KR0141223B1 - Image recording method for dvcr - Google Patents

Abstract

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image recording method of DVCR, which uses a high compression property which is an advantage of interframe coding, and simultaneously solves the problem of data amount fixing and enables trick play and the like. The low frequency data of the intra frame corresponding to the portion is regularly and repeatedly recorded in one group of pictures (GOP), and a residual data recording area is provided at the end of each GOP. In addition, by grouping small areas on the image regularly or arbitrarily and recording low frequency data of all small areas belonging to the same group together in the low frequency data recording area of each small area, the quality of the image is improved during trick play.

Description

Image recording method of DVCR

1 is a block diagram showing the configuration of a general video encoder.

2 is an example of a recording format of a tape according to the present invention.

3 is another example of a recording format of a tape according to the present invention.

4A to 4C show an image composed of a plurality of small regions.

5 is a diagram illustrating one type of sync block for trick play according to the present invention.

* Explanation of symbols for the main parts of the drawings

100: recording medium 110: high frequency data recording area

120: low frequency data recording area 130: residual data recording area

BACKGROUND OF THE INVENTION 1. Field of the Invention [0001] The present invention relates to digital BC (hereinafter referred to as DVCR). In particular, the present invention relates to a digital CR (highly compressible property), which is an advantage of an interframe coding scheme, while maintaining a constant data rate and trick play ( The present invention relates to a DVCR image recording method capable of trick play and capable of improving the quality of a reproduced image during trick play.

The rapid development of digital video technology in the 1980s has made it possible to use digital video compression technology for applications in a variety of electronic communications fields such as video conferencing, digital broadcast codecs, and video telephony. In particular, the signal band compression method of DVCR, which has been developed over the past several years, has been mainly dominated by intraframe coding without considering redundancy between images. However, since data compression using the intraframe coding method removes redundancy in a frame, data compression efficiency is not high.

Therefore, in recent years, an attempt has been made to apply an interframe coding method, which is a motion video compression method that removes redundancy in the temporal direction, to DVCR data compression. Since the interframe coding method can remove the redundancy of image calculation, the signal compression rate is higher than that of the intraframe coding method.

For example, sampling an NTSC signal at a sample rate of 4: 2: 2 produces about 158 [Mbps] data. About 25 [Mbps] is needed to record this data to tape. It must be compressed within the data. According to the experimental results, when comparing the same amount of compressed result data, the compression result of the interframe coding method is superior to that of the intraframe coding method. Therefore, it is concluded that interframe coding should be used for data compression of DVCR.

However, despite the advantages of the interframe coding scheme, the interframe coding scheme has not been used for DVCR. The reason is due to the characteristics of the recording medium called tape. That is, when recording data on a recording medium such as a tape, a fixed data rate (constant data rate or fixed data rate) is necessarily required, because such data is difficult to fix.

An example of a means for interframe video coding proposed to solve the problem of data fixation is an encoder of Motion Picture Experts Group (MPEG).

1 shows an example of the configuration of a general MPEG video encoder. The operation of the MPEG video encoder is briefly described as follows.

In the encoder shown in FIG. 1, a frame re-order 10 receives a source input image and performs color coordinate transformation, subsampling, and block division.

The motion estimation unit 12 estimates similar parts between the front image and the rear image, detects the result of the position shift as vector data, and outputs image signal data, vector data, and mode data (I, B, P). Here, the mode data (I, B, P) indicate the type of picture (Picture or Frame). The I picture is an intra coded picture and removes only spatial redundancy. The P picture, that is, the predictive picture, is predicted from one picture and only encodes an error. Therefore, the P picture can be said to remove similar portions between the I picture and the P picture. On the other hand, the P picture has motion vector data generated during the prediction as additional data. There is a B picture between the I picture and the P picture, and the B picture is obtained by encoding a prediction error by performing bidirectional prediction from the I picture and the P picture. When comparing the data amounts, there are many pictures in the order of I picture, P picture, and B picture. Since the amount of data per image is nonuniform, a buffer 32 must be provided to adjust the amount of data in order to maintain a constant data transfer rate in the transmission path. As described above, the motion estimation unit 12 outputs mode data (I, B, P), image signal data and motion vector data according to the mode data.

The first adder 14 adds the video signal data output from the motion estimation unit 12 and the video signal data output from the frame store / predictor 28.

A discrete cosine transform (DCT) unit 16 performs energy compression of image signal data input through the first adder 14 to perform discrete cosine transform. ) To be concentrated in one place.

The quantizer 18 quantizes the video signal energy-compressed by the DCT unit 16 to a predetermined quantization level (Q-level or step-size), and runs length coding of the quantized video signal. Is done.

A variable length coding portion 20 is quantized by the quantizer 18 and compresses a run-length coded video signal. That is, among the signals represented by 8 bits (0 to 256 levels), the frequent data is represented by fewer bits, and the less frequent data is represented by more bits, thereby reducing the total number of bits representing the video signal.

On the other hand, the inverse quantizer 22, if the quantized video signal input from the quantizer 18 is restored to the signal before quantization.

The inverse DCT unit 24 restores the signal before the discrete cosine transform when the output of the inverse quantizer 22 is input.

The second adder 26 adds an output of the inverse DCT unit 24 and an output of a frame store and predictor 28. The frame store / predictor 28 outputs the motion vector of the previous image stored after being output from the second adder 26 and the motion vector data and mode data I, B, which are output from the motion estimation unit 12. , P) to compensate.

The multiplexer 30 selectively outputs image signal data compressed and output from the variable length encoder 20 and motion vector data and mode data I, B, and P output from the motion estimation unit 12.

Since the length of the data output from the multiplexer 30 is not constant, the buffer 32 temporarily stores the data and outputs the data at a constant speed. At this time, the buffer 32 determines the quantization lever based on the fullness of the data and controls the quantizer 18 through the regulator 34 based on this. If the data fullness of the buffer 32 is high, the quantization level is increased to reduce the data amount, and if the data fullness is low, the quantization level is decreased to increase the data amount.

The video signal data whose output amount is constantly adjusted is finally recorded on a recording medium such as a tape. By the way, recording on a tape divides and records video signal data of one frame into four video tracks due to the characteristics of DVCR.

However, the conventional recording method as described above has a problem in that it is impossible to reproduce an image during a trick play such as a high-speed search because the video signal is divided and recorded in several tracks.

In view of such a problem, in order to enable high-speed search even in the digital VCR described above, a method of recording video signal data of one frame separately in an expected driving track of the head has been proposed in trick play. However, this method requires accurate prediction of the running track and accurately tracks the running track at the time of trick play, resulting in technical difficulties and problems in practical use. In addition, there is a problem that the data amount between images is irregular because the data is not completely fixed.

On the other hand, the compression-coded data is different from the analog signal, so that the data amount of a constant image area does not match the constant size of the original image. For this reason, the image reconstruction capability may not be as high as that of the analog signal during the trick play. Restoring the digitally compressed data during trick play may result in an unrecovered area of the image, so it is necessary to add an appropriate signal to such area to prevent deterioration of image quality.

Accordingly, in view of the above problems, the present invention allows the amount of data to be recorded on the recording medium to be fixed while using high compression, which is an advantage of the interframe coding method, and reduces the data loss during trick play, thereby improving image quality of a reproduced image. The purpose is to provide a picture recording method of a DVCR.

In addition, another object of the present invention is to provide a DVCR image recording method for adding and recording certain image data on a tape so that the digitally compressed data effectively fills an image so that there is no empty space during trick play.

In order to achieve the above object, the image recording method of the DCVR according to the present invention is encoded using motion compensation prediction from I pictures, past and / or future reference pictures, encoded using information corresponding to the corresponding picture itself. An image recording method of recording on a digital VCR tape as a unit a picture group (GOP) configured by arbitrarily combining a P picture encoded using motion compensation prediction from a B picture and a past reference picture. Low-frequency data of each sync block of I picture is recorded at regular intervals on each track for I pictures, P pictures, and B pictures of the tape, and high-frequency data of each sync block of I picture is recorded on the track for I pictures. The data of the P picture are sequentially recorded in the blank area which is not recorded, and the low frequency data of the I picture is recorded on the track for the P picture, and then in the remaining area. The data of the B picture is sequentially recorded, and the low frequency data of the I picture is recorded on the track for the B picture and sequentially recorded in the remaining area.

In addition, the image recording method of the digital VCR to achieve another object of the present invention, by dividing the compressed digital video signal of each small area of the image divided into a predetermined number of small areas into low frequency data and high frequency data to divide the low frequency data In a digital VCR image recording method of recording a predetermined interval on a track of a tape and sequentially recording the high frequency data in the remaining areas of the tape, the predetermined number of small areas is classified into a plurality of groups according to a predetermined method. In the track area for recording low frequency data of each small region belonging to the same group, the low frequency data of all small regions belonging to the same group are recorded together.

Hereinafter, the present invention will be described first with reference to the attached FIGS. 1 to 3.

2 is an example of a tape format in which data is recorded by the DVCR image recording method according to the present invention. Referring to FIG. 2, the recording medium 100 records one frame of digital image data every four tracks.

A picture obtained by the moving picture compression method includes an I code coded using information corresponding only to the picture itself as an intraframe, i.e., B coded using motion-compensated prediction from past and / or future reference pictures. Although it may consist of the P picture encoded using the motion and the prediction predicted by the motion compensation from the I picture which is the past reference picture, and the said I picture and the P picture only, the present invention takes the former as an example.

In the case of consisting of I, B, and P images of the image to be recorded, the order of the images recorded on the track of the recording medium 100 may be various, but in the present invention, [I, B, B, P, B, B, P, ...., I] [I, B, B, P, B, B, P ...., I] The case of recording in image order will be described as an example. Of course, the digital data of each of the I, B, and P images is recorded on four tracks of the recording medium 100.

First, the image recording method according to the present invention will be described with reference to FIG. In FIG. 2, low frequency data representing the outline of an image and the like in an I picture is recorded in the low frequency data recording area 120 periodically formed in four tracks constituting the I picture. The low frequency data of the I picture is also regularly recorded in the low frequency data recording area 120 periodically formed in the tracks of the B picture and the P picture. The high frequency data of the I picture is recorded in the high frequency data recording area 110 periodically formed in four tracks constituting the I picture. The data of the B picture and the P picture are recorded in portions other than the area in which the low frequency data of the I picture is recorded.

Further, according to the DVCR image recording method according to the present invention, the data amount is fixed by grouping N images into one group. Here, N can be any integer of 2, 3, 4, .... This group is called a group of picture (hereinafter referred to as a GOP). At the end of each GOP, the remaining data recording area 130 is provided with a track length as long as the size of the buffer or the maximum amount of data in the buffer, and the data remaining in the buffer each time the GOP is encoded. Is recorded in the remaining data recording area 130 of the recording medium 100 as it is. In this way, the data amount can be fixed. In actual trick play, the data lost is reduced because the space of sufficient recording medium is used as the remaining data recording area.

As such, the low frequency data of the I picture is regularly recorded again on another track. Therefore, data loss is reduced even in any double speed mode. Of course, only low-frequency data is used, but the image quality deteriorates. However, even if the image quality is slightly deteriorated in any of the double speed modes, since the human visual characteristics are not well recognized, there is no problem in actual use.

If about 15 Mbps of data is consumed for encoding using the moving picture technique, the data of about 5 Mbps can be fully utilized for trick play. This is because if the total amount of data is within 20Mbps, this is a data rate that is very easy to achieve in a recording medium such as a tape.

Thus, in FIG. 2, low frequency data of an intra frame is recorded in a GOP in a regular repeating arrangement. Therefore, even if the recording medium passes at an arbitrary speed, data loss is reduced and images can be reproduced.

Hereinafter, another embodiment of the present invention will be described.

3 is a diagram showing a tape format according to another embodiment of the present invention. Since the tape format shown in FIG. 3 is the same as the tape format shown in FIG. 2 except for the following, the description of the same part is omitted.

The embodiment shown in FIG. 3 differs from the tape format shown in FIG. 2 in that a residual data recording area of the buffer is not separately placed at the end of the GOP. In general, independent encoding is performed in GOP units for encoding and decoding. Therefore, by appropriately adjusting the bit rate of the buffer, the residual data recording area 130 corresponding to the buffer of FIG. 2 can be avoided.

Next, an image recording method for achieving another object of the present invention will be described with reference to FIGS. 4 (a) to (c) and FIG.

As mentioned above, the present invention is to enhance a trick play function when a moving picture is compressed and recorded on a tape, and is a method for perfectly forming an image in a fast forward / backward search mode.

In general, even when digital compressed data is made so that one sync block represents a certain region of an image due to its characteristics, it is difficult to correctly configure an image during trick play. In view of such a problem, the present invention proposes a new recording method for increasing the amount of data read by the magnetic head during trick play, so that the reconstruction of the image is advantageous. That is, this recording method is based on the method of separately recording the low frequency data and the high data when recording one sync block on the tape as in the conventional recording method, but the low frequency data recording area 120 has a specific image on the image. Instead of recording only the low-frequency data of the part, a method of loading together the low-frequency data of various parts on the image is taken. In this way, even when performing trick play, the amount of information read from the magnetic head increases, and the information is evenly distributed over the entire image, thereby improving the quality of the trick-played image.

To this end, first, the image is divided into a plurality of small areas, and the small areas are grouped into several groups by a predetermined method. That is, one image is composed of a plurality of groups, and each group is composed of a plurality of small regions.

There are two ways to group small areas. One is to group according to a certain rule, the other is to group arbitrarily.

Here, the former method is demonstrated first.

In the case of 30 moving images per second, one frame can be divided into a plurality of small regions (Subimage or Macroblock) as shown in FIG. In the case of an image having a horizontal h pixel, a vertical v pixel, and an image having a horizontal H pixel and a vertical V pixel, there are L small areas per image as shown in the following equation.

[Α] here means the largest integer not exceeding α.

In the image divided into L small areas, the small areas are grouped as follows using the concept of surplus set. That is, as shown in FIG. 4, when the small areas are serialized from left to right and top to bottom, and these serial numbers are divided by divisors of a predetermined size, the smaller areas corresponding to the same surplus set Group them into the same group.

In this way, one image is divided into several groups, and in the low frequency data recording area 120 of each small area belonging to the same group, not only the low frequency data of the corresponding small area but also the low frequency data of other small areas belonging to the same group. Record even together. The small area has a sync block representing itself. Therefore, while the same data is recorded in the low frequency data recording area 120 of the sync block of each of the small areas belonging to the same surplus set, the data unique to each small area, that is, different data, is recorded in the high frequency data recording area 110. Is recorded. In other words, the low-frequency data recording area 120 of each small area belonging to a certain carp set is recorded so that not only its low-frequency data but also low-frequency data of another small area belonging to the surplus set including the small area itself are recorded together. . At this time, only the high frequency data of the corresponding small region is recorded in the high frequency data recording region 110 of each small region.

This will be described in detail with reference to the example of FIG. 4 (b) and FIG. 5 as follows. In other words, assume that there are images made up of 16 small areas as shown in FIG. 4 (b). FIG. 4 (b) relates to the case where the dividend, that is, the serial number L of the small region is 16. FIG. In the case of MOD (L, 16), the low frequency data recording area 120 of each sync block records its own low frequency data. For example, in the case of MOD (L, 8), the data of the low frequency data recording area 120 of sync block 1 is the same as that of sync block 9. In the case of MOD (L, 4), sync blocks 1, 5, 9 , All of the low frequency data recording areas 120 have the same data.

In generalization of the above example, when MOD (L, n) = α, the serial numbers are α, n + α, 2n + α, ...., mn + α (where mn = L and m , m, α are positive integers). The same data is all recorded in the low frequency data recording area 120 of the sync block of the small areas. 5 shows this. In FIG. 5, one sink block is composed of a sink area, an ID area, a low frequency data recording area 120, a high frequency data recording area 110, and a parity area. A sink pattern is recorded in the sink area, an identification number of a specific small area in the ID area, and an error correction code in the parity area. In addition, the high frequency data recording region 110 records the high frequency data of the specific small region itself, while the low frequency data recording region 120 not only the low frequency data of the specific small region itself but also the other small region classified into the same group as the small region. Record even the low frequency data of the field. In the above description, the low frequency data may be only a DC component or may include data of the low frequency component of the DC component and the AC component.

On the other hand, the mapping table may be used to form an arbitrary pair, as illustrated in FIG. 4 (c), without using the concept of the surplus set MOD (L, S), which is standing as another method of grouping the small regions. It may be. Using an arbitrary mapping table can be a more advantageous and flexible way to improve the quality of the reproduced image in that small areas are irregular and evenly mixed. In other words, if the low frequency data recorded together is a signal of a small area evenly distributed throughout the image, the quality of the trick reproduction image will be improved as much. And such a recording method will have a different effect depending on the recording density of the tape. When the recording density of the tape is sufficiently high due to the development of technology, it is possible to record more low frequency data of a small area in one low frequency data recording area 120, thereby improving the quality of the trick reproduction image. It becomes possible.

As described above, the image recording method of the DVCR according to the present invention can fix the amount of data recorded on the tape, thereby making it possible to effectively use the interframe coding method for the DVCR. In addition, in the present invention, even if a recording medium passes at a certain speed, if only one sync block is restored, the data of another small area classified into the same group as the sync block can be restored at the same time. In addition, the present invention can further improve the quality of a trick-playing image by sharing and recording low frequency data of each small area belonging to the same group.

Claims (8)

I pictures encoded using information of the corresponding picture itself, B pictures encoded using motion compensated predictions from past and / or future reference pictures, and P encoded using motion compensated predictions from past reference pictures In an image recording method in which a group of images (GOP) constituted by arbitrarily combining images is recorded on a digital VCR tape as a unit, the low frequency data of each sync block of the I image is an I image or a P image of the tape. And at a predetermined interval on each track for the B picture, the high frequency data of each sync block of the I picture is sequentially recorded in an unrecorded blank area on the track for the I picture, and the P picture data is The low frequency data of the I picture is recorded on the track for the picture and sequentially recorded in the remaining area, and the data of the B picture is recorded on the track for the B picture. The image recording method of a digital VCR according to claim sequentially written in the low-frequency data is recorded on the remaining area. 2. The image recording method according to claim 1, wherein the image recording method records data in a data recording area of a predetermined size separately provided in an end recording area of each picture group (GOP) on a track in order to make the data amount of each picture group constant. The image recording method of the DVR. The image recording method according to claim 2, wherein the predetermined size is a tape recording area corresponding to a data storage capacity of an output buffer for controlling an output bit rate of compressed coded data. Compressed digital video signals of each small area of the image divided into a predetermined number of small areas are divided into low frequency data and high frequency data, and the low frequency data is recorded at regular intervals on a track of a tape, and the high frequency data is recorded in the remaining areas of the track. In the image recording method of a digital VCR for recording data sequentially, the method for recording low frequency data comprises classifying the predetermined number of small areas into a plurality of groups according to a predetermined method, and all the small belonging to each same group. And recording the low frequency data of the area together in the low frequency data recording area of each small area belonging to the same group. The digital VCR image recording method according to claim 4, wherein when the predetermined number of small areas is classified into a plurality of groups, the small areas distributed evenly over the image are included in the same group. The method according to claim 4, wherein when the predetermined number of small regions is classified into a plurality of groups, the predetermined number of small regions is assigned a serial number from the left to the right of the image, and the respective serial numbers are divided by a predetermined size. and a division obtained by dividing into (divisor) divides the small area of the same serial number into the same group. The method of claim 4, wherein the low frequency data includes only a DC component. 5. The image recording method according to claim 4, wherein the low frequency data includes a low frequency data portion of a direct current component and a high flow component.