JPH06303572A - Image data recorder - Google Patents

Abstract

PURPOSE:To obtain an image data recorder capable of performing recording in which the deterioration of picture quality on a screen when reproduction is performed can be prevented from being prominent even when the compression of image data is performed. CONSTITUTION:Input image data from an input terminal 10 is supplied to memory 115 via a data input/output interface 111 in a shuffling part 11. The output of a write address generating part 113 and that of a readout address generating part 114 are supplied to the memory 115 via an address data switching selection part 112. An address switching selection signal supplied via an input terminal 18 is supplied to the data input/output interface 111 and the address data switching selection part 112, which controls the data flow of each part. Output image data in which shuffling is applied is outputted via the data input/output interface 111 and an output terminal 19 after receiving such control.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data recording apparatus for shuffling and recording an input digital data, particularly an orthogonal transform block obtained by sub-sampling image data.

[0002]

2. Description of the Related Art As a conventional digital VTR,
A D-1 format VTR based on a so-called 4: 2: 2 component coding system, a D-2 format VTR based on a system for digitizing and recording a composite video signal of the current broadcast, and a 1/2 inch tape. D based on a method of digitally recording the composite video signal using
There are -3 format VTRs and the like. All of these digital VTRs have error correction (hereinafter referred to as ECC).
Signals are recorded / reproduced while processing.

However, in the digital VTR, the presence of an error may be detected but may not be corrected when the ECC process is performed. In order to prevent such error correction and to suppress image quality deterioration due to, for example, the concentration of corrected pixels on the screen, the digital VTR performs so-called shuffling processing such as switching the order of image code generation and the order of recording code. Is going.

Actually, as the shuffling process of the digital VTR, a process of placing adjacent samples at distant positions on the tape is performed by placing importance on concealment with respect to head clogs and scratches on the tape. . By performing such a corresponding process, the digital VTR does not concentrate the code error and makes the image quality deterioration inconspicuous.

In addition, the digital VTR divides the input video signal into a digital image code and divides it into a plurality of channels included in the digital VTR to reduce the bit rate per channel, and also to prevent errors from concentrating in a certain screen. In addition, consideration is given to evenly distributing the image code corresponding to pixels to each channel.

[0006]

By the way, a user is often required to have a component signal as an input / output signal of a digital VTR. However, the D-1 format digital VTR capable of handling component signals is a very expensive device. Therefore, the user needs to use the conventional VTR to improve the cost performance.
There is a strong demand for low-priced digital VTRs that can reproduce cassettes. Further, the user also desires to reduce the size of the tape cassette used and increase the recording time. In order to satisfy these user requirements, it becomes necessary for the digital VTR to compress the input video signal into digital image data and perform compression.

However, when image data is block-coded by orthogonal transform in a digital VTR, for example, discrete cosine transform (hereinafter referred to as DCT) to compress and record the data, the digital VTR performs image compression by performing data compression. The code error of appears as a block error on the screen. The block error on this screen is
Since the influence is exerted over a wide area, the deterioration of image quality becomes conspicuous.

Therefore, the present invention has been made in view of the above situation, and an image can be recorded in which deterioration of the image quality on the screen during reproduction is not noticeable even when the image data is compressed. The purpose is to provide a data recording device.

[0009]

According to an image data recording apparatus of the present invention, a plurality of orthogonal transform blocks are constructed by data obtained by sub-sampling input image data, and after the orthogonal transform blocks are shuffled. An image data recording device for performing data compression by performing an orthogonal transformation process, by recording the orthogonal transformation blocks of the overlapping area at the time of the sub-sampling corresponding to different head channels of the rotary head, The above problems are solved.

Here, the fact that the concealment is possible by the sub-sampling means that even if there is an error in some sampling image data when sub-sampling, there is no error in the left and right sample data. In this image data recording apparatus, the orthogonal transformation block by sub-sampling is composed of 4 × 4 image data, and the orthogonal transformation block is shuffled. An example of the orthogonal transform block is a discrete cosine transform. In this image data recording device, one field is composed of 6 tracks. Further, the image data recording apparatus uses at least four heads when recording / reproducing via the magnetic head.

In the image data recording apparatus, the orthogonal transformation blocks in the overlapping area are separately recorded in the first half and second half sectors of the track. In the orthogonal transform block, the orthogonal transform blocks arranged in the vertical direction are recorded on the same track, and the orthogonal transform blocks adjacent in the vertical direction are recorded on the same track.

Here, the sector of one track has a format structure having at least two sectors.

In the image data recording apparatus, the data recorded on the same channel of the head channel is positioned on the left of the orthogonal transform block used for recording the one track on the track next to the one track. Orthogonal transform blocks are recorded.

Further, one sync block in which a plurality of the orthogonal transform blocks are collected is treated as one equal length unit. Further, in this image data recording apparatus, two color difference signals forming a color difference signal by recording the orthogonal transformation block of the luminance signal and the orthogonal transformation block of the color difference signal corresponding to the same pixel on the screen in different sync blocks on the tape, respectively. The orthogonal transform blocks of are recorded in the same sync block.

Here, the one sync block is composed of 20 orthogonal transform blocks. One sector is formed by collecting 114 sync blocks. Also,
The number of orthogonal transform blocks in one field is (114 × 2 × 6) because one track is divided into upper and lower two sectors and six tracks form one field.

[0016]

The image data recording apparatus according to the present invention forms a plurality of orthogonal transform blocks with data obtained by sub-sampling input image data, and performs shuffle processing on the orthogonal transform blocks by orthogonal transform, for example, By performing the discrete cosine transform before performing the discrete cosine transform, for example, the number of data bits is 16 bits, which is required after the discrete cosine transform process, to perform processing with 10 bits to reduce data bits. Then, the image data recording device is
The plurality of orthogonal transformation blocks obtained by subsampling the image data in the overlapping area by the subsampling are recorded corresponding to different head channels of the rotary head, and the magnetic head used for one channel fails. Even if occurs, an error between fields is compensated for through the magnetic head of the channel on the other side.

Further, by recording the orthogonal transform block obtained by the sub-sampling separately in a plurality of sectors constituting one track, even if one of the sectors has an abnormality, a sector having no abnormality is recorded. Conceal using the data of the orthogonal transformation block.

By recording on the same track vertically aligned in the orthogonal transform block, the same track in the next field is recorded by different magnetic heads, for example. Therefore, even if one of the magnetic heads is damaged, the other magnetic head is used to perform concealment between fields.

Burst errors are dealt with by recording the orthogonally-adjacent orthogonally-transformed blocks divided into a plurality of sectors provided along the track for each block.

Further, recording with the same head is performed as the same channel recording, and when recording with this same channel, an orthogonal transformation block located to the left of the currently recorded data block is recorded on the next track, and orthogonal transformation is performed. Block samples are separated so that they are not next to each other.

1 which is a collection of a plurality of the orthogonal transform blocks
By treating the sync block as one equal length unit, the orthogonal transform block affected by the sync error is reduced.

The orthogonal transformation block of the luminance signal and the orthogonal transformation block of the color difference signal corresponding to the same pixel on the screen are respectively recorded in different sync blocks on the tape, and the equalization efficiency is increased as different equalization units, Even if the orthogonal conversion blocks of the two color difference signals forming the color difference signal are recorded in the same sync block and the color difference signals are arranged between the fields, it is possible to prevent the color from being different from the original color.

[0023]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the image data recording apparatus according to the present invention will be described below with reference to the drawings.

FIG. 1 shows a schematic structure of an image data recording / reproducing apparatus in this embodiment. The image data recording / reproducing apparatus shown in FIG.
Discrete cosine transform (DCT) unit 12, variable length coding unit 1
3, an error correction encoding unit 14, a recording unit 15, and a rotary head 16. Further, since the reproduction system performs the reverse processing of the recording system, the rotary head 16, the reproduction unit 21, the error correction decoding unit 22, the variable length decoding unit 23,
An inverse discrete cosine transform (hereinafter referred to as IDCT) unit 24,
It is composed of the deshuffling portion 25.

The operation of the image data recording / reproducing apparatus will be briefly described with reference to FIG. In the image data recording / reproducing apparatus, a digital input image signal is supplied to the shuffling portion 11 via the input terminal 10. Since the shuffling unit 11 performs a discrete cosine transform, which is one of orthogonal transforms arranged in the subsequent stage, it sub-samples a digital input image signal and handles it as a block of data, that is, a DCT block. is doing. The shuffling unit 11 uses the DCT block as a unit for performing shuffling processing.

The DCT unit 12 performs DCT conversion processing and outputs image data to the variable length coding unit 13. The variable length coding unit 13 supplies the image data that has been subjected to variable length coding according to the compression rate to the error correction encoding unit 14. The error correction encoding unit 14 adds a code for error correction and outputs it to the recording unit 15. Image data that has been subjected to various kinds of signal processing in the recording unit 15 is supplied to the rotary head 16, and the rotary head 16 receives the image data.
It is recorded on the magnetic tape 17 via the.

At the time of reproduction, the image data recording / reproducing apparatus performs the reverse processing at the time of recording the image data read from the magnetic tape 17 via the rotary head 16, so that the reproducing unit 2
1, error correction decoding unit 22, variable length decoding unit 23, I
The signal is output from the output terminal 26 via the DTCT section 24 and the deshuffling section 25.

A schematic block configuration of the shuffling portion 11 in this image data recording / reproducing apparatus will be described with reference to FIG. In FIG. 2, the digital input image signal is used as input image data via the input terminal 10 and the data input / output interface unit 1 of the shuffling unit 11.
11 are being supplied. Further, an address switching selection signal is supplied to the data input / output interface section 111 and the address data switching selection section 112 via the input terminal 18. The shuffling unit 11 includes a write address generation unit 113, a read address generation unit 114, and a memory 115.
have. Data input / output interface unit 111
Controls the input / output of the image data to the memory 115 or the output of the image data from the memory 115 described later by the address switching selection signal. Memory 1
Reference numeral 15 stores the image data supplied via the data input / output interface unit 111 at a write address position described later.

Here, the image data supplied to the memory 115 is written in units of time of the vertical synchronizing signal (V) and the horizontal synchronizing signal (H), for example.

The shuffling of the input image data is performed via the memory 115. Therefore, the write address generation unit 113 and the read address generation unit 114
Generates write address data and read address data for realizing shuffling, which will be described later. The write address generation unit 113 and the read address generation unit 114 output the generated write address data and read address data to the address data switching selection unit 112. The address data switching selection unit 112 supplies one of the write address data and the read address data to the memory 115 according to an address switching selection signal for performing shuffling according to the format. At this time, in the memory 115, the supplied address is in the enabled state.

The memory 115 outputs the data stored at the supplied read address position as output image data via the data input / output interface section 111 and the output terminal 19. The output image data is read out in the order in which it is recorded on the tape. The memory 115
Requires the capacity for two fields for efficient processing using the video synchronization signal as described above. With such a configuration, the system delay of the shuffling portion 11 becomes about 1 field.

Here, in the image data recording apparatus, as shown in FIG.
3 schematically shows the respective head positions of the rotary head 16 shown in FIG.
Shown in. The rotary head 16 includes four magnetic heads A, B,
It has C and D. The magnetic heads A and C are arranged at positions facing each other by 180 ° on the first channel CH1. The magnetic heads B and D are magnetic heads A and C, respectively.
Corresponding to the second channel CH2, and the magnetic heads B and D are also arranged at positions facing each other by 180 °. Adjacent heads have azimuth with respect to each other.

As shown in FIG. 4, this image data recording apparatus employs a recording system that uses 6 tracks per field. Therefore, when recording and reproducing with the four magnetic heads shown in FIG. 3, the image data recording apparatus uses the A head to record the first track in the N-1th field, which is one field before the Nth field. For recording on one track,
The C head will be used. In this way, recording on the corresponding track for each field is performed by different magnetic heads. Therefore, even if one of the two magnetic heads that perform one-channel recording fails, the other magnetic head can be used to perform the conceal between fields.

Next, subsampling in this image data recording apparatus will be described with reference to FIG.
This image data recording device is supplied with image data corresponding to pixels on the screen by sampling an analog input video signal as shown in FIG. In the supplied image data, for example, a 4 × 8 pixel area D _C is treated as an area where subsamples overlap. Subsampling is
In this case, the pixel data of already sampled pixels is a process of sampling pixels of symbols in which symbols ◯ and × are alternately arranged, for example. By this sub-sampling process, the image data of FIG.
For example, a DCT block is composed of 4 × 4 image data of only the symbols ◯ or × shown in FIGS. 5B and 5C. That is, in the image data of FIG. 5A, each image data is represented by the data of two systems of symbol ◯ or × in the overlapping area. The 4 × 4 image data is shuffled in this unit as a unit of the DCT block described above. The discrete cosine transform process is performed after the shuffling process is performed for each DCT block. The discrete cosine transform process is also performed in DCT block units.

Here, the reason why the shuffling process is performed before the discrete cosine transform process will be briefly described. If the image data before the DCT conversion process is image data consisting of 10 bits, the image data after the DCT conversion process needs about 16 bits, which increases the number of bits. The circuit configuration that performs the shuffling process on the increased image data after the DCT conversion process has a larger circuit scale than the circuit configuration that performs the shuffling process before the discrete cosine transform process described above. In fact, in the configuration in which the shuffling process is performed before the DCT conversion process as in the present embodiment, it is possible to reduce the number of bits of the memory 115 of FIG. become. Further, even if an error occurs in the DCT block, the error pixel data can be reproduced from the pixel data of the left and right adjacent samples recorded in different DCT blocks, which enables concealment.

The principle of the shuffling process used in the image data recording apparatus thus configured will be sequentially described with reference to the drawings. First, in the present embodiment, the image data recording apparatus is the magnetic tape 17 as described above.
In order to reduce the bit rate for recording, the rotary head 16 having a plurality of magnetic heads is used for recording. The rotary head 16 has the four magnetic heads A, B, C and D shown in FIG. Since the magnetic heads A and C are the first channel CH1 and the magnetic heads B and D are the second channel CH2 as described above, the relationship between these magnetic heads and recording channels is 6 tracks per field. When the recording method used is adopted, the relationship shown in FIG. 6 is obtained. That is, the first track of the field is the first track depending on the setting of the rotary head 16.
When recorded by the magnetic head A of the channel CH1 of
Image data is recorded on the second track of the second channel of the magnetic head B having different azimuths. As is clear from the recording order on each track shown in FIG. 6, it is understood that the odd-numbered tracks are recorded as the first channel CH1 and the even-numbered tracks are recorded as the second channel CH2.

As described above, the image data recording apparatus associates the plurality of DCT blocks obtained by subsampling the image data in the overlapped area at the time of subsampling with different head channels of the rotary head. Is recorded.

The image data of the area surrounded by the broken line in FIG. 5A is the data in the overlapping area composed of, for example, 4 × 8 symbols ◯ and ×. Shuffling part 1
1 sub-samples this image data, and FIG.
DCT blocks divided into 4 × 4 pixels of (b) and (c) for each symbol are formed and recorded in different head channels.

Here, for example, if there is a damaged magnetic head and recording or reproduction on a corresponding track cannot be performed, but an adjacent magnetic head is normal,
Even if there is no 4 × 4 image data in FIG.
It is possible to conceal the image data of FIG. 5B by arithmetic processing such as taking an average from the 4 × 4 image data of (c).

This is because the DCT blocks formed by sub-sampling the overlapping areas by sub-sampling are recorded on each track in association with different head channels of the rotary head 16.
It enables concealment by subsampling.

Another principle of shuffling performed by the shuffling section 11 is that the DCT block obtained by sub-sampling the overlapping area as described above is divided into the first half section and the second half section of the track. To be recorded.

This principle will be described with reference to FIGS. 5 and 7. An area including the symbols ◯ and × in FIG. 5A, that is, an overlapping area D _C is shown in FIGS. 5C and 5B.
As shown in FIG. 7, the DCT blocks DS1 and DS2 are recorded separately in the first half (or lower) side sector S _A1 and the second half (or upper side) sector S _A2 of the track. By doing so, even if, for example, the tape is scratched in the longitudinal direction or one of the sectors is not reproduced due to a burst, the recorded DCT block on the normal side has a defect D.
The CT block data can be concealed.

Another principle of shuffling, which will be described below, is that the DCT blocks are arranged in the vertical direction in the D direction.
The CT block is recorded on the same track.
As described above, the data between fields are always recorded by different magnetic heads. This relationship uses different magnetic heads for recording on each track of the first field and the second field of the Nth frame, for example, as shown in FIG. In an actual digital VTR,
Although a specific example will be described later, the DCT block includes a first channel CH1 and a second channel CH2.
Each DCT block is labeled with a label number, and each group consists of 12 groups.

The shuffling portion 11 has the second field of the Nth frame even if the data in the previous field, that is, the first field is damaged due to the damage of the magnetic head.
Since the magnetic head of the field is recorded on each track by a magnetic head different from that of the first field, the damaged DCT block can be concealed.

Further, the shuffling section 11 records the DCT blocks arranged in the vertical direction on the same track, so that the shuffling section 11 can be arranged on every other track in the vertical direction in the same channel of the head channel. For each array, the label number of the DCT block is "L1, L2, L1, L2, ...", "L5,"
L6, L5, L6, ... "," L3, L4, L3, L
4 ... "are arranged repeatedly. The reason for this will be described in detail later. Therefore, the DCT blocks arranged in the vertical direction have the same magnetic head in the same head channel. Will be recorded by.
In the array format of FIG. 8, for example, in the channel CH1, recording of the DCT blocks CH1-L1 and CH1-L2 is performed by the magnetic head A, and recording of CH1-L5 and CH1-L6 is performed by the magnetic heads C, CH1-L3, and CH1-L4. Is set in the magnetic head A, and the same channel is recorded in the vertical direction.

However, D written in the sector S _{A2 of} FIG.
DCT block C instead of CT block CH1-L2
When H1-L4 is recorded, the magnetic head used must use a magnetic head C different from the magnetic head A. When the magnetic head is "damaged and an error occurs near the boundary of the vertical DCT block, it is difficult to know whether the data necessary for concealment by both magnetic heads A and C is surely recorded.

Therefore, as described above, by setting the DCT block array in which recording is performed on each sector in the same track by the magnetic head of the same channel, that is, by setting the DCT block array in the vertical direction, the magnetic head of one channel is set. The recording state in one track can be determined only by examining. As a result, for example, even if an abnormality occurs in the magnetic head A and the first track of the first field cannot be recorded in the Nth frame, the magnetic head C of which the first track of the second field is different. Recording is performed, and concealing can be performed using the DCT blocks recorded in the two sectors provided in the first half and the second half of this track.

By setting in this way, it is possible to conceal errors occurring in DCT blocks between fields and errors near sector boundaries.

The operation of the shuffling portion 11 as described above will be described more specifically.
DCT blocks (L1, L2) and (L
5, L6) and (L3, L4) are recorded as a pair in two sectors of the track. The DCT blocks shown in FIG. 9 are alternately arranged in the vertical direction. The DCT blocks are labeled and classified, and the DCT blocks with the same label on different channels are recorded in different sectors. In this way, the DCT block having the same label is divided into the sectors S _A1 and S _A2 which divide the track (see FIG. 14 shown later), so that the sector on one side is damaged. Not only can the scratches in the direction be dealt with, but also the concealment area for burst errors can be dispersed.

Another shuffling principle described below is as follows.
This is a method of recording a DCT block located on the left side of the DCT block used for recording the above-mentioned one track on the track next to one track for the data recorded on the same channel of the head channel. FIG. 10 shows the recording position relationship regarding the shuffling principle.
This will be described with reference to the schematic diagram of FIG.

In FIG. 10, the DCT block array in the horizontal direction is arrayed as "..., L5, L1, L3, L5, L1". The first DCT block starts at label L1. Further, as is apparent from FIG. 10, the track numbers are “1, 3, 5” for recording on the same channel.
Is selected. Recording is performed on each track using the magnetic heads A and C, respectively. In this way, the DCT blocks are arranged in the horizontal direction according to the track numbers in the field.

When actually recording on the same channel, FIG.
On the first track of 0, the data of the DCT block with label L1 is recorded by the magnetic head A. On the next third track, D of label L5 located to the left of label L1
The data of the CT block is recorded by the magnetic head C. Further, on the fifth track, the data of the DCT block of the label L3 located on the left side of the label L5 is recorded by the magnetic head A. In this way, when recording on the same channel, the DCT block located to the left of the previous DCT block is recorded next.

The reason for using this DCT block array will be described with reference to FIG. Here, the overlapping area due to the first sampling is (4 × 8)
It is composed of individual image data. The image data recording device subsamples the (4 × 8) DCT blocks for each column and divides the DCT blocks into symbols (∘) and (×) in FIG. 11 to form DCT blocks. The DCT block of the data (◯) obtained by sub-sampling indicates the channel CH1. The DCT block of the data (×) obtained by subsampling is the channel CH.
2 is shown. For example, in FIG. 11A, the DCT blocks are arranged in the order of labels L3 and L5.

If there is an error in the image data (fifth track) indicated by the symbol () near the boundary within the DCT block, the image data recording device determines the channels located on both sides of the image data (). DC different from CH1
The data can be reproduced by averaging and concealing the image data (x) of the T block CH2-L3 and the image data (x) of the DCT block CH2-L5. These DCT
The data (x) of the block CH2-L3 and the data (x) of the DCT block CH2-L5 are recorded on the fourth track and the sixth track, respectively, as is apparent from FIG. Here, the rotary head 16 is provided with two sets of magnetic heads, that is, the set of magnetic heads A and B and the set of C and D adjacent to each other, as described above.

However, when a head clog or the like occurs instantly and a defect occurs in the pair of magnetic heads, for example, in the above-mentioned case, the DCT block CH1-L5 of the fifth track and the DCT block CH2 of the sixth track.
-The two tracks in which L5 are arranged side by side, the magnetic head C,
Playback will not be possible with the D group. Therefore, the sixth track cannot be reproduced. In order to avoid such a problem, it is understood that concealment should be performed using tracks at positions not adjacent to the two tracks even if the pair cannot be reproduced by a head clog or the like. Therefore, the shuffling setting may be performed by setting the arrangement of the DCT blocks in the order of “L1, L5, L3” as described above so that the DCT blocks to be recorded are not adjacent to each other.

Even if the image data (fifth track) in the boundary area actually has an error like the image data (●), the second track shown in FIG. 10 is used without using the track using the pair of magnetic heads. DCT block CH of track
2-L1 and the DCT block CH2-L of the fourth track
5 and 5 can be used for concealing. If all the shuffling principles described above are satisfied, the image data and D
The relationship of CT blocks and the relationship of recorded sectors are obtained. An example of this specific shuffling process will be described in detail later.

Further, the image data recording apparatus records an audio signal and a video signal using a common format called a sync block (not shown). This one sync block is configured by using 20 DCT blocks. In addition, one sector is composed of 114 sync blocks. Further, since one field is divided into two sectors and one track is composed of six tracks, one field is composed of (114 × 2 × 6) = 1368 sync blocks for buffering.

In this image data recording apparatus, the variable length coding unit 13 performs variable length coding according to the compression rate after performing the DCT conversion processing as described above.
By this variable length encoding process, the image data is equalized in length. If a sync error occurs as a result of this processing, the effect of this error appears in the DCT block. So-called error countermeasures are required to reduce the number of DCT blocks affected by this. Furthermore, it is also necessary to take measures against errors in special reproduction of shuttle images, which will be described later.
These error countermeasures are also a major factor in determining the shuffling method.

Therefore, as a measure against the image quality deterioration due to the sync error among the above-mentioned errors, the image data recording / reproducing apparatus has a DCT block in which the screen is subdivided in advance without equalizing the entire screen of a unit such as one field. A plurality of sync blocks are treated as one equal length unit.

Therefore, as described above, one sync block is treated as one equal length unit, and the number of data blocks of the luminance signal Y and the color difference signals (RY) and (BY) is made equal. . By performing this operation, when one sync block has a sync error, the screen is affected only by the sync block area. Therefore, by using this method, the image data recording apparatus can suppress the DCT block having an error due to the equal length and reduce the area where the image quality deteriorates. Furthermore, this method can also perform on-screen dispersion for the disappearance of consecutive DCT blocks.

Further, the image data recording apparatus has a so-called shuttle reproduction function, which is one of the special reproductions, in which the head is moved in the longitudinal direction to perform variable speed reproduction. Since the adjacent DCT blocks are arranged side by side during the shuttle reproduction, when scratches occur in the longitudinal direction of the magnetic tape, the scratches deteriorate the image quality only in a certain area of the screen. That is, scratches in the longitudinal direction concentrate the places where errors occur. In order to avoid this, the shuffling portion 11 has an offset in the recording order for each sector. This operation will disperse the positions of the data blocks recorded on the tape. In order to disperse the DCT blocks at equal intervals, the shuffling unit 11 gives an offset for each sector shown in Table 1.

[0062]

[Table 1]

At this time, the values in Table 1 are set so as to maximize the distance between overlapping data blocks in consideration of the burst error in the longitudinal direction. As is clear from Table 1, sectors 1A and 1B and sectors 2A and 2B are separated by 114 blocks, that is, by one sector. By making the settings as described above, it is possible to increase the image update rate during so-called shuttle reproduction and to make the device resistant to scratches in the longitudinal direction.

Further, the shuffling section 11 in the image data recording apparatus records the DCT block of the luminance signal and the DCT block of the color difference signal corresponding to the same pixel on the screen in different sync blocks on the tape to form the color difference signal. The DCT blocks of the two color difference signals are recorded in the same sync block.

For example, FIG. 12 shows a so-called digital VT.
FIG. 6 is a diagram showing that one pixel (pixel) is represented by a luminance signal Y, color difference signals (RY), and (BY) in the standard of 4: 2: 2 of R. According to the above standard, the number of pixels of the luminance signal Y of 360 pixels is such that the color difference signals (RY) and (B-
The number of pixels in Y), 180 pixels, has a double relationship. The luminance signal and the color difference signal are recorded in different sync blocks. On the other hand, there are two color difference signals (RY),
(BY) is recorded in the same sync block.

Generally, the luminance signal Y and the color difference signal (R-
Since Y) and (B−Y) have a correlation with each other, it is desirable that the amount of information included in the equalization unit configured to increase the efficiency of the above-described equalization over the entire screen is approximately the same. Therefore, the DCT blocks of the luminance signal Y and the color difference signals (RY) and (BY) including the samples at the same position are recorded in different sync blocks as different units of equal length.

Two color difference signals (RY) and (B-
Y) has a correlation, but when so-called shuttle reproduction is performed, for example, the color difference signals between the fields are mixed, which may cause a difference from the original color displayed on the screen.

In this way, two color difference signals (RY),
By recording (BY) in the same sync block, it is possible to prevent abnormal coloring and display the original color screen even in so-called shuttle reproduction.

A more specific embodiment of the shuffling in the image data recording device based on the above-mentioned setting will be described with reference to FIGS. 13 to 16. Here, the Recommendation 601 (CC
An example will be given of the case where it is applied to a digital VTR system according to the PAL system (625/50) of the IR REC.601) standard. The basic structure of the digital VTR is the block shown in FIG. 1, and common portions are given the same reference numerals.

This digital VTR system has a luminance signal Y, color difference signals (RY), (B) defined by the above standards.
-Y) so-called 4: 2: 2 signal is satisfied. FIG.
The pixels of the luminance signal Y shown in (a) are 304 × 720 pixels,
The pixels of the color difference signals (RY) and (BY) are 304 × 36.
The screen is composed of 0 pixels. Here, the number in parentheses indicates the number of pixels of the color difference signal.

The shuffling section 11 in the digital VTR system sub-samples the pixels in the screen for each column by using the screen areas shown in FIG. 13A as overlapping areas. The pixel indicated by the symbol (◯) in FIG. 13A is the image data of the channel CH1, the symbol (×).
The pixel indicated by indicates the image data of the channel CH2. By sub-sampling the overlapping area of FIG. 13A, each channel is classified into 304 × 360 pixels of a luminance signal and 304 × 180 pixels of a color difference signal as shown in FIG. 13B. In this way, screen pixels are sub-sampled and channel distribution is performed. In the shuffling unit 11 in the digital VTR system, the pixels of the luminance signal and the color difference signals are combined into one block of image data of 4 × 4 to form a DCT block. In FIG. 13 (c), the screens of the channels CH1 and CH2 are 76
× 90 blocks, 76 blocks × 45 in the color difference signal
The block is shown as a block.

The DCT blocks for both channels are shown in FIG.
Labels L shown in (c) are attached in order from the upper left corner. D
As for the label of the CT block, as shown in FIG. 13C, the first line is “L1, L5, L3, L1, ..., L1, L.
5, L3 ", the second line is" L2, L6, L4, L2, ...
.., L2, L6, L4 ", and this DCT block array is repeated in the vertical direction. In addition, considering that the horizontal DCT block array (that is, the column direction) represents one field in 6 tracks," L1 , L5, L3 "," L
It is arranged to repeat 2, L6, L4 ". This DC
The T block array is adopted for recording a data block located on the left side of a data block recorded on the same front track of the head channel on the next track of the same channel after track recording in the same channel recording. .

The shuffling portion 11 divides the six labels L1 to L6 into groups according to the labels attached to the formed DCT blocks shown in FIG. 13 (d).
By this grouping for each label, one group is composed of 38 × 30 DCT blocks for the luminance signal Y and 38 × 15 DT for the color difference signals (RY) and (BY).
It is composed of CT blocks. The shuffling portion 11 is
1 screen is sub-sampled and divided into 2 channels, 1
Each channel is composed of 6 groups, that is, one screen is classified into channels and labels and grouped into 12 groups.

FIG. 14 is a diagram showing a result of arranging these 12 groups of DCT blocks in accordance with the above-described plurality of shuffling principles set in the shuffling section 11. The six tracks shown in FIG. 14 record image data for one field on the screen. Here, the rotary head 16 moves along each track to the sector S _A1.
From the sector S _A2 (arrow S). While scanning in this direction, the rotary head 16 is moved in the direction of arrow M.

As a principle of shuffling in the shuffling section 11, a plurality of DCT blocks obtained by sub-sampling one screen as described above are recorded with different head channels for recording for each track. , One track is divided into two tracks each composed of two sectors and recorded.
In a block, DCT blocks that are aligned in the vertical direction are recorded on the same track, and are recorded separately in the first half sector and the second half sector of this track. When recording on the same channel of the head channel, the next track is recorded. By satisfying the principle of recording the DCT block at the position on the left side of the above, the DCT blocks labeled with the above DCT block in advance and classified, and the DCT blocks with the same label of different head channels are shifted by 3 tracks at a time. Each sector is recorded separately. That is, according to this method, the shuffling unit 11 records the DCT blocks CH1-L1 and CH2-L1 of the label L1 of the channels CH1 and CH2 on the first track and the fourth track, respectively.

In other words, by performing such a shuffling setting, the shuffling section 11 associates the image data with the DCT block and the DCT block with the sector including the recorded track. DCT
Block CH1-L5 and DCT block CH2-L3
When concealing using sub-sampled image data, recording tracks are written adjacent to each other, but there is no relation of image data used for concealing each other.

By satisfying such a shuffling principle, the shuffling section 11 can prevent the sample data near the boundary of the DCT block from including the track and the sector. When the DCT block arrangement according to the above-mentioned rule is performed, if there is an error at the boundary of the DCT block, if there is a scratch in the longitudinal direction, one of the magnetic heads provided on the rotary head 16 or the adjacent magnetic heads. If the pair is head clogged or a burst error occurs, it can be concealed by coping with it.

Digital V for compressing image data
The TR lengthens the DCT block. Considering the method of minimizing the influence of the sync error when performing the equal length, the way of the shuffling process is affected. As a method for reducing the influence of sync error as much as possible, for example, in many sync blocks such as 1 field, 1 block is used without equalization.
There is a method of using a sync block as a unit of equal length. By making the lengths equal, the digital VTR can reduce the influence of the image quality deterioration due to the DCT block in which an error has occurred, and can also increase the image update rate during shuttle reproduction.

The DCT block formation in the digital VTR is performed by the luminance signal Y of the image data, as described above.
This is also performed for the color difference signals (RY) and (BY).
Luminance signal Y is 13.5 MHz, color difference signals (RY), (B-
Y) is 6.75 MHz, and the data lengths of the luminance signal and the color difference signals are different by sampling at a so-called 4: 2: 2 ratio. In order to maintain the correlation between the luminance signal and the color difference signal and improve the efficiency of equalization of DCT blocks over the entire screen, it is desirable that the image data amount included in the equalization unit to be configured be approximately the same. Therefore, the data blocks of the luminance signal Y and the color difference signals (RY) and (BY) are recorded in different sync blocks as different equal length units. On the other hand, the color difference signals (RY) and (BY) are recorded in the sync block of the same equalization unit.

By making this setting, it is possible to improve the efficiency of equalizing the entire screen. Further, by recording in the same sync block, it is possible to prevent a problem in which color reproducibility is deteriorated by mixing color difference signals between fields during conventional shuttle reproduction.

Next, the order of the DCT blocks recorded in each sector shown in FIG. 14 will be described with reference to FIG. As described above, each group of each channel has the 38 × 30 = 1140 DCT blocks shown in FIG. The numbers of the DCT blocks of one group are numbered by 1 from the DCT block number 0 in the upper left corner shown in FIG. 15 to the data block number 29 in the horizontal direction, and then the lines are changed to sequentially number the numbers. Finally, the DCT block in the lower right corner is numbered 1139.

For example, the DCT block Z of the luminance signal Y
_Y is recorded as N _Y- th satisfying the equation (1) of Z _Y = (1087 × N _Y )% 1140 (1) using% as a symbol representing the remainder of division. The value 1087 in Expression (1) is relatively prime to the value 1140 in Expression (1). Therefore, the value Z _Y of the equation (1) also takes a value of 0 to 1139 while the N _Y changes from 0 to 1139.

For the color difference signals (RY) and (BY), D
Since CT number of blocks 38 × 15 = 570 pieces is, N _C satisfying the formula (1) Z _C = for calculating the same remainder as the _{(1087 × N C)% 570} (2) of formula (2) Record second.

Addresses are easily generated by performing the operations of the equations (1) and (2), and errors on the screen are dispersed even if continuous DCT blocks on the magnetic tape disappear. Can be made inconspicuous.

However, if the data calculated by the above equations (1) and (2) are used as they are, adjacent DCT blocks will be arranged side by side on the tape. At this time, the scratch BE due to the burst error formed in the longitudinal direction of the magnetic tape shown in FIG. 16A, for example, on the sector S _A1 side
, Etc., the location of the error along the scratch is “L2,
You will concentrate on L6, L4 ".
In order to avoid this error concentration displayed on the screen, the shuffling unit 11 provides an offset at an initial position where a certain 2 × 3 DCT block on the screen is extracted and recorded, as shown in FIG. 16B. Are dispersed. In this way, the shuffling unit 11 offsets the sync block position for recording. By recording in this way, the error occurrence positions on the tape can be dispersed.

In practice, the shuffling section 11 uses 114 sync blocks per sector in order to equalize the error distribution, and the shuffling section 11 is separated by approximately equal intervals and each of the sectors is approximately one sector. (See Table 1). For example, the data block with label L2 is offset by "58" for both channels CH1 and CH2 and shifted by approximately one sector.

Here, the number of DCT blocks of one sync block is 20 in total, including 10 data blocks for the luminance signal Y and 5 data blocks for the color difference signals (RY) and (BY). It is composed of. Since the DCT blocks of the luminance signal and the color difference signal on the screen are recorded in different sync blocks, the luminance signal number N _Y of the above equation (1) is started from zero, while the number of the color difference signal of the equation (2) is changed. Number N _C
Has an offset to start with N _C = 10.

In this embodiment, P of 625/50
Although it was explained based on the AL method, NTS of 525/60
The C method can be similarly performed. The setting is PA
304 samples in the vertical direction in the L system are NTSC
With the number of 256 in the method, basically the same can be done except that the offset value is changed by changing the value 1087 and the value 1140 of Expression (1) and Expression (2) to 913 and 960, respectively.

By performing the shuffling process for arranging the DCT blocks as described above, in the digital VTR, for example, when there is an error at the boundary of the DCT block, when there is a scratch in the longitudinal direction, the rotary head is provided. Even if there is an error such as one of the magnetic heads or a pair of adjacent magnetic heads clogging a head or a burst error occurs, concealing can be performed to make the trouble appearing on the screen inconspicuous.

A sync block is formed by collecting a plurality of the above DCT blocks, and one sync block can be treated as one equal length unit to disperse image data errors on the screen so as to be inconspicuous. .

Further, the DCT block of the luminance signal and the DCT block of the color difference signal corresponding to the same pixel on the screen are recorded in different sync blocks on the tape, respectively, and the data blocks of the two color difference signals forming the color difference signal are synced with each other. It is possible to prevent the color reproducibility from deteriorating even if the data is recorded in the block and the shuttle reproduction is performed.

In this embodiment, 525/60 NTSC is used.
It can also be applied to the method.

[0093]

According to the image data recording apparatus of the present invention, the data obtained by sub-sampling the input image data forms a plurality of orthogonal transform blocks, and the orthogonal transform blocks are shuffled and then orthogonal transform is performed. An image data recording apparatus that performs processing to compress data, and by recording the orthogonal transform blocks of the overlapping area in the subsampling in association with different head channels of the rotary head, an error is generated. Even if there is, it can be concealed and the defects that appear on the screen can be made inconspicuous. If the plurality of shuffling principles described above are satisfied, for example, if there is an error at the boundary of the DCT block, if there is a scratch in the longitudinal direction, one of the magnetic heads provided in the rotary head, or an adjacent magnetic head. Even if there is an error such as a head clog of the pair, a burst error occurs, or the like, concealing can be performed to make the trouble appearing on the screen inconspicuous.

By treating one sync block in which a plurality of the orthogonal transform blocks are combined as one equal length unit, it is possible to disperse image data errors on the screen and make them inconspicuous.

Further, the DCT block of the luminance signal and the DCT block of the color difference signal corresponding to the same pixel on the screen are recorded in different sync blocks on the tape, respectively, and the DCT blocks of the two color difference signals forming the color difference signal are synced with each other. It is possible to prevent the color reproducibility from deteriorating even if the data is recorded in the block and the shuttle reproduction is performed.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an image data recording apparatus according to the present invention.

FIG. 2 is a block diagram showing a configuration of a shuffling portion in the image data recording device.

FIG. 3 is a schematic diagram showing an arrangement of magnetic heads in the rotary head used in this embodiment.

FIG. 4 is a diagram showing a track configuration for writing 1-field data in 6 tracks and a magnetic head corresponding to each track in the image data recording device.

FIG. 5 is a diagram for explaining subsampling performed corresponding to image data in the image data recording device.

FIG. 6 is a diagram for explaining a relationship among tracks, recording channels and a magnetic head of the image data recording device.

FIG. 7 is a diagram for explaining an advantage of dividing a track in the image data recording device into two sectors.

FIG. 8 is a diagram for explaining that DCT blocks having the same head channel CH1 or CH2 used in the rotary head are vertically arranged in the same track and the possibility of concealment by this arrangement.

FIG. 9 shows DCT blocks (L1, L) arranged in a vertical direction.
FIG. 3 is a diagram for explaining that 2), (L5, L6), and (L3, L4) are divided into two sectors on a track and recorded in pairs.

FIG. 10 is a diagram illustrating a case in which, when recording is performed on the same head channel of an image data recording apparatus, the next track of the same head channel after track recording is positioned to the left of the DCT block recorded on the previous track. It is a figure for explaining recording a DCT block.

11 is a schematic diagram for explaining that concealment becomes possible by using the DCT block arrangement shown in FIG.

FIG. 12 is a so-called digital VTR 4: 2: 2 standard in which one pixel (pixel) has a luminance signal Y and a color difference signal (R−).
It is a figure showing what is expressed by Y) and (B-Y).

FIG. 13 is a diagram for explaining a procedure of subsampling image data in a digital VTR to form a DCT block and classifying the DCT block.

FIG. 14 is a diagram showing a pattern when 12 groups of DCT blocks for one field are arranged on a magnetic tape according to a plurality of proposed shuffling principles.

FIG. 15 is a schematic diagram showing a DCT block array of sync blocks in the image data recording device.

FIG. 16 is a diagram showing an example of a DCT block in the image data recording apparatus, in which a DCT block is provided with an offset at a recording position when the DCT block is extracted.
It is the figure which showed the array which dispersed the T block.

[Explanation of symbols]

10, 18 ... Input terminal 11 ... Shuffling section 12 ... Discrete cosine transform (DCT) section 13 ... Variable length code Conversion unit 14 ... Error correction encoding unit 15 ... Recording unit 16 ... Rotating head 17 ... Magnetic tape 19, 26 ... Output terminal 21 ... Reproduction unit 22 ... Error correction decoding unit 23 ... Variable length decoding unit 24 ...・ ・ ・ ・ ・ Inverse discrete cosine transform (IDCT)
25: Deshuffling unit 111: Data input / output interface 112: Address switching selection unit 113: Write address generation unit 114 .... Read address generation unit 115 ... Memory S _A1 , S _A2 ... Sectors L1 to L6 ... Labeled data blocks

[Procedure amendment]

[Submission date] June 15, 1993

[Procedure Amendment 1]

[Document name to be amended] Statement

[Name of item to be corrected] 0003

[Correction method] Change

[Correction content]

However, in the digital VTR, the presence of an error may be detected but may not be corrected when the ECC process is performed. In order to avoid such an error correction impossible or to suppress image quality deterioration due to concentration of the screen of modified pixels, for example, the digital VTR performs a so-called shuffling process such as switching the order of generation of image codes and the order of recording codes. Is going.

[Procedure Amendment 2]

[Document name to be amended] Statement

[Name of item to be corrected] 0005

[Correction method] Change

[Correction content]

[0005] In addition, the digital VTR, the digital VTR and the input video signal to digitized video code
In addition to reducing the bit rate per channel by dividing it into a plurality of channels, the image code corresponding to pixels is evenly distributed to each channel so that errors do not concentrate in a certain screen.

[Procedure 3]

[Document name to be amended] Statement

[Correction target item name] 0011

[Correction method] Change

[Correction content]

In both image data recording devices, the orthogonal transformation blocks in the overlapping area are separately recorded in sectors in the first half and the second half of the track. In the orthogonal transform block, the orthogonal transform blocks aligned in the vertical direction are recorded on the same track, and the orthogonal transform blocks adjacent to each other in the vertical direction are recorded on the front half and the rear of the track.
It is recorded in half .

[Procedure amendment 4]

[Document name to be amended] Statement

[Correction target item name] 0020

[Correction method] Change

[Correction content]

Further, recording with the same head is performed as the same channel recording, and when recording with the same channel, an orthogonal transformation block located to the left of the currently recorded data block is recorded in the next track, and the mounting angle is set. Small of
Even if two heads become abnormal at the same time, the samples of the orthogonal transform block are separated so that they do not adjoin each other.

[Procedure Amendment 5]

[Document name to be amended] Statement

[Correction target item name] 0040

[Correction method] Change

[Correction content]

This is because the DCT block formed by sub-sampling the overlapping region when sub-sampling is recorded on each track corresponding to different head channels of the rotary head 16, concealment by sub-sampling is performed. It is possible.

Claims (7)

[Claims] 1. Image data recording in which data obtained by subsampling input image data constitutes a plurality of orthogonal transform blocks, and the orthogonal transform blocks are shuffled and then subjected to orthogonal transform processing to compress the data. An apparatus for recording image data, characterized in that the orthogonal transform blocks in an overlapping area are recorded in correspondence with different head channels of a rotary head when the subsampling is performed. 2. The image data recording apparatus according to claim 1, wherein the orthogonal transform blocks in the overlapping area are separately recorded in sectors in the first half and the second half of the track. 3. The image data recording apparatus according to claim 1, wherein, in the orthogonal transform block, the orthogonal transform blocks aligned in the vertical direction are recorded on the same track. 4. The image data recording apparatus according to claim 3, wherein the orthogonally-transformed blocks that are vertically adjacent to each other are recorded separately in each of the first half and the second half of the track. 5. An orthogonal transformation block located to the left of the orthogonal transformation block used for recording the one track in a track next to one track for data recorded in the same channel of the head channel. The image data recording apparatus according to claim 1, 2, 3, or 4, wherein the image data is recorded. 6. The image data recording apparatus according to claim 1, wherein one sync block in which a plurality of the orthogonal transform blocks are collected is treated as one equal length unit. 7. An orthogonal transformation of two color difference signals forming a color difference signal by recording an orthogonal transformation block of a luminance signal and an orthogonal transformation block of a color difference signal corresponding to the same pixel on a screen in different sync blocks on a recording medium, respectively. 2. The blocks are recorded in the same sync block.
2. The image data recording device according to 2, 3, 4, 5, or 6.