 83.

This mosaic playback decoding ROM 83 contains information shown in Fig. 14 (A).
), among the transform coefficients of the discrete cosine transform, only the spatial frequency DC component (u = v = 0) is decoded as it is during normal reproduction (at a magnification of 1), and all other frequency components are rOJ The value to be decoded (at a magnification of 0) is set in . As a result, a mosaiced screen containing only the direct current component of the spatial frequency is displayed.

FIG. 14(B) is a conceptual diagram showing another example of the decoding ROM 83 for mosaic reproduction. In this example, the spatial frequency DC component (u=
v=0) as is during normal playback, and
For the next lowest frequency component, 1/2 of that for normal playback.
, and for all frequency components higher than this,
0" is set as the value to be decoded. As a result, a mosaiced screen containing only low spatial frequency components is displayed.

FIG. 14(C) is a conceptual diagram showing still another example of the decoding ROM 83 for mosaic reproduction. In this example, the spatial frequency DC component (
In addition to decoding only U=V=O) as is, the next lowest frequency component is decoded at 1/2 of the normal reproduction rate, the next lowest frequency component is decoded at 1/4 of the normal reproduction rate, and all higher frequency components are Regarding the frequency component, a value to be decoded is set in rOJ. As a result, a mosaiced screen containing only low spatial frequency components is displayed.

2. "Scroll out" This scroll out is performed using a motion vector, which is one of the control codes.

When the main control unit 28 receives a scroll-out instruction from the dialogue device 29, it instructs the data insertion unit 30 to insert a pseudo control code indicating interframe predictive coding, and also instructs the data insertion unit 30 to insert a pseudo control code indicating interframe predictive coding. commands a scroll-out operation. Upon receiving this command, the data insertion unit 30 outputs a pseudo control code indicating interframe predictive coding by switching the switch 39 in place of the control code included in the frame type information of the picture header in each frame. .

On the other hand, when the sequencer 64 in the variable-length compressed image data decoding section 23 receives the command for sfloral 1 from the terminal I5, it switches the selector 86 in the quantization transform coefficient decoding ROM (FIG. 11) for normal reproduction. Decoding R
Zero data insertion register 85 in place of the output of OM81
The zero data being held is outputted to the output terminal 04. At the same time, the sequencer 64 performs motion vector decoding RO
By switching the selector 95 in M (FIG. 12), the output of the scroll-out decoding ROM 92 is passed to the output terminal 05 instead of the output of the normal reproduction decoding ROM 91.

That is, as illustrated in FIG. 15(A), during normal playback, the motion vector is a value (Vx, Vy) corresponding to the movement within the display screen, and the quantization transform coefficient corresponds to the prediction error signal between frames. The value Q1. Q2. Q3...
It is. On the other hand, when scrolling out, as illustrated in FIG. 1), and the quantization transform coefficients are all zero. In this example, the display screen moves downward (-Y direction) at a frame period and at a predetermined speed corresponding to the absolute value 1 of the motion vector, and scrolls out downward. The direction and speed of this scroll out are determined by (V x, V y
) is specified by a combination of the polarity and absolute value of the motion vector to be replaced, and is specified by a part of the address supplied from the input terminal I9 to the scroll-out ROM 92.

The motion vector output from the motion vector decoding ROM shown in FIG. 12 is supplied to the frame memory 52 via the control code input terminal 3 in the inverse predictive encoding section 26 shown in FIG. In the frame memo IJ52, the pixel data of the preceding frame is read out at a timing shifted in the X and Y directions by a time corresponding to the direction and magnitude of the motion vector, and the pixel data of the preceding frame is displayed on the display screen of the current frame. Pixels that have moved from the frame are displayed.

As described above, a decoding ROM for special playback is installed in the inverse variable length encoding section, and a decoding ROM for normal playback is performed according to the command for special playback.
An example of a configuration for switching with OM is illustrated.

However, in the special playback operation that generates a zero prediction error signal, the decoding ROM that outputs this zero data
It is also possible to have a configuration in which the ROM is installed in the inverse quantization section or the inverse discrete cosine transform section, and the decoding ROM for special playback is switched to in accordance with the command for special playback.

Furthermore, a configuration in which a data insertion section for inserting a pseudo control code is installed before the inverse variable length encoding section is illustrated. However, instead of installing this data insertion section, a decoding ROM for outputting a pseudo control code may be installed in the inverse variable length encoding section.

As explained in detail above, according to the recording/playback method of DCT compressed video data according to the present invention, in addition to the conventional compression method using predictive encoding, quantization, and variable length encoding,
Since the configuration is such that hybrid compression is performed by combining the T compression methods, -layer data compression can be realized.

Further, according to the recording/reproducing method of the present invention, the features of interframe predictive coding and DCT compression method are used to freeze,
Because it is configured to realize a wide variety of special playback functions such as fade out, mosaic playback, and scroll out,
The convenience of the interactive playback device is greatly improved.

[Brief explanation of drawings]

Fig. 1 is a block diagram showing the configuration of a compression/recording system to which a DCT compressed video data recording/playback method according to an embodiment of the present invention is applied, and Fig. 2 explains a block division method using the compression/recording system. Figure 3 is a conceptual diagram illustrating the arrangement of prediction error signals in a block, and Figure 4 shows the arrangement in frequency space of transform coefficients obtained by discrete cosine transform of the prediction error signal in a block. FIG. 5 is a conceptual diagram illustrating the data format of DCT compressed video data recorded on an optical disc. FIG. 6 is a conceptual diagram illustrating the structure of the index, picture header, and block header in FIG. 5. Figure 7 shows the DC according to the above embodiment.
FIG. 8 is a block diagram illustrating the configuration of the reproducing system for T-compressed video data, FIG. 8 is a block diagram illustrating the configuration of the inverse predictive encoding unit (26) in FIG. 1, and FIG. 9 is the inverse variable length code in FIG. 1. (VL)
D) A block diagram illustrating the configuration of the unit (23), FIG. 10 is a flowchart for explaining the operation of the sequencer (64) in FIG. 9, and FIG.
Figure 12 is a block diagram showing the configuration of the quantized transform coefficient decoding ROM (62n) of the decoding ROM (62n) in Figure 9.
2a to 62n), FIG. 13 is a conceptual diagram for explaining the configuration of the fade-out decoding ROM 82 in FIG. 11,
FIG. 14 is a conceptual diagram for explaining the configuration of the mosaic reproduction decoding ROM 83 shown in FIG. 11, and FIG. 15 is a data format diagram for explaining the scroll-out operation using motion vectors. 11... Predictive coding unit, 12... Discrete cosine transform (DCT) unit, 13... Scalar quantization (SQ) unit, 14... Variable length coding (V L Ci unit, 15...・
・Buffer memory, 16...Record data assembly section, 1
7... Optical disc recording unit, 21... Optical disc reproducing unit, 22... Buffer memory, 23... Inverse variable length coding (VLD) unit, 24... Inverse quantization (Q-' ) part,
25... Inverse discrete cosine transform (DCT-') section, 26
... Inverse predictive encoding unit, 27... Display unit, 51...
Adder, 52...Frame memory, 53...Delay unit for one pixel, 62a-62b...Decoding R of variable length code
OM, 64... Sequencer, 81... Quantized transform coefficient decoding ROM for normal playback, 82... Quantized transform coefficient decoding ROM for fade-out, 83... Quantized transform coefficient decoding ROM for mosaic playback, 85 . . . Register for zero data insertion, 91 . . . Motion vector decoding ROM for normal reproduction, 92 . . . Motion vector decoding ROM for scroll out, 94 .

Claims (5)

[Claims] (1) Each of a plurality of consecutive frames constituting a video is divided equally into a plurality of blocks containing a predetermined number of pixel data groups arranged in rows and columns, and each block is divided into corresponding pixel data in the preceding frame. Block prediction is performed by performing interframe predictive coding to obtain an interframe prediction error signal by calculating the difference between the previous pixel and the previous pixel in the frame, or intraframe predictive coding to obtain an intraframe prediction error signal by calculating the difference from the previous pixel within the frame. Create a prediction error signal in each block, apply discrete cosine transform to the prediction error signal in each block to obtain a transform coefficient, quantize this to obtain a quantized transform coefficient, and variable-length encode this to obtain a variable-length quantum At the same time, a control code indicating the compression method applied during the compression is variable-length encoded to obtain a variable-length control code, which is added to a variable-length quantization conversion coefficient and then recorded on an optical disc as DCT compressed video data. a compression/recording means for recording the DCT compressed video data read from the optical disk; and a variable length quantized transform coefficient included in the DCT compressed video data read from the optical disk, which is inversely variable length encoded to obtain a quantized transform coefficient, and which is dequantized to obtain a transform coefficient. obtained,
A reproduction means for performing inverse discrete cosine transform on this to obtain a prediction error signal and inverse predictive encoding to reproduce a pixel data group; and a display means for displaying the reproduced pixel data group, the reproduction means outputs a zero prediction error signal in place of the prediction error signal in accordance with a freeze command input from the dialogue device to freeze the display screen, and performs interframe predictive coding in place of the control code. 1. A method for recording and reproducing DCT compressed video data, characterized by comprising a special reproduction means for freezing that outputs a pseudo control code for displaying a message. (2) During the compression/recording, blocks whose interframe prediction error signal is zero are skipped from the subsequent compression/recording targets, and the consecutive number of skipped blocks is used as the block address as a control code for the block immediately after this skipping ends. and for the consecutive number of blocks included in the block address and displayed at the block address during the reproduction, an inter-frame prediction error signal of zero is reproduced, and the freeze special reproduction means reproduces the blocks according to the freeze command. The DCT according to claim 1, characterized in that means for inserting a value equal to or greater than the upper limit value as pseudo data in place of an address is provided at a stage upstream of the means for outputting the zero interframe prediction error signal. A recording/playback method for compressed video data. (3) In accordance with a fade-out command inputted from the dialogue device to fade out the display screen, the reproduction means replaces the conversion coefficients in each block with a predetermined value of negative polarity for only the DC component and zero for the AC component. The DC according to claim 1 or 2, characterized in that the DC is equipped with a fade-out special reproduction means that outputs a certain conversion coefficient.
A recording/playback method for T-compressed video data. (4) The reproduction means performs mosaic reproduction for outputting transformation coefficients whose alternating current components are suppressed in place of the transformation coefficients in each block in accordance with a mosaicization command that instructs mosaicization of the display screen input from the dialogue device. 4. A recording and reproducing method for DCT compressed moving image DCT according to any one of claims 1 to 3, characterized in that the DCT compressed moving image DCT is equipped with a special reproduction means. (5) The compression/recording means includes in the control code a motion vector indicating from which point in the previous frame each point in the current frame has moved during interframe predictive encoding, and during the reproduction, the motion vector is included in the control code. The reproducing means reproduces a pseudo motion vector indicating movement in a predetermined direction instead of the motion vector in accordance with a command to scroll out the display screen inputted from the dialogue device. , a DCT compressed video DCT according to any one of claims 1 to 4, further comprising scroll-out special reproduction means for reproducing a zero interframe prediction error signal instead of an interframe prediction error signal. recording/playback method.