JP2002171524A - Data processing apparatus and method - Google Patents

Abstract

(57) [Summary] (With correction) [PROBLEMS] To enable stable operation even when an invalid code that does not exist in the VLC table is input when decoding a variable length code. SOLUTION: A variable-length-encoded stream of DC
If a VLC table mismatch occurs in the header section before the T block (S11), the header section is replaced with a default value, the DCT block is replaced with a DC component block that performs gray display, an EOB is added, and a macro block is added. Is terminated (S12). If the VLC table mismatch does not occur in the header portion but occurs in the DC coefficient portion (S14), the DCT block of the DC component is replaced with a block that performs gray display, an EOB is added, and the macro block is aborted (S15). ). If the VLC table mismatch does not occur in the DC coefficient part but occurs in the AC coefficient part (S16), EOB is added from that position and the macroblock is aborted (S17).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is capable of performing stable processing when decoding a variable length code with reference to a table, even if an invalid variable length code not present in the table is input. Such a data processing apparatus and method.

[0002]

2. Description of the Related Art In recent years, MPEG (Moving Pictures Exposure) has been used as a method for compressing and encoding digital video signals.
erts Group) is widely used. MPEG2 uses DCT (Discrete Cosine Transfor
m) and a video compression standard using predictive coding. DCT obtained by dividing one frame of image data into macroblocks of a predetermined size, performing predictive coding using a motion vector in macroblock units, and further dividing the macroblock.
DCT is performed for each block, and variable-length coding is performed.
At present, MPEG2, which has higher expandability and can obtain high image quality, is mainly used.

[0003] MPEG2 data is composed of a data stream having a hierarchical structure. The layers are a sequence layer, a GOP (Group Of Picture) layer, a picture layer, a slice layer, and a macroblock layer from the top, and each layer includes one or more lower structures. Each layer has a header part. In addition, each layer except the macroblock layer includes
A start code is provided prior to the header.

A macro block is a block composed of 16 pixels × 16 pixels, and one or more macro blocks constitute one slice. On the other hand, one picture corresponds to one screen, and a slice cannot cross pictures. In addition, the slice header always comes at the left end of the screen. The slice start code includes vertical position information of the slice, and the slice header stores extended slice vertical position information, quantization scale information, and the like.

In recent years, for example, in a broadcasting station or the like, transmission of a digital video signal between devices or recording of a digital video signal on a recording medium such as a magnetic tape is performed using the digital video signal as an MPEG stream. A system that does this has emerged. In such a system, for example, an MPEG elementary stream is transmitted on an SDTI (Serial Data Transport Interface), and is transmitted from the SDTI to the MPE on the receiving side.
The G elementary stream is extracted, and a predetermined process is performed.

[0006]

By the way, MPEG2
Is encoded using variable length codes (hereinafter, VLC: Variable Length Codes) as described above. The variable length coding is performed by assigning a code length according to the frequency of occurrence of the data based on a predetermined table (hereinafter, referred to as a VLC table). At the time of decoding, the VLC table is referenced to decode the VLC.

[0007] When an error occurs, the VLC
Unreliable until the next header (start code) is detected. This will be described with reference to FIG. FIG. 51A shows a normal stream. MPEG
Is a stream having a data width of 8 bits (1 byte) as described above. 32 bits (4 bytes) indicating the start of the slice layer in the data string 400
Start code (slice_sta)
rt_code) and a 5-bit quantizer_sca which is a parameter of the slice header.
le_code, extra_bit consisting of 1 bit
_Slice, macro_address_increm of a variable-length code that is a parameter of the macroblock layer
ent and macroblock_type are arranged, followed by predetermined parameters.

Slice_star in the data string 400
some DC in the slice layer indicated by t_code
In the data sequence 401 of the T block, in the example of FIG. 51A, the VLC is a predetermined VLC table (for example, DCT Coefficients Tab specified in MPEG).
Based on le 1), each line has the following meaningful stream.

... 0100 0000 — 0001 — 0010 — 0 0001 — 010 0010 — 0110 — 1 100.

It is assumed that bit inversion occurs in such a stream at the position of the hatched portion, and an error occurs in the position of FIG. 51B. Then, the stream changes as shown in FIG. 51C. When this is applied to the above-mentioned predetermined VLC table, the following is obtained.

... 0100 0000 — 0101 — 0010 — 0000 — 1010 —
0010 0110 1100 ・ ・ ・ ・ ・

In the changed stream, the VLC table (DCT Coefficients) described above is used.
Based on Table 1), the second line is Escape
(0000_01) + run 18 (01_0010)
+ Level 162 (0000_1010_001
0), and the third line means EOB (End Of Block). As described above, the VLC may be solved as a VLC completely different from the original stream by bit inversion of only one bit. Therefore, after the VLC in which an error has occurred, even if the VLC is unreliable, it is not reliable, and data up to the next start code 402 needs to be discarded. Since the start code is a 32-bit unique code byte-aligned as shown in FIG. 51, the start code can be detected and can recover from an error.

Therefore, in an MPEG decoder for decoding an MPEG stream, an illegal MP such as a bit inversion or a broken stream may occur.
When an EG stream is input, there is a problem that a failure such as a hang-up may occur due to a VLC that does not exist in the syntax defined by MPEG, that is, does not exist in the VLC table referred to during decoding. Was. As a result, there is a problem that it is difficult to realize a stable system.

Therefore, an object of the present invention is to provide an MPEG
It is an object of the present invention to provide a data processing apparatus and method that can operate stably even when an invalid VLC that is not present in a VLC table referred to when decoding a VLC is input in a system that handles a stream.

[0015]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention provides a data processing apparatus for processing a stream which has been subjected to variable-length encoding. Detecting means for detecting a code that does not correspond to the parameter representing the variable-length code, correcting means for correcting the stream according to the content of the detection result by the detecting means, and instructing to start the next processing after detection by the detecting means A data processing device comprising:

According to the present invention, there is provided a data processing method for processing a variable-length coded stream, wherein a variable-length code of the variable-length coded stream does not match a parameter representing the variable-length code. Detecting a stream, correcting the stream according to the content of the detection result in the detection step, and instructing to start the next process after the detection in the detection step. This is a data processing method characterized by the following.

As described above, the present invention detects a code that does not correspond to a parameter representing a variable-length code from a variable-length code of a stream that has been subjected to variable-length coding in a predetermined block unit, and Since the stream is modified accordingly and the start of the next process is instructed after the detection, even if a stream including an irregular variable length code is input, it can operate stably.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described with reference to a digital VT.
An embodiment applied to R (Video Tape Recorder) will be described. This embodiment is suitable for use in a broadcast station environment.

In this embodiment, the compression method is as follows.
For example, the MPEG2 system is adopted. MPEG2 is a combination of motion-compensated predictive coding and DCT-based compression coding. The data structure of MPEG2 has a hierarchical structure. FIG. 1 schematically shows a hierarchical structure of a general MPEG2 data stream. As shown in FIG. 1, the data structure includes a macroblock layer (FIG. 1E), a slice layer (FIG. 1D), and a picture layer (FIG.
C), GOP layer (FIG. 1B) and sequence layer (FIG. 1)
A).

As shown in FIG. 1E, the macroblock layer is composed of DCT blocks, which are units for performing DCT. The macro block layer includes a macro block header and a plurality of DCT blocks. Fig. 1
As shown in D, it is composed of a slice header section and one or more macroblocks. As shown in FIG. 1C, the picture layer includes a picture header section and one or more slices. A picture corresponds to one screen.
As shown in FIG. 1B, the GOP layer includes a GOP header, I pictures that are pictures based on intra-frame coding, and P and B pictures that are pictures based on predictive coding.

An I-picture (Intra-coded picture) uses information that is closed only in one picture when it is encoded. Therefore, at the time of decoding, decoding can be performed using only the information of the I picture itself. A P-picture (Predictive-coded picture: a forward predictive coded picture) uses a previously decoded I-picture or P-picture which is temporally previous as a predicted picture (a reference picture for taking a difference). . Whether to encode the difference from the motion-compensated predicted image, to encode without taking the difference,
The more efficient one is selected for each macroblock. A B picture (Bidirectionally predictive-coded picture) is a temporally previous I-picture or P-picture which is temporally preceding, and a temporally backward I-picture, We use three types of I-pictures or P-pictures already decoded, as well as interpolated pictures made from both. Among the three types of difference coding after motion compensation and intra coding, the most efficient one is selected for each macroblock.

Therefore, as the macroblock type,
Intra-frame coding (Intra) macroblocks, predicting the future from the past (Forward) Interframe predicting macroblocks, and predicting the future from the future (Backward)
There are an inter-frame prediction macro block and a bi-directional macro block predicted from both forward and backward directions. All macroblocks in an I picture are intra-coded macroblocks. The P picture includes an intra-frame coded macro block and a forward inter-frame predicted macro block. The B picture includes all four types of macroblocks described above.

A GOP includes at least one I picture, and P and B pictures are allowed even if they do not exist. As shown in FIG. 1A, the uppermost sequence layer includes a sequence header section and a plurality of GOPs.

In the MPEG format, a slice is one variable-length code sequence. A variable-length code sequence is a sequence from which a data boundary cannot be detected unless the variable-length code is correctly decoded.

At the head of the sequence layer, GOP layer, picture layer and slice layer, a start code having a predetermined bit pattern arranged in byte units is arranged. The start code arranged at the top of each layer is called a sequence header code in the sequence layer and a start code in other layers, and the bit pattern is [00 00 01 xx] (hereinafter []).
Is a hexadecimal notation). Two digits are shown, and [xx] indicates that a different bit pattern is arranged in each of the layers.

That is, the start code and the sequence header code are composed of 4 bytes (= 32 bits), and can identify the type of subsequent information based on the value of the 4th byte. Since these start codes and sequence header codes are arranged in byte units, they can be captured only by performing 4-byte pattern matching.

Further, the upper 4 bits of 1 byte following the start code are identifiers of the contents of the extended data area described later. The content of the extension data can be determined from the value of the identifier.

It should be noted that such an identification code having a predetermined bit pattern aligned in byte units is not allocated to the macro block layer and the DCT block in the macro block.

The header section of each layer will be described in more detail. In the sequence layer shown in FIG. 1A, a sequence header 2 is arranged at the head, and a sequence extension 3, an extension and user data 4 are arranged subsequently. Sequence header 2
Is arranged at the head of the sequence header code 1. Although not shown, a predetermined start code is also arranged at the head of each of the sequence extension 3 and the user data 4. From the sequence header 2 to the extension and user data 4 are used as the header part of the sequence layer.

As shown in FIG. 2, the sequence header 2 includes a sequence header code 1, an encoded image size including the number of horizontal pixels and the number of vertical lines, an aspect ratio, a frame rate, and a bit. Information set in units of a sequence, such as a rate, a VBV (Video Buffering Verifier) buffer size, and a quantization matrix, is assigned a predetermined number of bits and stored.

In sequence extension 3 after the extension start code following the sequence header, as shown in FIG.
Additional data such as a profile, a level, a chroma (color difference) format, and a progressive sequence used in MPEG2 are specified. Extension and user data 4
As shown in FIG. 4, the sequence display () can store the information of the RGB conversion characteristics and the display image size of the original signal, and the sequence scalable extension () specifies the scalability mode and the scalability layer. be able to.

Following the header of the sequence layer, GOP
Is arranged. At the head of the GOP, a GOP header 6 and user data 7 are arranged as shown in FIG. 1B.
The GOP header 6 and the user data 7 are used as a GOP header. As shown in FIG. 5, the GOP header 6 includes a GOP start code 5, a time code, a GOP
Flags indicating the independence and the validity of each are assigned a predetermined number of bits and stored. As shown in FIG. 6, the user data 7 includes extension data and user data. Although not shown, a predetermined start code is arranged at the head of each of the extension data and the user data.

Following the header section of the GOP layer, a picture is arranged. As shown in FIG. 1C, a picture header 9, a picture coding extension 10, and extension and user data 11 are arranged at the head of the picture. At the head of the picture header 9, a picture start code 8 is arranged. A predetermined start code is provided at the head of the picture coding extension 10 and the extension and user data 11, respectively. The part from the picture header 9 to the extension and user data 11 is the header part of the picture.

In the picture header 9, as shown in FIG. 7, a picture start code 8 is arranged, and coding conditions for a picture are set. In the picture coding extension 10, as shown in FIG. 8, the range of the motion vector in the front-back direction and the horizontal / vertical direction and the picture structure are specified. In the picture coding extension 10, the setting of the DC coefficient accuracy of the intra macroblock,
Selection of an LC type, selection of a linear / non-linear quantization scale, selection of a scanning method in DCT, and the like are performed.

In the extension and user data 11, as shown in FIG. 9, setting of a quantization matrix, setting of a spatial scalable parameter, and the like are performed. These settings can be made for each picture, and encoding can be performed according to the characteristics of each screen. Further, in the extension and user data 11, it is possible to set the display area of the picture. Further, copyright information can be set in the extension and user data 11.

Following the header part of the picture layer, slices are arranged. At the head of the slice, as shown in FIG. 1D, a slice header 13 is arranged, and the slice head 1
3, a slice start code 12 is provided.
As shown in FIG.
Contains vertical position information of the slice. The slice header 13 further stores extended slice vertical position information, quantization scale information, and the like.

Following the header portion of the slice layer, a macro block is arranged (FIG. 1E). In the macro block, a plurality of DCT blocks are arranged following the macro block header 14. As described above, the macroblock header 1
4 has no start code. As shown in FIG. 11, the macroblock header 14 stores the relative position information of the macroblock, and instructs the setting of the motion compensation mode, the detailed setting related to the DCT coding, and the like.

Following the macroblock header 14, DC
A T block is provided. As shown in FIG. 12, the DCT block includes a variable-length coded DCT coefficient and D
Data on CT coefficients is stored.

In FIG. 1, a solid line delimiter in each layer indicates that data is aligned in byte units, and a dotted line delimiter indicates that data is not aligned in byte units. That is, up to the picture layer, FIG.
As shown in an example in FIG. 3A, the boundaries of the codes are separated in units of bytes, whereas in the slice layer, only the slice start code 12 is separated in units of bytes, and each macroblock is shown in FIG. 13B. As an example is shown, it can be divided on a bit-by-bit basis. Similarly, in the macroblock layer, each DCT block can be divided in bit units.

On the other hand, in order to avoid signal deterioration due to decoding and encoding, it is desirable to edit the encoded data. At this time, the P picture and the B picture require a temporally preceding picture or a preceding and succeeding picture for decoding. Therefore, the editing unit cannot be set to one frame unit. In consideration of this point, in this embodiment, one GOP is made up of one I picture.

For example, a recording area in which recording data for one frame is recorded is a predetermined area. MPEG2
Since the variable length coding is used, the amount of generated data for one frame is controlled so that data generated during one frame period can be recorded in a predetermined recording area. Further, in this embodiment, one slice is composed of one macroblock so as to be suitable for recording on a magnetic tape, and one macroblock is applied to a fixed frame having a predetermined length.

FIG. 14 shows the MPE in this embodiment.
The header of the G stream is specifically shown. As can be seen from FIG. 1, the respective header portions of the sequence layer, the GOP layer, the picture layer, the slice layer, and the macroblock layer appear continuously from the head of the sequence layer. FIG. 14 shows an example of a data array that is continuous from the sequence header portion.

From the beginning, a sequence header 2 having a length of 12 bytes is arranged, followed by a sequence extension 3 having a length of 10 bytes. Subsequent to the sequence extension 3, extension and user data 4 are arranged.
A 4-byte user data start code is arranged at the head of the extension and user data 4, and information based on the SMPTE standard is stored in the subsequent user data area.

The header part of the sequence layer is followed by the header part of the GOP layer. A GOP header 6 having a length of 8 bytes is arranged, followed by extension and user data 7. At the beginning of the extension and user data 7, a 4-byte user data start code is arranged, and in the subsequent user data area, information for ensuring compatibility with other existing video formats is stored.

The header part of the GOP layer is followed by the header part of the picture layer. A picture header 9 having a length of 9 bytes is arranged, followed by a picture coding extension 10 having a length of 9 bytes. After the picture coding extension 10, the extension and user data 11 are arranged. The extension and user data are stored in the first 133 bytes of the extension and user data 11, followed by a user data start code 15 having a length of 4 bytes. Following the user data start code 15, information for compatibility with another existing video format is stored. Further, a user data start code 16 is provided.
Data based on the TE standard is stored. Next to the header part of the picture layer is a slice.

The macro block will be described in more detail. The macro block included in the slice layer is a set of a plurality of DCT blocks, and the coded sequence of the DCT block is obtained by converting the sequence of quantized DCT coefficients into the number of consecutive 0 coefficients (run) and the non-zero sequence immediately after it. (Level) is variable-length coded as one unit. Macro blocks and DCT blocks within the macro block include:
Identification codes aligned in byte units are not added.

The macro block is composed of one screen (picture).
It is divided into a grid of 6 pixels × 16 lines. A slice is formed by connecting these macroblocks in the horizontal direction, for example. The last macroblock of the previous slice of a continuous slice and the first macroblock of the next slice are continuous, and it is not allowed to form a macroblock overlap between slices. When the size of the screen is determined, the number of macroblocks per screen is uniquely determined.

The numbers of macroblocks in the vertical and horizontal directions on the screen are mb_height and m, respectively.
Called b_width. The coordinates of the macroblock on the screen are the vertical position number of the macroblock, mb_row counted from 0 based on the upper end, and the horizontal position number of the macroblock, and mb_col counted from 0 based on the left end.
mn. To represent the position of the macroblock on the screen with one variable, macro
block_address is set to macroblock
_Address = mb_row × mb_width +
mb_column is defined in this way.

It is specified that the order of slices and macroblocks on a stream must be in the order of smaller macroblock_address. That is, the streams are transmitted from top to bottom of the screen and from left to right.

In MPEG, one slice is usually composed of one stripe (16 lines), and variable length coding starts from the left end of the screen and ends at the right end. Therefore, V
When the MPEG elementary stream is recorded as it is by the TR, at the time of high-speed reproduction, the reproducible portion is concentrated on the left end of the screen and cannot be uniformly updated. Further, since the arrangement of the data on the tape cannot be predicted, if the tape pattern is traced at a constant interval, the screen cannot be uniformly updated. Furthermore, if an error occurs even at one location, it affects the right edge of the screen and cannot return until the next slice header is detected. For this purpose, one slice is composed of one macroblock.

FIG. 15 shows an example of the configuration of a recording / reproducing apparatus according to this embodiment. At the time of recording, the digital signal input from the terminal 100 is SDI (Serial Data Inter
face) It is supplied to the receiving unit 101. SDI is (4:
2: 2) SMPT for transmitting component video signals, digital audio signals and additional data
The interface specified by E. SDI
In the receiving unit 101, a digital video signal and a digital audio signal are respectively extracted from the input digital signal, the digital video signal is supplied to an MPEG encoder 102, and the digital audio signal is
The signal is supplied to the ECC encoder 109 via the delay 103. The delay 103 is for eliminating a time difference between the digital audio signal and the digital video signal.

The SDI receiver 101 extracts a synchronization signal from the input digital signal and supplies the extracted synchronization signal to the timing generator 104. An external synchronization signal can be input to the timing generator 104 from a terminal 105. The timing generator 104 generates a timing pulse based on a specified signal among the input synchronization signal and a synchronization signal supplied from an SDTI receiving unit 108 described later. The generated timing pulse is supplied to each section of the recording / reproducing apparatus.

The input video signal is supplied to the MPEG encoder 1
In 02, the signal undergoes DCT (Discrete Cosine Transform) processing, is converted into coefficient data, and the coefficient data is subjected to variable length coding. The variable length coded (VLC) data from the MPEG encoder 102 is an elementary stream (ES) compliant with MPEG2. This output is supplied to one input terminal of a recording-side multi-format converter (hereinafter, referred to as MFC) 106.

On the other hand, through the input terminal 107, the SDTI
(Serial data Transport Interface) format data is input. This signal is transmitted to the SDTI receiving unit 10
At 8, synchronization is detected. Then, it is buffered in the frame memory 170, and the elementary stream is extracted. The readout timing of the extracted elementary stream is controlled by a signal ready supplied from the recording side MFC 106, and the frame memory 170 is read out.
, And supplied to the other input end of the recording MFC 106. The synchronization signal obtained by synchronous detection by the SDTI receiving unit 108 is the same as that of the timing generator 1 described above.
04.

In one embodiment, for example, MPEG ES
In order to transmit (MPEG elementary stream), SDTI (Serial Data Transport Interface) -C
P (Content Package) is used. This ES is 4:
2: 2 components, and as described above, all are I-picture streams, and 1 GOP = 1
It has a picture relationship. In the SDTI-CP format, the MPEG ES is separated into access units and is packed in packets in frame units. The SDTI-CP uses a sufficient transmission band (27 MHz or 36 MHz at a clock rate and 270 Mbps or 360 Mbps at a stream bit rate), and can transmit ES in bursts in one frame period.

That is, system data, a video stream, an audio stream, and AUX data are arranged after the SAV of one frame period until the EAV.
There is no data in the entire one frame period, but data exists in a burst form for a predetermined period from the beginning. SDTI-CP streams (video and audio) can be switched in stream states at frame boundaries. The SDTI-CP has a mechanism for establishing synchronization between audio and video for content using SMPTE time code as a clock reference. Further, the format is determined so that SDTI-CP and SDI can coexist.

The SDTI receiving unit 108 outputs an enable signal “enable” indicating a valid section of the data based on the input stream. The enable signal is, for example, a signal that becomes “H” level in a data valid section and “L” level in an invalid section. The enable signal is supplied to, for example, the recording-side MFC 106.

In the interface using the SDTI-CP, the encoder and the decoder are VBV (Video B-Video), as in the case of transferring a TS (Transport Stream).
buffer Verifier) Buffers and TBs (Transport Buf
fers), so there is less delay. Further, the fact that the SDTI-CP itself is capable of extremely high-speed transfer further reduces the delay. Therefore, in an environment where synchronization exists to manage the entire broadcast station, SDTI-
It is effective to use CP.

The SDTI receiving unit 108 further extracts a digital audio signal from the input SDTI-CP stream. The extracted digital audio signal is supplied to the ECC encoder 109.

The recording MFC 106 has a built-in selector and stream converter. Recording MFC 106
Is shared by, for example, switching the mode with a reproduction-side MFC 114 described later. The processing performed in the recording MFC 106 will be described. MPEG as described above
One of the MPEG ESs supplied from the encoder 102 and the SDTI receiving unit 108 is selected by a selector and supplied to a stream converter.

In the stream converter, the DCT arranged for each DCT block in accordance with the MPEG2
The coefficients are grouped for each frequency component through a plurality of DCT blocks constituting one macroblock, and the grouped frequency components are rearranged. When one slice of the elementary stream is one stripe, the stream converter makes one slice consist of one macroblock. Further, the stream converter limits the maximum length of variable length data generated in one macroblock to a predetermined length. This can be achieved by setting the higher-order DCT coefficient to zero.

As will be described in detail later, in the stream converter, when the above-described DCT coefficients are rearranged, the supplied MPEG ES variable length code is solved with reference to a predetermined VLC table, and rearrangement is performed. After that, variable length coding is performed again. When solving this variable length code, it is detected whether or not there is a code in the stream that does not exist in the VLC table and that does not match the VLC table. If there is a code that is a mismatch of the VLC table in the stream, the process of decoding the variable length code is immediately stopped, and the input stream is discarded until the next start code is found. Also, a stream in which a VLC mismatch exists is corrected and output in a predetermined manner.

The converted elementary stream rearranged in the recording MFC 106 is the ECC encoder 1
09. The ECC encoder 109 is connected to a large-capacity main memory (not shown), and includes a packing and shuffling unit, an outer code encoder for audio, an outer code encoder for video, an inner code encoder, a shuffling unit for audio, and shuffling for video. Built-in parts. In addition, the ECC encoder 109
It includes a circuit for adding an ID in sync block units and a circuit for adding a synchronization signal. The ECC encoder 109 is composed of, for example, one integrated circuit.

In one embodiment, the error correction code for video data and audio data is
The product code is used. The product code encodes an outer code in a vertical direction of a two-dimensional array of video data or audio data, encodes an inner code in a horizontal direction thereof, and encodes data symbols doubly. Reed-Solomon code (Reed-Solomon code)
de) can be used.

The processing in the ECC encoder 109 will be described. Since the video data of the converted elementary stream is variable-length coded, the data length of each macroblock is not uniform. In the packing and shuffling sections, macroblocks are packed in fixed frames. At this time, the overflow portion that protrudes from the fixed frame is sequentially packed into an area that is vacant with respect to the size of the fixed frame.

Further, system data having information such as an image format and a version of a shuffling pattern is supplied from a system controller 121 described later, and is input from an input terminal (not shown). The system data is supplied to a packing and shuffling unit, and undergoes recording processing in the same manner as picture data. System data is video A
Recorded as UX. Also, shuffling is performed in which the macroblocks of one frame generated in the scanning order are rearranged and the recording positions of the macroblocks on the tape are dispersed. Shuffling can improve the image update rate even when data is reproduced in pieces during variable speed reproduction.

Video data and system data from the packing and shuffling unit (hereinafter, also referred to as video data even when system data is included unless otherwise required) are obtained by encoding the video data by outer coding. , And an outer code parity is added. The output of the outer code encoder is shuffled by a video shuffling unit to change the order in sync block units over a plurality of ECC blocks. The shuffling in sync block units prevents errors from concentrating on a specific ECC block. Shuffling performed by the shuffling unit may be referred to as interleaving. The output of the video shuffling unit is written to the main memory.

On the other hand, as described above, the SDTI receiving unit 1
08 or the digital audio signal output from the delay 103 is supplied to the ECC encoder 109. In this embodiment, an uncompressed digital audio signal is handled. The digital audio signal is not limited to these, and may be input via an audio interface. An audio AUX is supplied from an input terminal (not shown). Audio AUX
Is auxiliary data, which is data having information related to audio data such as the sampling frequency of audio data. The audio AUX is added to the audio data and is treated equivalently to the audio data.

The audio data to which the audio AUX is added (hereinafter, unless otherwise required, the audio data including the AUX is also simply referred to as audio data) is an audio outer code for encoding the audio data with an outer code. Supplied to the encoder. The output of the audio outer code encoder is supplied to the audio shuffling unit, and undergoes shuffling processing. As audio shuffling, shuffling in sync block units and shuffling in channel units are performed.

The output of the audio shuffling unit is written to the main memory. As described above, the output of the video shuffling unit is also written in the main memory, and the main memory mixes audio data and video data into one-channel data.

Data is read from the main memory, added with an ID having information indicating a sync block number, and supplied to the inner code encoder. The inner code encoder encodes the supplied data with the inner code. A synchronization signal for each sync block is added to the output of the inner code encoder, thereby forming recording data in which the sync blocks are continuous.

The recording data output from the ECC encoder 109 is supplied to an equalizer 110 including a recording amplifier and the like, and is converted into a recording RF signal. The recording RF signal is supplied to a rotating drum 111 provided with a rotating head in a predetermined manner, and is recorded on a magnetic tape 112. Actually, a plurality of magnetic heads having different azimuths of heads forming adjacent tracks are attached to the rotating drum 111.

The recording data may be scrambled as required. Further, digital modulation may be performed at the time of recording, and a partial response class 4 and Viterbi code may be used. Note that the equalizer 110 includes both a configuration on the recording side and a configuration on the reproduction side.

FIG. 16 shows an example of a track format formed on a magnetic tape by the rotary head described above. In this example, video and audio data per frame are recorded on four tracks. One segment is composed of two tracks having different azimuths. That is, four tracks are composed of two segments. Track number corresponding to azimuth for a set of tracks constituting a segment

[0] and a track number [1] are assigned. In each of the tracks, video sectors on which video data is recorded are arranged at both ends,
An audio sector in which audio data is recorded is arranged between video sectors. FIG. 16 shows the arrangement of sectors on the tape.

In this example, four channels of audio data can be handled. A1 to A4
Indicates audio data channels 1 to 4 respectively. The audio data is recorded with its arrangement changed in segment units. In this example, in this example, data of four error correction blocks is interleaved for one track, and Upper Side and Low
It is divided into sectors of r Side and recorded.

A system area (SYS) is provided at a predetermined position in the video sector of the lower side.
The system area is provided alternately for each track, for example, at the beginning and end of a lower side video sector.

In FIG. 16, SAT is an area in which a servo lock signal is recorded. A gap having a predetermined size is provided between the recording areas.

FIG. 16 shows that the data per frame is 4
In this example, data is recorded on a track. Depending on the format of data to be recorded / reproduced, data per frame can be recorded on eight tracks, six tracks, or the like.

As shown in FIG. 16B, the data recorded on the tape is composed of a plurality of blocks divided at equal intervals called sync blocks. FIG. 16C schematically shows the configuration of a sync block. The sync block is
It is composed of a SYNC pattern for synchronous detection, an ID for identifying each sync block, a DID indicating the content of subsequent data, a data packet, and an inner code parity for error correction. Data is handled as packets in sync block units. That is, the smallest data unit to be recorded or reproduced is one sync block. Many sync blocks are arranged (Fig. 1
6B) For example, a video sector is formed.

Referring back to FIG. 15, at the time of reproduction, a reproduction signal reproduced from the magnetic tape 112 by the rotary drum 111 is supplied to the reproduction side configuration of the equalizer 110 including a reproduction amplifier and the like. The equalizer 110 performs equalization and waveform shaping on the reproduced signal. Further, demodulation of digital modulation, Viterbi decoding, and the like are performed as necessary.
The output of the equalizer 110 is supplied to the ECC decoder 113.

The ECC decoder 113 performs the above-described ECC
It performs processing reverse to that of the encoder 109, and includes a large-capacity main memory, an inner code decoder, a deshuffling unit for audio and video, and an outer code decoder. Further, the ECC decoder 113 includes a deshuffling and depacking unit and a data interpolation unit for video. Similarly, an audio AUX separation unit and a data interpolation unit are included for audio. The ECC decoder 113 is composed of, for example, one integrated circuit.

The processing in the ECC decoder 113 will be described. The ECC decoder 113 first detects synchronization and detects a synchronization signal added to the beginning of a sync block, and cuts out the sync block. The reproduction data is supplied to the inner code decoder for each sync block, and error correction of the inner code is performed. The ID interpolation processing is performed on the output of the inner code decoder, and the ID of the sync block in which the error occurred due to the inner code, for example, the sync block number, is interpolated. The playback data with the interpolated ID is separated into video data and audio data.

As described above, the video data is the MPE
G means DCT coefficient data and system data generated in intra coding of G, and audio data is PCM
(Pulse Code Modulation) Data and audio AU
Means X.

The separated audio data is supplied to an audio deshuffling unit, and performs a process reverse to the shuffling performed by the recording-side shuffling unit. The output of the deshuffling unit is supplied to an outer code decoder for audio, and error correction by the outer code is performed. The audio outer code decoder outputs error-corrected audio data. An error flag is set for data having an uncorrectable error.

The audio AUX is separated from the output of the audio outer code decoder by the audio AUX separation unit, and the separated audio AUX is output from the ECC decoder 1.
13 (the path is omitted). Audio A
The UX is supplied to, for example, a system controller 121 described later.
The audio data is supplied to a data interpolation unit. In the data interpolating unit, an erroneous sample is interpolated. As the interpolation method, it is possible to use an average value interpolation for interpolating with the average value of correct data before and after in time, a previous value hold for holding a previous correct sample value, and the like.

The output of the data interpolation unit is the ECC decoder 11
The audio data output from the ECC decoder 113 is output to the delay 117 and the SDTI output unit 115. The delay 117 is provided to absorb a delay caused by processing of video data in the MPEG decoder 116 described later. The audio data supplied to the delay 117 is supplied with a predetermined delay to the SDI output unit 118.

The separated video data is supplied to a deshuffling unit, and the reverse processing of the shuffling on the recording side is performed. The deshuffling unit performs a process of restoring the shuffling in sync block units performed by the shuffling unit on the recording side. The output of the deshuffling unit is supplied to the outer code decoder, and error correction by the outer code is performed. When an error that cannot be corrected occurs, an error flag indicating the presence or absence of the error is set to indicate the presence of the error.

The output of the outer code decoder is supplied to a deshuffling and depacking unit. The deshuffling and depacking unit performs processing for restoring shuffling in macroblock units performed by the packing and shuffling unit on the recording side. In the deshuffling and depacking unit, the packing performed at the time of recording is disassembled. That is, the length of the data is returned in units of macroblocks, and the original variable length code is restored. Further, in the deshuffling and depacking unit, the system data is separated, output from the ECC decoder 113, and supplied to the system controller 121 described later.

The output of the deshuffling and depacking unit is supplied to a data interpolation unit, and data for which an error flag is set (that is, an error exists) is corrected.
That is, if it is determined that there is an error in the macroblock data before the conversion, the DCT coefficients of the frequency components after the error location cannot be restored. Therefore, for example, the DCT coefficient of the frequency component after the error point is set to zero. Similarly, at the time of high-speed reproduction, only the DCT coefficients up to the length corresponding to the sync block length are restored.
Replaced with zero data. Further, in the data interpolation section, when the header added to the head of the video data is an error, processing for recovering the header (sequence header, GOP header, picture header, user data, etc.) is also performed.

The video data and error flag output from the data interpolation unit are the output of the ECC decoder 113, and the output of the ECC decoder 113 is output to a reproduction-side multi-format converter (hereinafter, abbreviated as a reproduction-side MFC) 114. Supplied. The ECC decoder 113
Is output from the recording side MFC1 described above.
06 corresponds to the transformed elementary stream in which the DCT coefficients of the MPEG stream are rearranged.

The reproduction-side MFC 114 is the recording-side MFC 114 described above.
It performs the reverse process of the FC 106 and includes a stream converter. The reproduction-side MFC 114 is composed of, for example, one integrated circuit. In the stream converter,
An end-of-block (EOB: End Of Block) is added to video data having an error at an appropriate position using the error flag from the data interpolation unit, and the data is terminated.

Since the DCT coefficients are arranged from the DC component and the low-frequency component to the high-frequency component across the DCT block, even if the DCT coefficients are ignored from a certain point onward, the macro block , DCT coefficients from DC and low-frequency components can be distributed evenly to each of the DCT blocks constituting.

[0093] In the stream converter, a process reverse to that of the stream converter on the recording side is performed. That is, the DCT coefficients arranged for each frequency component across the DCT blocks are rearranged for each DCT block.
Further, the reproduction side MFC 114 detects the sequence extension 3 from the supplied stream and extracts the information of the chroma format. When the DCT coefficients are rearranged in the stream converter, predetermined timing control is performed based on the extracted chroma format information.
Thereby, the reproduced signal is converted into an elementary stream conforming to MPEG2.

As will be described later in detail, in the stream converter, when the above-described DCT coefficients are rearranged, the variable length code of the supplied converted stream is solved with reference to a predetermined VLC table, and rearranged. After that, variable length coding is performed again. When solving this variable length code, it is detected whether or not there is a code in the stream that does not exist in the VLC table and that does not match the VLC table. If there is a code that is a mismatch of the VLC table in the stream, the process of decoding the variable length code is immediately stopped, and the input stream is discarded until the next start code is found. Also, VLC
The stream having the mismatch is output after being corrected in a predetermined manner.

The input and output of the stream converter are as follows.
As with the recording side, a sufficient transfer rate (bandwidth) is secured according to the maximum length of the macroblock. If the length of the macroblock (slice) is not limited, it is preferable to secure a bandwidth three times the pixel rate.

The output of the stream converter is the reproduction side MF
The output of the reproduction side MFC 114 is output from the SDTI output unit 115 and the MPEG decoder 11.
6.

[0097] The MPEG decoder 116 decodes the elementary stream and outputs video data. That is, the MPEG decoder 116 performs the inverse quantization process and the inverse D
CT processing is performed. The decoded video data is supplied to the SDI output unit 118. As described above, audio data separated from video data by the ECC decoder 113 is supplied to the SDI output unit 118 via the delay 117. The SDI output unit 118 maps the supplied video data and audio data into an SDI format, and converts the data into a stream having a data structure in the SDI format. SDI output unit 11
8 is output from the output terminal 120 to the outside.

On the other hand, the SDTI output section 115 is supplied with the audio data separated from the video data by the ECC decoder 113 as described above. The SDTI output unit 115 converts the supplied video data and audio data as elementary streams into SDDT
I format and converted to a stream having a data structure of SDTI format. The converted stream is output from the output terminal 119 to the outside.

In FIG. 15, the system controller 121 is composed of, for example, a microcomputer and controls the entire operation of the recording / reproducing apparatus. Further, the servo 122 performs the running control of the magnetic tape 112 and the drive control of the rotating drum 111 while communicating with the system controller 121.

Here, the chroma format will be schematically described. FIGS. 17, 18 and 19 are diagrams for explaining the chroma formats 4: 4: 4, 4: 2: 2 and 4: 2: 0, respectively. Of these,
FIGS. 17A, 18A and 19A show the size of the luminance signal Y and the color difference signals Cb and Cr and the sampling phase. In the figure, “x” indicates the phase of the luminance signal Y, and two overlapping “「 ”indicate the phases of the color difference signals Cb and Cr.

FIG. 17 shows the chroma format 4: 4: 4.
As shown in A, the color difference signals Cb and Cr and the luminance signal Y
Have the same size and sampling phase. Therefore, in the case of a macro block including four DCT blocks each including 8 × 8 pixels, FIG.
As shown in (1), the matrix of the color difference signals Cb and Cr consists of four blocks of the same size as the matrix of the luminance signal Y in both horizontal and vertical dimensions.

On the other hand, chroma format 4:
2: 2 is the color difference signal Cb, as shown in FIG. 18A.
The size of Cr is half the size of the luminance signal Y in the horizontal direction. Therefore, considering the macroblock, the matrix of the color difference signals Cb and Cr is half the matrix of the luminance signal Y in the horizontal dimension.

Further, the chroma format 4: 2: 0
In FIG. 19A, the sizes of the color difference signals Cb and Cr are ｂ each of the size of the luminance signal Y in both the horizontal and vertical directions, as shown in FIG. 19A. Therefore, considering the macro block, the color difference signals Cb, Cr
Are half the matrix of the luminance signal Y in both the horizontal and vertical directions.

Note that FIG. 17B, FIG.
As shown in FIG.
The DCT blocks that make up the black block are shown in the figure.
Numbers 1, 2, 3 and 4 from the upper left
Each is represented. Each macro block shown in FIGS.
The coding order of the blocks in the lock is
Is 4: 4: 4, and as shown in FIG. 17B,
Y₁, Y₂, Y₃, Y ₄, Cb₁, Cr₁, Cb₂, C
r₂, Cb₃, Cr₃, Cb₄And Cr₄In order
You. Similarly, for chroma format 4: 2: 2,
As shown in FIG. 18B, Y₁, Y₂, Y₃, Y₄,
Cb₁, Cr₁, Cb₂And Cr₂In the order of
Format 4: 2: 0, as shown in FIG.
Y₁, Y₂, Y₃, Y₄, Cb₁And Cr
₁It becomes the order of.

FIG. 20A shows the order of DCT coefficients in video data output from the DCT circuit of the MPEG encoder 102. M output from SDTI receiving section 108
The same applies to PEG ES. In the following, MPE
A description will be given using the output of the G encoder 102 as an example. DC
Starting from the upper left DC component in the T block, DCT coefficients are output in a zigzag scan in a direction in which the horizontal and vertical spatial frequencies increase. As a result, FIG.
As shown in FIG. 1, a total of 64 (8 pixels × 8 lines) DCT coefficients are obtained by being arranged in order of frequency components.

This DCT coefficient is equal to the V of the MPEG encoder.
Variable length coding is performed by the LC unit. That is, the first coefficient is fixed as a DC component, and the next component (AC
From the component), codes are assigned corresponding to the run of zero and the subsequent level. Therefore, the variable-length coded output for the coefficient data of the AC component is converted from the low (low-order) coefficient of the frequency component to the high (high-order) coefficient of AC ₁ ,
AC ₂ , AC ₃ ,... The elementary stream includes DCT coefficients subjected to variable length coding.

In the recording-side stream converter built in the recording-side MFC 106, the DCT coefficients of the supplied signal are rearranged. That is, in each macroblock, DCT coefficients arranged in order of frequency components for each DCT block by zigzag scan are rearranged in order of frequency components over each DCT block constituting the macroblock.

FIG. 21 schematically shows rearrangement of DCT coefficients in the recording-side stream converter.
In the case of a (4: 2: 2) component signal, one macroblock is composed of four DCT blocks (Y ₁ , Y ₂ , Y ₃ and Y ₄ ) based on a luminance signal Y and color difference signals Cb and Cr. Two DCT blocks (Cb ₁ , Cb ₁
b ₂ , Cr ₁ and Cr ₂ ).

As described above, the MPEG encoder 10
In 2, the zigzag scan is performed in accordance with the rules of MPEG2, and as shown in FIG. 21A, the DCT coefficients are arranged in order of frequency components from DC components and low-frequency components to high-frequency components for each DCT block. When scanning of one DCT block is completed, scanning of the next DCT block is performed, and similarly, DCT coefficients are arranged.

That is, in the macro block, DCT blocks Y ₁ , Y ₂ , Y ₃ and Y ₄ , DCT block C
For each of b ₁ , Cr ₁ , Cb ₂ and Cr ₂ , D
The CT coefficients are arranged in order of frequency from the DC component and the low frequency component to the high frequency component. Then, [DC, AC ₁ , AC ₂ , AC
_3, a..], So that codes are assigned, are variable length coded.

In the recording-side stream converter, the variable-length coded and arranged DCT coefficients are once decoded by decoding the variable-length code to detect a break of each coefficient, and are straddled over each DCT block constituting a macroblock. Summarize for each frequency component. This is shown in FIG. 21B. First, the DC components of the eight DCT blocks in the macroblock are summarized, the AC coefficient components of the eight DCT blocks having the lowest frequency components are summarized, and the AC coefficients of the same order are grouped in order. The coefficient data is rearranged across the DCT blocks.

The rearranged coefficient data is DC
(Y ₁ ), DC (Y ₂ ), DC (Y ₃ ), DC (Y ₄ ), D
C (Cb ₁ ), DC (Cr ₁ ), DC (Cb ₂ ), DC
(Cr ₂ ), AC ₁ (Y ₁ ), AC ₁ (Y ₂ ), AC
₁ (Y ₃ ), AC ₁ (Y ₄ ), AC ₁ (Cb ₁ ), AC ₁ (C
r ₁ ), AC ₁ (Cb ₂ ), AC ₁ (Cr ₂ ),. Here, DC, AC ₁ , AC ₂ ,... Are the respective codes of the variable length codes assigned to the set consisting of the run and the subsequent level, as described with reference to FIG. .

The conversion elementary stream in which the order of the coefficient data is rearranged by the recording-side stream converter is supplied to a packing and shuffling unit built in the ECC encoder 109. The data length of the macroblock is the same for the converted elementary stream and the elementary stream before conversion. MPE
In the G encoder 102, even if the length is fixed in GOP (one frame) units by bit rate control, the length varies in macroblock units. In the packing and shuffling unit, the data of the macroblock is applied to a fixed frame.

FIG. 22 schematically shows a macroblock packing process in the packing and shuffling unit. The macro block is applied to a fixed frame having a predetermined data length and is packed. The data length of the fixed frame used at this time matches the data length of the payload, which is the data storage area of the sync block, which is the minimum unit of data during recording and reproduction. This is to simplify the processing of shuffling and error correction coding. In FIG. 22, for simplicity, it is assumed that one frame includes eight macroblocks.

As shown in an example in FIG. 22A, the lengths of eight macroblocks are different from each other due to the variable length coding. In this example, as compared with the length of the data area (payload) of one sync block, which is a fixed frame, the data of macro block # 1, the data of # 3 and the data of # 6 are each longer, and the data of macro block # 2 is longer. , # 5, # 7 and # 8 are short. The data of the macro block # 4 has a length substantially equal to the payload.

By the packing process, the macro block is packed in a fixed length frame of the payload length. Data can be packed without excess or shortage because the amount of data generated in one frame period is controlled to a fixed amount. As shown in an example in FIG. 22B, a macroblock longer than the payload is divided at a position corresponding to the payload length. Of the divided macroblocks,
Portion outside the payload length (overflow portion)
Are packed in an area that is empty from the beginning, that is, after a macroblock whose length is less than the payload length.

In the example of FIG. 22B, macro block # 1
The portion that extends beyond the payload length is first packed after the macroblock # 2, and when it reaches the length of the payload, it is packed after the macroblock # 5. Next, the portion of the macroblock # 3 that extends beyond the payload length is packed after the macroblock # 7. In addition, the portion of the macro block # 6 that extends beyond the payload length is packed after the macro block # 7,
Further, the protruding portion is packed behind macro block # 8. Thus, each macroblock is packed into a fixed frame of the payload length.

The length of the variable-length data corresponding to each macroblock can be checked in advance by the recording-side stream converter. This allows the packing unit to know the end of the macroblock data without decoding the VLC data and checking the contents.

As described above, in this embodiment, processes such as rearrangement of DCT coefficients in a macroblock and packing of macroblock data into a payload in units of one picture are performed. Therefore, even if an error occurs beyond the error correction capability of the error correction code due to, for example, tape dropout,
Image quality degradation can be suppressed to a small extent.

The effects of coefficient rearrangement and packing will be described with reference to FIGS. 23 and 24. Here, an example in which the chroma format is 4: 2: 2 will be described. FIG. 23 shows a case where DCT blocks and DCT coefficients are supplied according to MPEG ES. In this case, as shown in FIG. 23A, followed by a slice header or a macroblock (MB) header, DCT blocks the luminance signal _Y 1 to Y _4, in the order of the color difference signals _{Cb 1,} Cr _{1, Cb 2,} and Cr ₂ Are lined up. In each block,
The DCT coefficients are arranged from the low order to the high order of the DC component and the AC component.

Here, for example, in an ECC decoder or the like, the error correction capability of the error correction code exceeds
Timing position A of A, i.e., an error occurs in the high-order coefficients of the block Cb _1. As mentioned above,
In MPEG, a slice forms one variable-length code sequence. Therefore, once an error occurs, data from the error position to the detection of the next slice header is not reliable. Therefore, in this stream composed of one slice = 1 macroblock, it is not possible to decode data subsequent to position A in this macroblock.

As a result, as shown in an example in FIG. 23B, the blocks Cr ₁ , Cb ₂ and Cr
No. ₂ cannot reproduce even the DC component. Therefore, the portion B corresponding to the blocks Y ₁ and Y _2, since the block order and other color difference signal blocks Cb ₁ can not be reproduced, the image of the abnormal color obtained by low-order coefficient block Cb ₁ turn into. The portion C corresponding to the block Y ₃ and Y _4, the luminance signal only is reproduced, resulting in a monochrome image.

FIG. 24 shows a transform stream in which DCT coefficients are rearranged according to this embodiment. Also in this example, it is assumed that an error has occurred at the position A, as in the case of FIG. In the transform stream, as shown in an example in FIG. 24A, a block in which DCT coefficients are grouped for each component across DCT blocks, following a slice header or a macroblock header, is divided into a DC component and A block.
The C components are arranged from the lower order to the higher order.

Even in this case, data after the error position cannot be relied on until the next slice header is detected.
The data after the error position A in this macro block is
Not reproduced. However, in this converted stream, the data that cannot be decoded due to the error is each DCT
The higher order side of the AC component in the DCT coefficient in the block, and the lower order side of the DC component and the AC component in the DCT coefficient of each DCT block are obtained evenly.
Therefore, as shown in FIG. 24B, the high-order AC component is not reproduced, so that the detailed part of the image is lacking. However, as in the case of the above-mentioned MPEGES, the image becomes monochrome or the two color difference components In most cases, an abnormal color in which one of them is missing can be avoided.

Thus, even if the data stored in another fixed frame length cannot be reproduced by the above-mentioned packing, a certain image quality can be secured. Therefore, deterioration of image quality at the time of high-speed reproduction or the like can be suppressed.

If an error occurs in the VLC, data becomes unreliable until the next header (start code) is detected. In the VLC, a data sequence is converted using a predetermined table to which a code length having a length corresponding to the appearance frequency of data is assigned. Therefore, even if only one bit in the variable-length-encoded data string is bit-inverted, it may be resolved as a VLC having a different meaning. Therefore, the VLC after the occurrence of the error is unreliable even if it is unraveled, and it is necessary to discard the data until reliable data appears. As described above, the start code of each layer except the macroblock layer is
It consists of a unique code with code boundaries separated by bytes. Therefore, it is possible to recover from an error by detecting this start code.

FIG. 25 shows the ECC encoder 10 described above.
9 shows a more specific configuration. In FIG. 25, 164
Is an interface of the main memory 160 external to the IC. The main memory 160 is composed of an SDRAM. The interface 164 arbitrates an internal request for the main memory 160, and performs write / read processing on the main memory 160. The packing and shuffling unit 137 is constituted by the packing unit 137a, the video shuffling unit 137b, and the packing unit 137c.

FIG. 26 shows an example of the address configuration of the main memory 160. The main memory 160 is, for example, 64
It is composed of an M-bit SDRAM. Main memory 16
0 has a video area 250, an overflow area 251 and an audio area 252. Video area 250
Are four banks (vbank # 0, vbank # 1,
vbank # 2 and vbank # 3). Each of the four banks can store a digital video signal of one equal length unit. One equalization unit is a unit for controlling the amount of generated data to a substantially target value, for example, one unit of a video signal.
This is a picture (I picture). Part A in FIG.
Indicates a data portion of one sync block of the video signal. Data of a different number of bytes is inserted into one sync block depending on the format. In order to support a plurality of formats, the data size of one sync block is equal to or larger than the maximum number of bytes and is a number of bytes convenient for processing, for example, 256 bytes.

Each bank in the video area is further divided into a packing area 250A and an output area 250B to the inner encoding encoder. Overflow area 251
Consists of four banks corresponding to the above-mentioned video area. Further, the main memory 160 has an area 252 for audio data processing.

In this embodiment, the packing unit 1 is referred to by referring to the data length indicator of each macro block.
Reference numeral 37a stores the fixed frame length data and the overflow data, which is a portion beyond the fixed frame, in separate areas of the main memory 160 and stores them. The fixed frame length data is data that is shorter than the length of the data area (payload) of the sync block, and is hereinafter referred to as block length data. The area for storing the block length data is the packing processing area 250A of each bank. If the data length is shorter than the block length, an empty area is created in the corresponding area of the main memory 160. The video shuffling unit 137b performs shuffling by controlling the write address. Here, the video shuffling unit 137b shuffles only the block length data, and the overflow portion is written to the area assigned to the overflow data without shuffling.

Next, the packing section 137c performs processing of packing and reading the overflow portion into the memory for the outer code encoder 139. That is, 1 is prepared from the main memory 160 to the outer code encoder 139.
The data of the block length is read into the memory of the ECC block, and if there is a free area in the data of the block length, an overflow portion is read there to block the data to the block length. When data for one ECC block is read, the reading process is temporarily suspended, and the outer code encoder 139 generates parity of the outer code. The outer code parity is the outer code encoder 13
9 is stored in the memory. When the processing of the outer code encoder 139 is completed for one ECC block, the outer code encoder 1
From 39, the data and the outer code parity are rearranged in the order of performing the inner code, and are written back to the packing processing area 250A of the main memory 160 and another output area 250B.
The video shuffling unit 140 performs shuffling on a sync block basis by controlling an address at the time of writing back the data on which the encoding of the outer code has been completed to the main memory 160.

As described above, the block length data and the overflow data are separated and the data is written into the first area 250A of the main memory 160 (first packing process), and the overflow data is packed into the memory of the outer code encoder 139. (The second packing process), the generation of the outer code parity, and the data and the outer code parity in the second area 250 of the main memory 160.
The process of writing back to B is performed in units of one ECC block.
Since the outer code encoder 139 includes the memory having the size of the ECC block, the frequency of access to the main memory 160 can be reduced.

A predetermined number of Es contained in one picture
When the processing of the CC block (for example, 32 ECC blocks) is completed, the packing of one picture and the encoding of the outer code are completed. The data read from the area 250B of the main memory 160 via the interface 164 is processed by the ID addition unit 148, the inner code encoder 149, and the synchronization addition unit 150, and the output data of the synchronization addition unit 150 is output by the parallel / serial conversion unit 124. Is converted to bit serial data. The output serial data is processed by the partial response class 4 precoder 125. This output is digitally modulated as required, and is supplied via a recording amplifier 110 to a rotating head provided on a rotating drum 111.

It is to be noted that a sync block called null sync, in which valid data is not arranged, is introduced into the ECC block, and an ECC block is provided for the difference in recording video signal format.
The configuration of the C block is made flexible.
The null sink is generated in the packing unit 137a of the packing and shuffling block 137, and written into the main memory 160. Therefore, since the null sync has a data recording area, it can be used as a recording sync for the overflow portion.

In the case of audio data, even-numbered samples and odd-numbered samples of one-field audio data form different ECC blocks. Since the ECC outer code sequence is composed of audio samples in the input order, the outer code encoder 136 generates an outer code parity each time an outer code sequence audio sample is input. The address control when writing the output of the outer code encoder 136 to the area 252 of the main memory 160 causes the shuffling unit 137 to perform shuffling (in units of channels and units of sync blocks).

Further, a CPU interface indicated by 126 is provided, receives data from an external CPU 127 functioning as a system controller, and sets parameters for internal blocks. In order to support a plurality of formats, it is possible to set many parameters including a sync block length and a parity length.

The “packing length data” as one of the parameters is sent to the packing units 137a and 137b, and the packing units 137a and 137b determine the fixed frame (“payload length” in FIG. 22A). At the indicated length).

[0138] "Pack number data" as one of the parameters is sent to the packing unit 137b, and the packing unit 137b determines the number of packs per sync block based on the data and the data for the determined number of packs. Is supplied to the outer code encoder 139.

"Video outer code parity number data" as one of the parameters is sent to outer code encoder 139, and outer code encoder 139 outputs the outer code of the video data from which the number of parities is generated based on this. Perform encoding.

Each of “ID information” and “DID information” as one of the parameters is sent to the ID adding section 148, and the ID adding section 148
The ID information is added to the unit-length data string read from the main memory 160.

Each of “parameter number data for video inner code” and “parity number data for audio inner code” as one of the parameters is an inner code encoder 149.
And the inner code encoder 149 encodes the inner code of the video data and the audio data from which the number of parities are generated based on these. It should be noted that “sink length data”, which is one of the parameters, is also sent to the inner code encoder 149, whereby the unit length (sink length) of the inner-coded data is regulated.

The shuffling table data as one of the parameters is stored in a video shuffling table (RAM) 128v and an audio shuffling table (RAM) 128a. The shuffling table 128v performs an address conversion for shuffling the video shuffling units 137b and 140. The shuffling table 128a performs an address conversion for the audio shuffling 137.

As described above, in MPEG, a stream is compression-encoded using a variable length code (VLC) to which a predetermined code length is assigned according to the appearance rate of data. Variable-length encoding of a stream is performed by referring to a VLC table in which data values and variable-length codes are associated as parameters representing variable-length codes. VL
Various C tables are prepared according to the type of data to be subjected to variable-length coding. FIG. 27 to FIG.
4 shows an example of a VLC table used in MPEG. FIGS. 27 to 39 show ITU-T Re.
c. H. 262 (1995E).

FIGS. 27 to 29 show VLC tables related to parameters included in the macroblock header 14. FIG. 27 shows macroblock_addresses.
It is a VLC table for s_increment. VLC codes are assigned according to the values indicated in increment_value. The same notation is used in the following VLC table.
FIGS. 28 and 29 are VLC tables for macroblock_type in I and P pictures, respectively.

FIGS. 30 to 39 are VLC tables related to parameters included in the DCT block. FIG.
And FIG. 31 respectively show dct_dc_size_l
uminance and dct_dc_size_ch
It is a VLC table for romance.

FIGS. 32 to 35 and FIGS. 36 to 3
Reference numeral 9 denotes a VLC table relating to the run and level of the DCT coefficient, which has already been described with reference to FIG. FIG.
FIG. 35 is a series of diagrams, each showing a DCT coefficient table 0.
(DCT coefficients Table
2 shows a VLC table called 0). Similarly, FIG.
FIG. 39 to FIG. 39 are a series of diagrams, each showing a DCT coefficient table 1
(DCT coefficients Table
1 shows a VLC table referred to as 1). In these tables, one VLC for a pair of runs and levels
A code is assigned.

In DCT coefficient table 0, VLC
The code “10” is determined to be the EOB indicating the end of the DCT block, and the VLC code “0000 — 01” is determined to be the escape code. Similarly, in DCT coefficient table 1, VLC code “0110”
It is determined that the EOB indicates the end of the T block. The escape code is the same as DCT coefficient table 0,
The LC code is “0000 — 01”.

FIGS. 40 and 41 show tables of fixed-length codes (FLC), respectively. The DCT coefficient tables 0 and 1 described above show that for run and level pairs,
As shown in FIGS. 32 to 35 and FIGS. 36 to 39, 113 VLCs having a high frequency of occurrence are prepared. For run and level pairs not provided in DCT coefficient tables 0 and 1, VL
The run and level FLs shown in FIGS. 40 and 41, respectively, following the C code escape code
Expressed by C.

In decoding the VLC, for example, the above-described VLC table is referred to. Here, a case where a VLC that does not exist in the VLC table is encountered when decoding a variable-length-encoded stream and a mismatch occurs in the VLC table will be described with reference to FIG.

FIG. 42A shows a normal stream.
Macroblock, DC following slice start code
Transmitted in the order of the T coefficient, and when the macroblock ends,
The next slice start code is transmitted. FIG. 42
In the example, the DCT coefficient is subjected to variable-length coding using the DCT coefficient table 1 described above, and the variable-length code of the portion using the DCT coefficient table 1 is as follows.

... 0100 0000 — 0001 — 0010 — 0 0001 — 010 0010 — 0110 — 1 100.

In the stream shown in FIG. 42A, if the stream is momentarily interrupted at a position A due to a transmission error or the like, the stream becomes as shown in FIG. 42B. become.

... 0100 0000_0001_0010_0 0001_010 0010_01 ...

In the variable length code, the code “0010 — 01” in the fourth row is a VLC that does not exist in the DCT coefficient table 1. Therefore, the variable length code of this stream cannot be solved.
When a code that does not exist in such a predetermined VLC table is input to some decoders, a hang-up or the like may occur as an unexpected code is input. Once it hangs up,
The decoder cannot return to a normal state unless it is initialized by, for example, turning off the power.

Therefore, according to the present invention, when a variable length code of an MPEG stream is encountered, if a code that does not exist in the VLC table is encountered, the process of decoding the variable length code is immediately stopped, and until the next start code is detected. Discard the input MPEG stream. In the example of FIG. 42B described above, the code from the code “0010 — 01” that does not exist in the VLC table corresponding to this stream to immediately before the next start code is discarded. Here, the slice start code [00 00 0101
(~ AF)] is discarded.

By this processing, for example, it is possible to avoid inputting a code that does not exist on the VLC table to a decoder connected at the subsequent stage.

However, there is a possibility that a problem may occur if the stream after the code which does not exist in the VLC table is cut off by the above-described processing and is directly output to the subsequent signal processing block. For example, an EOB indicating the end of the block is added to each DCT block, and by counting the number of EOBs in the macro block,
It is possible to know whether or not the macro block has ended. If the stream is partially discarded by the above processing, the number of EOBs in one macroblock becomes smaller than the specified number, and there is a possibility that the syntax of the MPEG is violated.

In the subsequent signal processing block, when it is expected that the input stream contains all the specified EOBs, if the number of EOBs is small, the operation of this signal processing block stops. There is a risk that it will. For example, the signal processing block may not proceed to the next processing until a specified number of EOB blocks are detected in one macroblock.

According to the present invention, in addition to the above-described process of discarding a stream from a code that does not exist in the VLC table to the next header, the stream is corrected and output so that there is no MPEG syntax error.

Although the details will be described later, in actuality, the VLC decoding unit detects whether or not a mismatch of the VLC table has occurred in the input stream, and based on the detection result, the VLC mismatch occurs. The signal table_mismatch indicating the current position is VLC
Output from the decoding unit. For example, the signal table_mi
When smatch is in the “H” state, it indicates that a VLC mismatch has occurred at that position. This signal tabl
The stream is discarded and corrected based on e_mismatch.

[0161] The stream correction processing according to this embodiment will be described. The processing is different between a converted stream in which DCT coefficients of an MPEG stream are rearranged and an MPEG stream.

The processing differs depending on the target VLC table. In this embodiment, the target VLC table is defined as (1) dct_coefficients
VLC table of (DCT coefficient table 0 or 1 described above), (2) dct_dc_size_luminan
VLC table of ce, (3) dct_dc_size
VLC table of _chrominance, (4) m
VLC table of acroblock_type and (5) macroblock_address_inc
The VLC table is a VLC table of the element.

Among them, dct_coef of (1)
The VLC table of “fients” is a process related to the DCT coefficient of the DCT block. Dct_d of (2)
VLC table of c_size_luminance and dct_dc_size_chromin of (3)
The VLC table of annce is a process related to a DCT block of DC coefficients in a luminance block and a chrominance block, respectively. In addition, the macroblock of (4)
VLC table of k_type and macr of (5)
The VLC table of obblock_address_increment is a process related to a header.

First, the processing for the mismatch of the VLC table shown in the above (1) to (5) in the case of the converted stream will be described with reference to FIGS. 43 to 45.
FIG. 43 is a diagram illustrating the dct_coefficients of (1).
An example in which a mismatch has occurred in the VLC table of FIG. As shown in FIG. 43A, a DCT coefficient is transmitted following a slice header and a macroblock header. Since this is a transformed stream, the coefficients are
The CT coefficient DC is placed at the head, and then the DC of the AC component
When the T coefficient changes from a low frequency component to a high frequency component, the DCT coefficient A
C _1, AC _2, ···, it is arranged with the AC _63. Note that all of the AC components are not always arranged.

[0165] The EOB is arranged after the DCT coefficient of the last AC component. D of DC and AC components
Each of the CT coefficient and EOB is further divided into DCT blocks Y ₁ , Y ₂ , Y ₃ and Y ₄ by the luminance component Y, and DCT blocks Cb ₁ by the color difference components Cb and Cr,
It consists of Cr ₁ , Cb ₂ and Cr ₂ . This arrangement is shown in FIG.
3, common in FIG. 44 and FIG.

[0166] Here, as an example in FIG. 43B is shown, the signal table_mismatch at the position of the DCT blocks Y ₄ of the luminance component Y in AC ₂ of the AC component
Is set to the “H” state, and a VLC mismatch exists in this DCT coefficient. Since the conversion stream is a VLC, the data thereafter is not reliable.

[0167] In this case, as exemplified in FIG. 43C is shown, in the DCT coefficients AC ₂ of the AC component VLC mismatch is present, the DCT blocks Y ₄ and each subsequent component of the luminance component VLC mismatch exists DCT
Blocks Cb ₁ , Cr ₁ , Cb ₂ , Cr ₂ , Y ₁ , Y ₂ , Y ₃
Is replaced with VLC indicating EOB in the DCT coefficient table. DCT coefficients of AC ₃ and later discarded, the macro block is terminated.

At this time, it is considered that EOB is added to all DCT blocks in the macro block. That is, the specified number of E
It is considered that OB is present. The stream thus modified is as shown in FIG. 43C.

FIG. 44 shows dct_dc_siz of (2).
e_luminance VLC table and (3)
7 shows an example in which a mismatch has occurred in the VLC table of dct_dc_size_chrominance. FIG. 44A is the same as FIG. 43A described above.

Here, as shown by the dotted line in FIG. 44B, table_mis is set at the position of the DCT block (in this example, DCT block Y ₄ ) of the luminance component Y of the DC component.
match is set to the “H” state, and dct_dc_siz
If there is a VLC mismatch in the e_luminance table, as shown in FIG.
DCT coefficient with LC mismatch (Y ₄ , DC)
Is, for example, dct_dc_size_l representing gray display.
uminance VLC and dct_dc_dif
replaced with ferential. DCT coefficients after that are, for example, dct_dc_size_ representing gray display.
Chrominance VLC and dct_dc_
replaced by differential. Further, EOB is added to each block, and the macro block is terminated.

It should be noted that table_misma is shown in FIG. 44B.
The tab at the position of the DCT coefficient Cr ₂ of the color difference component Cr ₂ in the DC component so that tch is set to the “H” state by the solid line.
le_mismatch is set to the “H” state, and this DC
Dct_dc_size_chromi in T coefficient
It is also conceivable that a nonce VLC mismatch exists.

As described above, when the VLC mismatch exists in only one of the color difference components Cr and Cb, only the one DCT block in which the VLC mismatch exists is replaced by the gray DCT block. Then, the block becomes a block of an abnormal color on the screen display. This is because, as a macroblock, the DCT block of the DC component of the other color difference block forming a pair is operated. In this embodiment, in order to avoid such a situation, FIG.
As shown in # 1, the other color difference component (C
Go back to the DCT block of b ₂ ), and replace it with the gray DCT block as described above.

Thereafter, EOB is further added to each DCT block.
Is added, and the macroblock is aborted. The modified stream is as shown in FIG. 44C.

FIG. 45 shows an example in which a VLC mismatch exists in a header portion, for example, a slice header or a macroblock header. This is based on the mac of (4) described above.
VLC table of roblock_type or macroblock_address_in of (5)
This corresponds to a case where a mismatch has occurred with respect to the VLC table of the “crement”. FIG. 45A is similar to FIG.
Is the same as

First, macroblock_t of (4)
A case where a VLC mismatch has occurred in the VLC table of type y will be described. More specifically, if the picture is an I picture, the VLC code is “00” and a VLC mismatch occurs according to FIG. 28 described above. table_mi
As shown in FIG. 45B, smatch is set to the “H” state at a corresponding position in the header section.

In this case, macroblock_type
e is a macro representing macroblock_intra
It is replaced by the VLC of obblock_type. For example, in the portion indicated by # 2 in FIG.
obblock_type is replaced with VLC code “1”. Thereafter, if necessary, dct_type is added. For example, in the part indicated by # 2 in FIG. 45C, dct_type is replaced with “1”.

Further, the DCT block Y _{1 of} the luminance component,
Dct_d in which Y ₂ , Y ₃ and Y ₄ represent, for example, gray display
VLC and dc of c_size_luminance
DCT blocks Cb ₁ , Cr ₁ , Cb ₂ and Cr ₂ which are replaced by t_dc_differential and have color difference components
Is, for example, dct_dc_size_ch representing gray display.
VLC and dct_dc_di of romance
Replaced with "ferential". In addition, each DCT
EOB is added to the block, and the macroblock is aborted. The modified stream is as shown in FIG. 45C.

(5) macroblock_addr
A case where a VLC mismatch has occurred in the VLC table of ess_increment will be described. In this case, the mb_row and mb of the macroblock
Using the continuity of the _column, the VL from the mb_row and mb_column of the immediately preceding macroblock
Mb_row of macroblock where C mismatch occurred
And mb_column. Mb_ found
macro based on row and mb_column
V of block_address_increment
LC is replaced. Then, macroblock_
For example, type is replaced with VLC “1” of macroblock_type representing macroblock_intra.

After that, if necessary, dct_type
Is added. For example, in the part indicated by # 2 in FIG. 45C, dct_type is replaced with “1”. Further, the DCT blocks Y ₁ , Y ₂ , Y ₃ and Y _{4 of} the luminance component are, for example, dct_dc_size_ representing gray display.
Luminance VLC and dct_dc_di
DCT blocks Cb ₁ , Cr ₁ , Cb ₂ and Cr _{2 of} the chrominance components are replaced by different, for example, dct_dc_size_chrominan representing gray display
ce VLC and dct_dc_different
ial. Furthermore, EO is assigned to each DCT block.
B is added and the macroblock is aborted. The modified stream is as shown in FIG. 45C.

Next, the processing for the mismatch of the VLC table shown in the above (1) to (5) in the case of the MPEG stream will be described with reference to FIGS. 46 to 48. FIG. 46 shows dct_coefficien of (1).
An example in which a mismatch has occurred in the VLC table of ts will be described. As shown in FIG. 46A, data of the luminance block Y ₁ by the luminance component Y Following the slice header and macroblock header is transmitted. Luminance block Y ₁
Followed by the luminance blocks Y ₂ , Y ₃ , Y ₄ and the color difference component C
The color difference blocks Cb ₁ , Cr ₁ , Cb ₂ and Cr ₂ by b and Cr are arranged.

The luminance blocks Y ₁ , Y ₂ , Y ₃ and Y ₄ and the color difference blocks Cb ₁ , Cr ₁ , Cb ₂ and Cr ₂
, With the DCT coefficient of the DC component at the top,
The DCT coefficients of the AC component are arranged from the lower order to the higher order, and finally EOB is added. This arrangement is the same in FIGS. 47 and 48.

[0182] Here, as an example in FIG. 46B is shown, DCT coefficients of the AC components in the luminance block Y ₃ A
Table_mismatch becomes "H" state at the position of C _3, it is assumed that the VLC mismatch occurs in the DCT coefficients. Since the MPEG stream is a VLC, the data thereafter is unreliable.

In this case, the DCT coefficient having the VLC mismatch is replaced with VLC representing EOB, and the macro block is aborted. For example, as exemplified in FIG. 46C is shown, DCT coefficients AC ₃ of the AC components in the luminance block Y ₃ which VLC mismatch exists EOB
Replaced by a block of the luminance block Y ₃ is aborted. For MPEG stream, this luminance block Y luminance signal follows the _third block Y ₄ and the color difference signals _{Cb 1, Cr 1, Cb 2} , the Cr ₂ blocks, it will not even be reproduced DC component.

Therefore, in this embodiment, MPEG
In order to avoid the syntax violation of the above, these blocks are replaced with DCT blocks of DC components representing gray display, for example, and EOB is added to each block,
The macroblock is aborted.

In the example of FIG. 46, when a mismatch of the VLC table occurs in the middle of the chrominance blocks Cb and Cr, a calculation is performed on a pair of chrominance blocks Cb and Cr in the macro block as described later. There is a possibility that the macro block will be displayed in an abnormal color. In order to avoid this, as will be described later, a pair of color difference blocks is retroactively replaced with a DCT block such as gray display and an EOB is added.

FIG. 47 shows dct_dc_siz of (3).
An example in which a mismatch has occurred in the VLC table of e_chrominance is shown. FIG. 47A is the same as FIG. 46A described above.

At the position shown in FIG. 47B, tabl
It is assumed that e_mismatch is in the “H” state, and a VLC table mismatch has occurred in the DCT coefficient DC of the DC component of the color difference block Cr ₂ . In this case, the DCT block is, for example, V of dct_dc_size_chrominance representing gray display.
It is replaced by LC and dct_dc_differential to make a DCT block of a DC component, and E
OB is added and the macroblock is aborted.

As described above, when the VLC mismatch exists in only one of the color difference components Cr and Cb, only the one DCT block in which the VLC mismatch exists is replaced by the gray DCT block. Then, the block becomes a block of an abnormal color on the screen display. This is because, as a macroblock, the DCT block of the DC component of the other color difference block forming a pair is operated. In this embodiment, in order to avoid such a situation, FIG.
As shown in # 1, D of the other color difference component forming a pair
Going back to the CT block, as described above,
Replace with a CT block. Thus, the area corresponding to the luminance blocks Y ₃ and Y ₄ of the macro block is
The color is not abnormal and can be reproduced as a monochrome image.

After that, EOB is further added to each DCT block.
Is added, and the macroblock is aborted. The modified stream is as shown in FIG. 47C.

[0190] Incidentally, the DCT blocks of the luminance component Y of the DC component, for example at the location of the DCT blocks Y ₄ table_
Mismatch is set to the “H” state, and dct_dc_
If a VLC mismatch exists in the size_luminance table,
The CT coefficient (Y ₄ , DC) is, for example, dc representing gray display.
Replaced by VLC of t_dc_size_luminance and dct_dc_differential. The subsequent DCT coefficients are, for example, d representing gray display.
ct_dc_size_chrominance VL
Replaced by C and dct_dc_differential. Further, EOB is added to each block, and the macro block is terminated. The modified stream is as shown in FIG. 47D.

FIG. 48 shows an example in which a VLC mismatch exists in a header portion, for example, a slice header or a macroblock header. This is based on the mac of (4) described above.
VLC table of roblock_type or macroblock_address_in of (5)
This corresponds to a case where a mismatch has occurred with respect to the VLC table of the “crement”. FIG. 48A is similar to FIG.
Is the same as

First, macroblock_t of (4)
A case where a VLC mismatch has occurred in the VLC table of type y will be described. More specifically, if the picture is an I picture, the VLC code is “00” and a VLC mismatch occurs according to FIG. 28 described above. table_mi
As shown in FIG. 48B, smatch is set to the “H” state at a corresponding position in the header section.

In this case, macroblock_type
e is a macro representing macroblock_intra
It is replaced by the VLC of obblock_type. For example, in the portion indicated by # 2 in FIG.
obblock_type is replaced with VLC code “1”.

After that, if necessary, dct_type
Is added. For example, in the portion indicated by # 2 in FIG. 48C, dct_type is replaced with “1”. Further, the DCT blocks Y ₁ , Y ₂ , Y ₃ and Y _{4 of} the luminance component are, for example, dct_dc_size_ representing gray display.
Luminance VLC and dct_dc_di
replaced with "ferential" and modified. In the color difference block, DCT blocks Cb ₁ , Cr ₁ ,
Dct_dc in which Cb ₂ and Cr ₂ represent, for example, gray display
_Size_chrominance VLC and d
Replaced with ct_dc_differential and modified. Further, EOB is added to each DCT block, and the macro block is aborted. The modified stream is as shown in FIG. 48C.

(5) macroblock_addr
A case where a VLC mismatch has occurred in the VLC table of ess_increment will be described. In this case, the mb_row and mb of the macroblock
Using the continuity of the _column, the VL from the mb_row and mb_column of the immediately preceding macroblock
Mb_row of macroblock where C mismatch occurred
And mb_column. Mb_ found
macro based on row and mb_column
V of block_address_increment
LC is replaced. Then, macroblock_
For example, type is replaced with VLC “1” of macroblock_type representing macroblock_intra.

After that, if necessary, dct_type
Is added. For example, in the portion indicated by # 2 in FIG. 48C, dct_type is replaced with “1”. Further, the DCT blocks Y ₁ , Y ₂ , Y ₃ and Y _{4 of} the luminance component are, for example, dct_dc_size_ representing gray display.
Luminance VLC and dct_dc_di
replaced with "ferential" and modified. In the color difference block, DCT blocks Cb ₁ , Cr ₁ ,
Dct_dc in which Cb ₂ and Cr ₂ represent, for example, gray display
_Size_chrominance VLC and d
Replaced with ct_dc_differential and modified. Further, EOB is added to each DCT block, and the macro block is aborted. The modified stream is as shown in FIG. 48C.

In the embodiment of the present invention, the conversion stream and the MPEG stream are corrected in this way.
In the above description, it has been described that the stream is corrected by replacing the block using the DCT block of the DC value for performing gray display, but this is not limited to this example. The DC value of the block used for correction is not limited to gray display,
The DC value of the immediately preceding block may be used.
Further, a block for displaying black or red may be used.
Note that the DC value here is d defined in MPEG.
ct_dc_size_luminance and dct_
dc_differential or dct_
dc_size_chrominance and dct_d
Indicates c_differential.

FIG. 49 is a flowchart showing an example of the above-described stream correction processing according to this embodiment. This stream correction process is a process that is completed for each frame. When the frame starts, decoding (VLD) of a variable length code is performed in the first step S10.
The input stream is solved with reference to a predetermined VLC table, and the variable length code is decoded.

In the next step S11, it is determined whether or not there is a mismatch in the VLC table in the header portion before the DCT block in the stream in which the variable length code has been decoded. If it is determined that a mismatch of the VLC table has occurred in the header portion, the process proceeds to step S
Then, the process proceeds to step 12, where the header part is replaced with the prepared header information. Further, the DCT coefficient of the DC component is replaced with, for example, data for displaying gray, and the DCT coefficient is also changed.
EOB is added immediately after the DCT coefficient of the component, and the macroblock is truncated. Then, the process proceeds to step S
Go to step 13.

On the other hand, if it is determined in step S11 that no VLC table mismatch has occurred in the header portion before the block, the process proceeds to step S14. In step S14, it is determined whether or not a VLC table mismatch has occurred at the position of the DCT block of the DC component. If it is determined that a mismatch of the VLC table has occurred at the position of the DCT block of the DC component, the process proceeds to step S15. Step S1
In step 5, the DCT block at the position where the mismatch has occurred in the VLC table and the DCT coefficients of the DC components thereafter are replaced with DCT coefficients for displaying gray. Furthermore, E
OB is added and the macroblock is aborted.
Then, the process proceeds to step S13.

In step S14, the DCT of the DC component
If it is determined that no mismatch in the VLC table has occurred at the coefficient position, the process proceeds to step S16. In step S16, it is determined whether or not a VLC table mismatch has occurred at the position of the DCT coefficient of the AC component. V at the position of the DCT coefficient of the AC component
If it is determined that a mismatch has occurred in the LC table, the process proceeds to step S17. Step S17
Then, D at the position where the mismatch occurs in the VLC table
The CT coefficient and the DCT block following the DCT coefficient are replaced with EOB. Then, the process proceeds to step S13.

On the other hand, in step S16, the DC of the AC component
If it is determined that no VLC table mismatch has occurred at the position of the T block, the process proceeds to step S13.

At step S13, the rearrangement of the DCT coefficients is performed. For example, in the reproduction side MFC 114, the stream is rearranged from the converted stream to the MPEG ES. Similarly, in the recording side MFC 106, the stream is rearranged from the MPEG ES to the converted stream. Then, in step S18, it is determined whether the processing up to the last macroblock of the frame has been completed. If it is determined that the processing up to the last macro block has been completed, a series of flowcharts is completed. On the other hand, if it is determined that the processing up to the last macroblock has not been completed, the processing returns to step S10, and the same processing is performed for the next macroblock.

The above-described processing is performed in the recording MFC 106 and the reproducing MFC 114, respectively. Since the recording side MFC 106 and the reproduction side MFC 114 can be realized with the same configuration, the following description will be made focusing on the reproduction side MFC 114. FIG. 50 shows a configuration example of the reproduction-side MFC 114. The recording MFC 106 and the reproduction MFC 1
14, the configuration can be shared.

At the time of reproduction, the converted stream output from the ECC decoder 113 is input to the reproduction side MFC 114 and supplied to the delay circuit 300 and the detection circuit 301.

Referring to FIG. 50, CPU_IF 310
An interface for controlling communication between the reproduction side MFC 114 and the system controller 121. Various commands and data output from the system controller 121 are stored in the CPU_IF
The signal is supplied to each section of the reproduction-side MFC 114 via 310.

The detection circuit 301 detects the slice start code 12 from the supplied conversion stream. The slice start code 12 is composed of four bytes separated by bytes.
It consists of bytes (32 bits), and the last one byte indicates the vertical position information of the slice, and [00 00
01 01] to [00 00 01 AF]. Therefore, the detection circuit 301 can detect the slice start code 12 by performing pattern matching for each byte, for example. As described above, in this embodiment, since one slice is one macroblock, the start of the macroblock is detected by detecting the slice start code 12.

The detection result of the detection circuit 301 is the signal sli
output as ce_start_code_det,
It is supplied to the timing generator 302. The timing generator 302 outputs the signal slice_start
Signal vld_tim, which is a signal that is reset by _code_det and is repeated for each macroblock
ings and the signal vlc_timings. These signals vld_timings and signal vl
c_timings is the type of block by each of the luminance signals Y ₁ , Y ₂ , Y ₃ and Y ₄ and the color difference signals Cb ₁ , Cr ₁ , Cb ₂ and Cr ₂ constituting the macroblock, and the DCT of each of these blocks It is a control signal indicating the DC component and the AC component in the coefficient, and the type of each header.

[0209] The signal vld_timing is output from the VLD3
03 and the mismatch timing generator 304
Respectively. Also, vlc_timings
Is supplied to a VLC 308 described later.

[0210] In the reproduction side MFC 114, a signal vld_timin indicating the data arrangement of the converted stream is provided.
gs is output from the timing generator 302, and a signal vld_timings indicating the data arrangement of the MPEG stream is output from the timing generator 302 in the recording-side MFC 106.

On the other hand, the converted stream supplied to the delay circuit 300 is given a predetermined delay in order to absorb the detection delay by the detection circuit 301 and the like, and is output after being phase-adjusted. The converted stream output from the delay circuit 300 is supplied to a variable length code decoder (VLD) 303 that decodes a variable length code.

A signal vld instructing the decoding mode of the variable length code from the system controller 121 to the reproduction side MFC 114.
_Settings is supplied. Signal vld_set
tings is the VLD 30 via the CPU_IF 310.
3 is supplied. In the VLD 303, the signal vld_se
The input stream is interpreted in accordance with the mode indicated by tentingsn. That is, the recording side MFC1
In 06, decoding is performed in accordance with a signal vld_timings indicating the data arrangement of the MPEG stream, and in the reproduction side MFC 114, decoding is performed in accordance with the signal vld_timings indicating the data arrangement of the converted stream.

[0213] The VLD 303 has a VLC table that is referred to for solving the variable length code of the input stream. For example, based on the signals vld_timings and vld_settings described above, a VLC table is selected in a predetermined manner, and the variable-length code of the input stream is solved using the selected VLC table.

At this time, it is detected whether or not a mismatch of the VLC table has occurred in the input stream. As a result of the detection, in the VLD 303,
A mismatch signal indicating that a VLC mismatch has occurred is generated as described below according to the position where the VLC mismatch has occurred.

(1) dct_coefficients
When a mismatch occurs in the VLC table of the signal dct_coefficients_mismatc
h is generated.

(2) dct_dc_size_lumi
If a mismatch occurs in the VLC table of the nonce, the signal dct_dc_luminance_mi
Smatch is generated.

(3) dct_dc_chrominan
If a mismatch occurs in the VLC table of ce, the signal dct_dc_chrominance_mi
Smatch is generated.

(4) When a mismatch occurs in the VLC table of macroblock_type, a signal macroblock_type_mismatch is generated.

(5) macroblock_addre
If a mismatch occurs in the VLC table of ss_increment, the signal macroblock_a
address_increment_mismatch
Is generated.

[0220] These signals generated by VLD 303 are supplied to mismatch timing generator 304. In the mismatch timing generator 304,
A signal mi, which is a timing signal indicating a timing for performing stream correction based on these mismatch signals supplied from the VLD 303 and the signal vld_timings supplied from the timing generator 302.
smart_timings is generated. Signal mi
smart_timing is supplied to the selector 306 as a selection control signal.

[0221] In the VLD 303, when a mismatch of the VLC table is detected, the decoding process of the variable length code is immediately stopped. The decoding process is performed by the detection circuit 3
01, the next slice start code is detected and the timing generator 302 is reset,
Will be resumed. The input stream is discarded until the detection circuit 301 detects the next slice start code after the VLD 303 detects a mismatch in the VLC table and stops decoding of the variable length code.

The stream obtained by decoding the variable length code of the converted stream by the VLD 303 is output from the VLD 303, input to the first selection input terminal of the selector 306, and supplied to the Cb / Cr tracing delay circuit 312. You. The output of the Cb / Cr tracing delay circuit 312 is input to the second selection input terminal of the selector 306. Selector 3
06 is connected to a third selection input terminal.
05 (described later) is supplied. The selector 306 outputs the signal mismat described above.
The first, second, and third selection input terminals are selected based on ch_timings, and input signals are switched.

In replacement data generating circuit 305, data for replacing a missing DCT block is prepared in advance. Header data such as a slice header and a macro block header is prepared in advance in the replacement data generation circuit 305. Also, the replacement data generation circuit 305
Include luminance blocks Y _{1 to} Y ₄ and color difference blocks Cb ₁ and Cb ₁ .
Data of DCT blocks of DC components representing, for example, gray display of each of r ₁ , Cb ₂ and Cr ₂ is prepared in advance. Further, the replacement data generation circuit 305 has EOB
Are prepared in advance. These prepared data are stored in a memory included in the replacement data generation circuit 305, for example.

The replacement data generation circuit 305 appropriately selects the previously prepared and stored data according to the signal mismatch_replace supplied via the CPU_IF 310 based on the instruction from the system controller 121 and supplies the data to the selector 306.

Signal mismatch_timings
Is in the “H” state, the third selection input terminal is selected by the selector 306, and the stream output from the VLD 303 is supplied from the replacement data generation circuit 305 to the signal mis.
The data is replaced with the above-described replacement data selected and supplied based on the match_replace.

When there is a mismatch in the color difference blocks Cr ₁ and Cr ₂ , as described above, the stream correction processing is performed by going back to the pair of color difference blocks Cb ₁ and Cb ₂ . At this time, the second and third selection input terminals are selected by the selector 306, and the stream correction processing is performed using the data whose output from the VLD 303 has been delayed by the Cb / Cr tracing delay circuit 312.

The stream output from selector 306 is temporarily written to memory 307 and memory 313. The stream written in the memory 307 is subjected to address control by a variable length encoder (VLC) 308, so that the data arrangement is converted into an MPEG stream and read. The read address of the memory 307 by the VLC 308 is controlled by the system controller 121 from the CPU.
Signal v supplied to the VLC 308 via the _IF 310
lc_settings and timing generator 3
02 supplied from the timing signal vlc_timin
gs

The memory 313 is a memory for delaying the slice header and the macro block header.
If there is a mismatch in the VLC table in these headers, the replacement data generation circuit 305 replaces the headers from the beginning of the headers with a prescribed value prepared in advance, as described in the third method. You. Memory 313 provides a delay for this processing.

The data whose arrangement has been converted to a predetermined value and which has been read from the memory 307 is supplied to the VLC 308. If there is a mismatch in the VLC table in the slice header or macroblock header, the replaced header data is delayed in the memory 313, and the VLC 30
8 is supplied. These data supplied to the VLC 308 are variable-length coded and 8 bits or 1 bit.
The data is bit-aligned into 6 bits and output as MPEG ES.

In the recording side MFC 106, VLD3
03 is supplied with a signal vld_timings indicating a time slot of an MPEG ES data sequence. The MPEG ES supplied to the VLD 303 has a variable-length code decoded based on the vld_timings. In addition, for the VLC 308, a signal vlc_tim indicating a time slot of the data sequence of the conversion stream.
ings is supplied. The stream supplied to the VLC 308 is converted in data sequence based on the signal vlc_timings, converted into a converted stream, and output.

[0231]

As described above, according to the present invention, when decoding a variable-length code, a mismatch of an input stream with respect to a VLC table is detected, and a stream of a stream corresponding to a position where a mismatch occurs in a VLC table is detected. Modifications are made. Therefore, even if an irregular stream that does not exist in the VLC table occurs, the stream can be corrected, and there is an effect that an MPEG syntax error can be avoided.

For example, instantaneous interruption of the stream occurs,
V whose input stream does not exist in the MPEG syntax
Even if the code is changed to an LC code, it is possible to avoid a syntax error of MPEG, and to avoid a situation where a decoder receiving this stream hangs up. Therefore, the present invention is applied to, for example, a VTR for broadcasting
By applying the present invention to a system, there is an effect that a stable system can be realized.

[Brief description of the drawings]

FIG. 1 is a schematic diagram schematically showing a hierarchical structure of MPEG2 data.

FIG. 2 is a schematic diagram showing the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 3 is a schematic diagram showing the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 4 is a schematic diagram showing the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 5 is a schematic diagram showing the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 6 is a schematic diagram showing the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 7 is a schematic diagram showing the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 8 is a schematic diagram showing the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 9 is a schematic diagram showing the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 10 is a schematic diagram illustrating the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 11 is a schematic diagram showing the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 12 is a schematic diagram showing the contents and bit allocation of data arranged in an MPEG2 stream.

FIG. 13 is a diagram illustrating alignment of data in byte units.

FIG. 14 is a schematic diagram specifically illustrating an MPEG stream header according to an embodiment.

FIG. 15 is a block diagram illustrating an example of a configuration of a recording / reproducing device according to an embodiment.

FIG. 16 is a schematic diagram illustrating an example of a track format formed on a magnetic tape.

FIG. 17 is a schematic diagram for explaining a chroma format.

FIG. 18 is a schematic diagram illustrating a chroma format.

FIG. 19 is a schematic diagram illustrating a chroma format.

FIG. 20 is a schematic diagram for explaining an output method of a video encoder and variable-length encoding.

FIG. 21 is a schematic diagram for explaining rearrangement of an output order of a video encoder.

FIG. 22 is a schematic diagram for explaining a process of packing data in a rearranged order into a sync block.

FIG. 23 is a schematic diagram for describing effects of coefficient rearrangement and packing.

FIG. 24 is a schematic diagram for describing effects of coefficient rearrangement and packing.

FIG. 25 is a block diagram showing a more specific configuration of the ECC encoder.

FIG. 26 is a schematic diagram illustrating an example of an address configuration of a main memory.

FIG. 27: macroblock_address_i
FIG. 14 is a schematic diagram illustrating a VLC table for ncrement.

FIG. 28 is macroblock_type of an I picture.
FIG. 14 is a schematic diagram illustrating a VLC table for e.

FIG. 29 is macroblock_type of a P picture
FIG. 14 is a schematic diagram illustrating a VLC table for e.

FIG. 30: dct_dc_size_luminanc
FIG. 14 is a schematic diagram illustrating a VLC table for e.

FIG. 31: dct_dc_size_chromina
FIG. 14 is a schematic diagram illustrating a VLC table for the nce.

FIG. 32: DCT coefficients Tab
FIG. 7 is a schematic diagram illustrating a VLC table for le 0.

FIG. 33: DCT coefficients Tab
FIG. 7 is a schematic diagram illustrating a VLC table for le 0.

FIG. 34: DCT coefficients Tab
FIG. 7 is a schematic diagram illustrating a VLC table for le 0.

FIG. 35: DCT coefficients Tab
FIG. 7 is a schematic diagram illustrating a VLC table for le 0.

FIG. 36: DCT coefficients Tab
FIG. 4 is a schematic diagram illustrating a VLC table for le 1;

FIG. 37. DCT coefficients Tab
FIG. 4 is a schematic diagram illustrating a VLC table for le 1;

FIG. 38: DCT coefficients Tab
FIG. 4 is a schematic diagram illustrating a VLC table for le 1;

FIG. 39. DCT coefficients Tab
FIG. 4 is a schematic diagram illustrating a VLC table for le 1;

FIG. 40 is a schematic diagram showing a table of fixed-length codes.

FIG. 41 is a schematic diagram illustrating a table of fixed-length codes.

FIG. 42 is a schematic diagram for explaining a case where a VLC that does not exist in the VLC table is encountered and a VLC table mismatch occurs.

FIG. 43: dct_coef in case of a conversion stream
FIG. 14 is a schematic diagram illustrating an example in which a mismatch has occurred with respect to the VLC table of the employees.

FIG. 44: dct_dc_s in case of a conversion stream
size_luminance and dct_dc_si
FIG. 14 is a schematic diagram illustrating an example in which a mismatch has occurred in a VLC table of ze_chrominance.

FIG. 45 is a schematic diagram illustrating an example where a VLC mismatch exists in a slice header and a macroblock header in the case of a converted stream.

FIG. 46: dct_coeff in the case of MPEG ES
FIG. 11 is a schematic diagram illustrating an example in which a mismatch has occurred with respect to a VLC table of the clients.

FIG. 47: dct_dc_si in the case of MPEG ES
FIG. 14 is a schematic diagram illustrating an example in which a mismatch has occurred in a VLC table of ze_chrominance.

FIG. 48 is a schematic diagram showing an example of a case where a VLC mismatch exists in a slice header and a macroblock header in the case of MPEG ES.

FIG. 49 is a flowchart illustrating an example of a stream correction process according to an embodiment;

FIG. 50 is a block diagram showing a configuration example of a reproduction-side MFC 114.

FIG. 51 is a schematic diagram for explaining that the VLC becomes unreliable until the next header is detected due to the occurrence of an error.

[Explanation of symbols]

1 ... sequence header code, 2 ... sequence header, 3 ... sequence extension, 4 ... extension and user data, 5 ... GOP start code, 6 ...
GOP header, 7: user data, 8: picture start code, 9: picture header, 10
..Picture coding extension, 11 ... extension and user data, 12 ... slice start code, 13 ...
・ Slice header, 14 ... Macro block header,
101: SDI receiver, 102: MPEG encoder, 106: Multi-format converter (MFC) on recording side, 108: SDTI receiver, 109
... ECC encoder, 112 ... Magnetic tape, 1
13: ECC decoder, 114: reproduction side MF
C, 115: SDTI output unit, 116: MPE
G decoder, 118... SDI output unit, 137a, 1
37c: packing unit, 137b: video shuffling unit, 139: outer code encoder, 140
..Video shuffling, 149 ... inner code encoder, 170 ... frame memory, 301 ... detection circuit, 302 ... timing generator, 303 ...
VLD, 304: mismatch timing generator, 305: replacement data generation circuit, 306
・ Selector, 307 ・ ・ ・ Memory, 308 ・ ・ ・ VLC

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Susumu Todo 6-35, Kita-Shinagawa, Shinagawa-ku, Tokyo Inside Sony Corporation (72) Inventor Hideyuki Matsumoto 6-35, Kita-Shinagawa, Shinagawa-ku, Tokyo Sony Corporation F-term (reference)

Claims (9)

[Claims] 1. A data processing apparatus for processing a variable-length coded stream, comprising: detecting a code that does not correspond to a parameter representing the variable-length code from a variable-length code of the variable-length coded stream. Means for correcting the stream in accordance with the content of the detection result by the detection means; and instruction means for instructing the start of the next processing after the detection by the detection means. apparatus. 2. The data processing device according to claim 1, wherein the stream is subjected to the variable-length coding in units of a predetermined block, and the correction unit sets the code indicating the position of the block by the detection unit to the parameter. If it is detected that the block does not correspond to the above, the correction is performed using a code representing the position of the block before the block. 3. The data processing apparatus according to claim 1, wherein the stream is subjected to the variable-length coding in units of a predetermined block, and the correction unit uses a code representing the coding type of the block by the detection unit. A data processing apparatus, wherein when it is detected that the block does not correspond to the parameter, the correction is performed using a parameter indicating that the block is completed by itself. 4. The data processing device according to claim 1, wherein the stream is subjected to the variable-length coding in a predetermined block unit, and the correction unit sets a code indicating a code length of a luminance component by the detection unit. A data processing apparatus characterized in that, when it is detected that the parameter does not correspond to a parameter, the code that does not correspond to the parameter is replaced with a code indicating a DC component of a predetermined value of luminance. 5. The data processing apparatus according to claim 1, wherein the stream is subjected to the variable-length coding in a predetermined block unit, and the correction unit sets a code indicating a code length of a color difference component by the detection unit. A data processing device, wherein when it is detected that the parameter does not correspond to a parameter, the code that does not correspond to the parameter is replaced with a code indicating a DC component of a predetermined value of color difference. 6. The data processing device according to claim 5, wherein the correction unit is configured to change a code indicating a code length of the color difference component of a pair of color difference components transmitted later in time among the pair of color difference components. A data processing device for performing the above-mentioned replacement by going back to a color difference component paired with the color difference component when it is detected that the color difference component does not correspond to the parameter. 7. The data processing apparatus according to claim 1, wherein the stream is subjected to the variable-length coding in units of a predetermined block, and the correction unit uses a code indicating a quantization coefficient in the block by the detection unit. A data processing device that, when it is detected that does not correspond to the parameter, replaces the code that does not correspond to the parameter with a code indicating the end of the block. 8. The data processing apparatus according to claim 1, wherein the stream is a stream based on MPEG. 9. A data processing method for processing a variable length coded stream, wherein a variable length code of a variable length coded stream that does not match a parameter representing the variable length code is detected. A detection step; a modification step of modifying the stream according to the content of the detection result of the detection step; and an instruction step of instructing the start of the next processing after the detection by the detection step. A data processing method characterized by the following.