JP3741964B2 - Video recording method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a video recording method for recording or transmitting a video signal with high efficiency encoding.
[0002]
[Prior art]
In recent years, high-efficiency coding technology has been actively developed in order to realize recording / reproduction at a limited transmission rate in digital signal processing of video from digital video equipment and the like. High-efficiency coding is a coding method that compresses the amount of data using the redundancy of a video signal. For example, data compression using spatial correlation within a frame or temporal correlation between frames is used. is there.
[0003]
Intraframe compression is a data compression method that utilizes the spatial correlation within a frame, that is, the feature that an arbitrary pixel and other neighboring pixels often have close values. In this compression method, a frame is divided into, for example, blocks of 8 pixels × 8 pixels (hereinafter referred to as orthogonal transform blocks), and compression is performed by performing orthogonal transform in units of orthogonal transform blocks. Compression efficiency can be improved by performing quantized and statistically defined variable length coding on the orthogonal transform coefficient calculated by the orthogonal transform.
[0004]
The inter-frame compression is a data compression method using a temporal correlation between frames, that is, a feature that a certain pixel block in a frame often exists at almost the same position in a neighboring frame. This compression method generally encodes difference data between pixel data of the frame and pixel data of neighboring frames. In particular, when a moving image is recorded, the compression efficiency can be increased by accurately predicting a position where a certain pixel block has moved in time from a neighboring frame and taking the difference from the predicted pixel block.
[0005]
Such a method is generally called motion compensation prediction. In the motion compensated prediction method, code data obtained by high-efficiency coding differential data and a motion vector indicating motion information are transmitted. In the video signal input here, the first frame encoded within the frame without using motion compensation prediction is an I frame, and one subsequent frame predicted using a past frame as a reference image is a P frame. Of the predictions using the past frame as the reference image, the prediction using the future frame as the reference image, or the prediction using both, the frame to which the prediction with the highest compression efficiency is applied is selected as the B frame.
[0006]
When recording / reproducing encoded data compressed using high-efficiency encoding as described above, since variable-length encoding is used, the encoded data cannot be decoded when an error occurs and cannot be reproduced. The demerit occurs. In addition, since the recording position and the position on the screen are indefinite, the image quality of high-speed playback in a VTR or the like is significantly deteriorated.
[0007]
As an example for solving these problems, there is a technique disclosed in Japanese Patent Laid-Open No. 6-268964. According to this example, an input video signal is subjected to high-efficiency encoding in units of K frames, which is 1 GOP, and recorded on L tracks on a recording medium. At this time, the first half of the L tracks, that is, N recording blocks having a fixed length with respect to the first frame recording area, and a block obtained by dividing the frame corresponding to the I frame into N blocks are highly efficient. Encode to generate N compressed blocks. Thereafter, N compressed blocks are recorded in association with N recording blocks. In addition, the compressed block data overflowing from the recording block is packed in a free area of another recording block and recorded, so that the recording position can be specified even for variable-length code data.
[0008]
In the remaining track, that is, the subsequent frame recording area, the encoded data of the first frame that could not be recorded in the first frame recording area, and the difference data of the second and subsequent frames corresponding to the P frame or the B frame ( Those data are recorded in the order of “including motion vectors if there are motion vectors”.
[0009]
Thus, when a transmission path error occurs, variable length coding is reset in N fixed-length recording blocks, that is, the first frame recording area, so that transmission of transmission path errors can be suppressed. In addition, since the data of N compressed blocks are recorded in N recording blocks having a fixed length, it is possible to sequentially reproduce video images independently for each recording block during high speed reproduction.
[0010]
[Problems to be solved by the invention]
However, in the conventional recording method described above, when a transmission path error occurs, variable-length coding is reset in N fixed-length recording blocks. When there is overflow data (for example, high-frequency data) from the compressed block, the code continuity after that is lost and cannot be restored. In this case, the reproduction image quality is equivalent to the image quality reproduced using only the low-frequency range of the variable length code (image quality with which block distortion or the like can be recognized). In addition, this influence has a problem that it extends over the entire screen and causes a reduction in image quality.
[0011]
In the recording area other than N recording blocks, the continuity of the code data after the location where the transmission path error occurs is interrupted and propagates until the first frame in the next K frame unit is decoded. For this reason, when K is a large value, that is, when the compressed frame unit is large, many frames cannot be restored.
[0012]
Also, since the data of the i-th frame (1 <i ≦ K) following the first frame is recorded in an area other than the specific N recording blocks, depending on the input image, the data of the first frame The amount of code data may be smaller than the recording capacity of N recording blocks. In this case, there is a problem that the recording block may be an empty area, and the compression efficiency is substantially reduced.
[0013]
The present invention has been made in view of the above-described conventional problems. Even when a transmission path error occurs, error propagation can be suppressed in the recording block for the data of the first frame. If there is an error in the data recorded in the free area, propagation can be suppressed to only M compressed blocks by the added header information and position information, and deterioration of the image quality of the entire screen is prevented. An object of the present invention is to realize a video recording method.
[0014]
[Means for Solving the Problems]
The invention of claim 1 of the present application collects input video signals for K frames, performs high-efficiency encoding in units of K frames, and outputs the K frames of high-efficiency encoded data on a recording medium. The first frame in the K frame is highly efficient encoded using only the data in the frame, and the i th frame (1 <i ≦ K) is in the frame. When performing high-efficiency encoding using only data or difference data between the i-th frame and a frame other than the i-th frame, for the first frame, the data in the frame is divided into N compressed blocks. High-efficiency encoding is performed, and N recording blocks are provided in the first frame recording area in the L tracks so as to correspond to the data recording areas of the N compressed blocks. Block data is recorded in each recording block of the first frame recording area, and data overflowing from each recording block is added with header information and position information for each M (0 <M ≦ N) compressed block units. The remaining data of the first frame, and for the i-th frame, only the data within the frame or the difference data between the i-th frame and the frame other than the i-th frame is divided into Yi compressed blocks to obtain a high Encoding is performed and Xi (0 <Xi ≦ Yi) header information and position information are added for each compressed block unit to obtain code data of the i-th frame, and the remaining data of the first frame is followed by the first data. A data stream is created by concatenating i-frame code data, and the data stream data is stored in an empty recording block in the first frame recording area. The remaining data of the data stream is recorded in a frame recording area other than the first frame recording area after at least a part of the data is recorded and all the empty recording blocks in the first frame recording area are filled. It is characterized by.
[0015]
DETAILED DESCRIPTION OF THE INVENTION
A video recording method according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of a video recording apparatus for realizing the video recording method of the present invention. This video recording apparatus includes a video input terminal 10, a differentiator 11, a high-efficiency encoding means 12, a memory 13, a recording block stuffing means 14, a data linking means 15, a recording means 16, and a recording medium 17.
[0016]
The operation of the video recording apparatus having such a configuration will be described. In the following, it is transmitted in the 525/60 system, and a luminance signal of one frame is composed of 720 pixels in the horizontal direction and 480 lines in the vertical direction. Such a video signal is independently encoded with high efficiency for each K frame which is a GOP unit. A case of recording will be described. Each frame is composed of a plurality of orthogonal transform blocks, with 64 pixels of 8 horizontal pixels and 8 vertical lines as orthogonal transform blocks. That is, the number of orthogonal transform blocks per frame is 5400.
[0017]
In addition, in the two color difference signals, that is, the RY signal and the BY signal, one orthogonal transform block is configured by pixels included in the same range as the orthogonal transform blocks of four adjacent luminance signals. A total of six orthogonal transform blocks including four orthogonal transform blocks of luminance signals in the same range and one orthogonal transform block of each of the RY signal and BY signal are called one macroblock. Therefore, one frame is composed of 1350 macroblocks. In the present embodiment, one macro block is associated with one compressed block. Therefore, it is assumed that N = 1350 compressed blocks are encoded with high efficiency for the I frame.
[0018]
The high-efficiency encoding means 12 in FIG. 1 first performs high-efficiency encoding using only the pixel data in the first frame as an I frame. Next, the high-efficiency encoding unit 12 performs motion compensation prediction as a P frame or a B frame in the second and subsequent frames. Of the pixel data in the frame or the difference data between the frame and another frame, the compression efficiency The better one is encoded with high efficiency. The memory 13 holds image data that has been highly efficiently decoded with respect to other frames in order to create difference data. The differentiator 11 generates difference data using the image data given from the video input terminal 10 and the image data held in the memory 13. Here, various methods can be considered for motion compensation prediction, and any method can be used as long as it has a function of searching for a pixel block of another reference image having a value close to the compressed block. it can. Also, how to select P frames and B frames after the second frame can be arbitrarily set.
[0019]
Now, the first frame is divided into 1350 compressed blocks and is encoded with high efficiency. The recording block stuffing means 14 shown in FIG. 1 firstly stores 1350 code data of 1350 compressed blocks at specific positions in L tracks on the recording medium 17, that is, 1350 pieces provided in the first frame recording area. Packs corresponding to the recording block. Here, since the code data of each compressed block is variable-length encoded, the data amount is larger or smaller than the recording capacity of the corresponding recording block. When the amount of data is larger than the recording capacity of the recording block, it overflows from the corresponding recording block. In this case, overflowing data (hereinafter referred to as residual data) is output to the data linking unit 15. When the data amount is smaller than the recording capacity of the recording block, all data is packed into the recording block, and an empty area is created in the recording block.
[0020]
Next, as described above, the high-efficiency encoding unit 12 performs high-efficiency encoding on the pixel data in the second frame and the subsequent frames (i-frames) or the difference data from other frames as described above. Further, as described above, each frame is divided into 1350 macroblocks to be highly efficient encoded, and the number of compressed blocks Yi is assumed to be 1350. The 1350 compressed blocks that have been input are encoded with high efficiency, resulting in variable-length code data, which are output to the data concatenation unit 15.
[0021]
The data concatenation unit 15 inputs the residual data of the first frame and the code data encoded with high efficiency after the second frame. First, for example, 3 macroblock units ( Header information and position information are added at M = 3). Secondly, the data concatenation unit 15 adds the data in which the header information and the position information are added, for example, in units of 5 macroblocks (Xi = 5) to the code data in the second frame and thereafter, and the remaining data in the first frame described above. Continuing on.
[0022]
The header information is a code that is incorporated in the data string and indicates a data breakpoint. The position information is information indicating what number the breakpoint is. With this information, it is possible to identify which position on the screen the data recorded in this section corresponds to. The data stream thus concatenated is supplied to the recording block stuffing means 14.
[0023]
The recording block stuffing unit 14 stuffs the concatenated data stream into the recording blocks in which the empty area remains after the code data of the first frame is packed. When the data stream is packed in all the empty areas of the recording block, the data stream input is stopped. The data packed in the recording block in this way is given to the recording means 16 and converted into recording data. Such recording data is recorded in each recording block in the first frame recording area in the L track of the recording medium 17.
[0024]
Next, after the data stream is stuffed into all free areas of the recording block in the recording block stuffing means 14, the data stream still remaining in the data linking means 15 is directly output to the recording means 16, Are sequentially recorded in the subsequent frame recording area in the L track.
[0025]
Next, a recording pattern on the recording medium 17 recorded by the above operation is shown in FIGS. FIG. 2 shows a case where K = 4, and shows a case where the amount of code data of the first frame is larger than the recording capacity of the first frame recording area L1. Here, the area of each recording block corresponds to the position of the compressed block on the screen.
[0026]
In FIG. 2, firstly, the code data I of the first frame is packed in each recording block of the first frame recording area L1 in the L tracks, and the remaining data of the first frame once overflowing from each recording block is the data connecting means 15. The header information and the position information are added by, and packed into each empty recording block as a data stream I ′. Then, the remaining data I ″ overflowed in each recording block, the code data D2 of the second frame, the code data D3 of the third frame, and the code data D4 of the fourth frame are recorded in the subsequent frame recording area L2.
[0027]
FIG. 3 is a diagram showing a case where the code data amount of the first frame is smaller than the recording capacity of the first frame recording area L1 with K = 4. First, the code data I of the first frame is packed in each recording block of the first frame recording area L1 in the L tracks. Then, the header information and position information are added to the remaining data of the first frame once overflowed from each recording block by the data connecting means 15, and are packed again into the first frame recording area L1 as the data stream I ′. Further, since there is a vacant area in a specific recording block in the first frame recording area L1, a part of the code data D2 of the second frame is continuously recorded. The remaining second frame code data D2, the third frame code data D3, and the fourth frame code data D4 are recorded in the subsequent frame recording area L2.
[0028]
Next, the operation of the recording block stuffing means 14 will be described in more detail. FIG. 4 is an explanatory diagram showing an example of a method for stuffing a recording block. The recording block is divided into four luminance signals (Y0 to Y3) and two color difference signals (RY, BY) constituting a macroblock which is a compression block. The auxiliary data required for decoding the orthogonal transform block and the code data indicating the low frequency are sequentially packed into the code data indicating the high frequency. In general, distortion of a component that visually represents a low frequency is easily detected, but distortion of a high frequency component is difficult to detect. Therefore, by packing in the order as described above, it is possible to make the image quality deterioration as inconspicuous as possible even in the case where code data indicating a high frequency band with low importance cannot be restored in each orthogonal transform block.
[0029]
In FIG. 4A, when the code data of the compressed block of the first frame is stuffed by such a stuffing method, the amount of data is larger than the recording capacity of each recording block. It is shown schematically. Here, the shaded portions in the orthogonal transform blocks Y0, Y2, and Y3 are residual data. These residual data are output from the recording block stuffing means 14 to the data linking means 15.
[0030]
Next, FIG. 4B schematically shows that a free area has been created because the code data of the compressed block of the first frame is packed and the code data of each compressed block is smaller than the recording capacity of each recording block. It is shown in. Here, the hatched portions in each of the orthogonal transform blocks Y0 and Y3 indicate that they are free areas. Therefore, the data stream from the data connection means 15 is sequentially packed in this empty area.
[0031]
If there is overflow data in a certain orthogonal transform block in the compressed block and there is a free area in another orthogonal transform block in the same compressed block, the overflow data is first packed in the free area. Anyway. In this case, after the stuffing in the same compressed block is completed, the data stream from the data connecting means 15 is sequentially stuffed into the vacant area of the recording block having the vacant area.
[0032]
In addition, when the remaining data of the first frame and the data amount of the first frame are smaller than the recording capacity of all the recording blocks, the code stream of the second and subsequent frames is sequentially packed in the data stream packed in the empty area. Can do.
[0033]
Next, details of the data stream generated by the data connection means 15 will be described again with reference to FIG. FIG. 5 is an explanatory diagram showing the data structure of the data stream. The data stream shown in the first column of the figure shows data that overflows when the code data of the first frame is packed in the recording block by the above-described operation in the recording block packing unit 14. Header information and position information are added to the remaining data.
[0034]
As described above, when 1350 macroblocks are encoded in the entire frame, header information and position information are added using the remaining data of M = 3 macroblocks as one delimiter unit. Position information is added following the header information, and the position information indicates the number of this delimiter unit. In this case, since the delimiter unit is 450 pieces, the position information is information that can represent 450 patterns.
[0035]
Further, although there is a case where no residual data is generated depending on the macro block, nothing is recorded for the macro block at that time. This shows a state in which there is no residual data of the second macroblock to be included in the delimiter unit at the second position of the data stream in the first column in FIG. Further, if no residual data is generated in all of the three macroblocks constituting the delimiter unit, there may be no residual data with only header information and position information. In such a case, it can be configured not to add header information and position information (skipping).
[0036]
Next, the data stream shown in the second column of FIG. 5 is the code data of the second frame to be concatenated following the residual data of the first frame of the first column. Here, when 1350 macroblocks are encoded in one entire frame, header information and position information are added using code data of macroblocks with Xi = 5 as one delimiter unit. Position information is added following the header information, and the position information indicates the number of this delimiter unit. In this case, since the delimiter unit is 270, the position information is also information representing 270 ways. After this position information, code data of five macroblocks are output. The third and subsequent frames have the same configuration as the second frame. However, the number of macroblocks Xi constituting the delimiter unit is not necessarily the same as that of the second frame or other frames, and can be arbitrarily determined in advance. For example, Xi = 15 may be determined.
[0037]
As a function of the data linking means 15, for example, using a memory, the remaining data of the first frame, the code data of the second frame, the code data of the third frame, and the code data of the i-th frame are stored in this order. Shall. However, it is not necessary to limit the storage order to the above order.
[0038]
In this embodiment, K = 4. However, the present invention is not limited to this, and the frame unit to be encoded can be arbitrarily set.
[0039]
In this embodiment, N = 1350. However, the present invention is not limited to this, and can be arbitrarily set depending on the information amount of the video signal, the encoding method, and the like.
[0040]
In the present embodiment, an explanation has been given of the operation in which one macro block is associated with one recording block. However, it is also possible to increase the compression ratio and associate a plurality of macro blocks with one recording block. It is.
[0041]
In this embodiment, M = 3. However, the present invention is not limited to this, and can be arbitrarily set as long as the condition (0 <M ≦ N) is satisfied.
[0042]
Although Yi = 1350 is set in the present embodiment, the present invention is not limited to this, and can be arbitrarily set depending on the information amount of video signals, the encoding method, and the like.
[0043]
In this embodiment, Xi = 5. However, the present invention is not limited to this, and can be arbitrarily set as long as the condition (0 <Xi ≦ Yi) is satisfied.
[0044]
In the present embodiment, the position information to be added indicates the order of the delimiter unit for the data stream generated by the data concatenation unit 15, but any information that can correspond to the position of the screen and the delimiter unit is used. Any format is acceptable. Further, information that can identify the frame number can be included. Such frame identification information can also be included in the header information. In the data stream generated by the data linking unit 15, when all the remaining data of the first frame is recorded in the recording block, the dummy data is placed after the remaining data of the first frame so as to fill all the recording blocks. It can also be inserted. Further, recording similar to this can be performed by the recording means 16. However, in this case, the compression efficiency is lower than the method described in the embodiment.
[0045]
In the present embodiment, a video recording method for a specific video signal has been described. However, the present invention can be applied to any other video signal. In the description, a video signal composed of a luminance signal and a color difference signal is described, but the present invention can also be applied to a composite signal, an RGB signal, and other image signals and audio signals.
[0046]
The recording medium can be applied to any medium such as a magnetic tape, a disk recording medium, and a semiconductor recording medium.
[0047]
In the present embodiment, the configuration as shown in FIG. 1 is shown as an example for realizing the present invention. However, various configurations other than that are possible and can be realized by software on a computer. is there.
[0048]
【The invention's effect】
As described above, according to the present invention, the code data of the N compressed blocks in the first frame is recorded in specific N recording blocks on the recording medium corresponding to the screen position, so that the recording data is recorded in units of recording blocks. Can be reproduced, and the image quality can be improved during high-speed reproduction.
[0049]
Also, since the code data of the N compressed blocks of the first frame is recorded in specific N recording blocks on the recording medium corresponding to the screen position, even if an error occurs in the transmission path, the error is In addition to the error being reset in the generated recording block, the residual data (data indicating the high frequency) of the first frame can also be suppressed in terms of the error propagation in the first unit.
[0050]
This is especially true when encoding is performed to increase the compression efficiency by causing the residual data of the compressed block overflowing from the recording block to flow around the entire frame, thereby increasing the compression efficiency. It is possible to prevent the continuity of the area data from being interrupted and the image quality from being deteriorated on the entire screen.
[0051]
Furthermore, since the code data can be recorded without any gap in the recording area, N pieces of code data of the first frame are encoded among the data encoded to record on the L tracks allocated to the K frame. It is possible to set smaller than the recording capacity of the recording block. For this reason, even when a scene change occurs after the second frame and the difference data becomes large, it is possible to allocate a sufficient code amount to the frame, and it is possible to improve the reproduction image quality.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a video recording apparatus for realizing a video recording method according to an embodiment of the present invention.
FIG. 2 is a conceptual diagram showing a recording pattern (No. 1) on a recording medium in the present embodiment.
FIG. 3 is a conceptual diagram showing a recording pattern (No. 2) on a recording medium in the present embodiment.
FIG. 4 is a conceptual diagram showing an example of a recording block in the present embodiment.
FIG. 5 is a conceptual diagram showing a configuration of a data stream in the present embodiment.
[Explanation of symbols]
10 Video input terminal
11 Differentiator
12 High-efficiency encoding means
13 memory
14 Recording block stuffing means
15 Data connection means
16 Recording means
17 Recording media

Claims (1)

Video that collects input video signals for K frames, performs high-efficiency encoding in units of K frames, and divides the K-frame high-efficiency encoded data into L tracks on a recording medium for recording A recording method,
The first frame in the K frame is highly efficient encoded using only the data in the frame, and the i-th frame (1 <i ≦ K) is the data in the frame only, or the frame other than the i-th frame and the i-th frame. When performing high-efficiency encoding using difference data with
For the first frame, the data in the frame is divided into N compressed blocks and subjected to high efficiency encoding. N recording blocks are provided in the first frame recording area in the L tracks, and the N frames Corresponding as a data recording area of the compressed block, the data of each compressed block of the first frame is recorded in each recording block of the first frame recording area,
The data overflowing from each recording block is the remaining data of the first frame by adding header information and position information for each M (0 <M ≦ N) compressed block units,
For the i-th frame, the data within the frame only, or the difference data between the i-th frame and the frame other than the i-th frame is divided into Yi compressed blocks and subjected to high-efficiency encoding, and Xi (0 < Header information and position information are added for each compressed block unit of Xi ≦ Yi) to obtain code data of the i-th frame,
Concatenating the code data of the i-th frame after the residual data of the first frame to create a data stream,
After at least part of the data stream data is recorded in the empty recording block of the first frame recording area and all the empty recording blocks of the first frame recording area are filled, a frame recording area other than the first frame recording area And recording the remaining data of the data stream.

Images