JPH0993537A - Digital video signal recording and reproducing device and digital video signal coding method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a reproduced image without deteriorating image quality by recording an information quantity produced through coding at a fixed quantization level for a 1st coding in the unit of GOP and controlling the code quantity based on an object information quantity set by a 2nd coding. SOLUTION: A digital video signal is received from a video signal input circuit 1 in a coding for a 1st pass in the unit of lines and fed to a memory circuit 2. The code quantity is not controlled for the coding for the 1st pass but quantized by a quantization circuit 5. Furthermore, an information quantity of video data by 1 GOP and a motion vector from a buffer memory 12 is detected and outputted to a code quantity allocation circuit 15 in the unit of GOP. The code quantity allocation circuit 15 records the code quantity for 1st pass received in the unit of GOP to decide the object information quantity allocated for each GOP by the coding for the 2nd pass when the coding is finished. Then the code quantity is controlled by the code quantity control section so as to attain equal allocated code quantity for each GOP for the 2nd pass.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital video signal recording / reproducing apparatus and a digital video signal encoding method for recording / reproducing a digital video signal.

[0002]

2. Description of the Related Art FIG. 16 shows, for example, ISO-IEC / JTC1 / SC29 / WG.
11 is a block circuit diagram showing a configuration example of a recording system in the conventional video signal encoding system shown in 11 MPEG 92 / N0245 Test Model 2. FIG. In the figure, the digital video signal input from the input terminal 1 is input to the memory circuit 901. Video signal 921 output from the memory circuit 901
Are provided to the first input of the subtractor 902 and the second input of the motion compensation prediction circuit 910. The output of the subtractor 902 is passed through a DCT (discrete cosine transform) circuit 903,
It is input to the quantization circuit 904. This quantization circuit 904
Is output to the input of the transmission buffer 906 via the variable length coding circuit 905. The output of the transmission buffer 906 is output from the output terminal 2. On the other hand, the output of the quantization circuit 904 is sent to the IDC via the inverse quantization circuit 907.
It is also input to the T circuit (inverse discrete cosine transform) 908.
The output of the IDCT circuit 908 is the first output of the adder 909.
Given to the input of. The output 922 of the adder 909 is provided to the first input of the motion compensation prediction circuit 910. The output 923 of the motion compensation prediction circuit 910 is given to the second input of the adder 909 and the second input of the subtractor 902.

FIG. 17 is a block circuit diagram showing a structural example of a reproducing system in a conventional video signal encoding method. In the figure, the video signal input from the input terminal 20 is input to the reception buffer 1001. Receive buffer 100
The output from 1 is input to the variable length decoding circuit 1002, inversely quantized by the inverse quantization circuit 1003, and ID
The inverse discrete cosine transform is performed by the CT circuit 1004, and the result is given to the first input of the adder 1006. On the other hand, the output of the reception buffer 1001 is also input to the prediction data decoding circuit 1005, and the output of the prediction data decoding circuit 1005 is the adder 1
006 to the second output. The output of the adder 1006 is output from the output terminal 30 via the memory circuit 1007.

FIG. 18 is a conceptual diagram for explaining the operation of the memory circuit in the conventional video signal encoding system. In the figure, the output 922 of the adder 909 is supplied to the input terminal 1201a, and the output 921 of the memory circuit 901 is supplied to the input terminal 1201b. Input terminal 1201
The signal 922 input from a is input to the frame memory 1204a or the frame memory 1204b via the switch 1203. Frame memory 1204a
The reference image output from the motion vector detection circuit 12
Applied to the first input of 05a. The video signal 921 input from the input terminal 1201b is applied to the second input of the motion vector detection circuit 1205a. The output of the motion vector detection circuit 1205a is the prediction mode selector 12
It is input to 06.

On the other hand, the reference image output from the frame memory 1204b is given to the first input of the motion vector detection circuit 1205b. Motion vector detection circuit 120
The video signal 921 input from the input terminal 1201b is applied to the second input of 5b. The output of the motion vector detection circuit 1205b is given to the second input of the prediction mode selector 1206. The video signal 921 input from the input terminal 1201b is applied to the third input of the prediction mode selector 1206. The first output of prediction mode selector 1206 is provided to the first input of switch 1207. The 0 signal is provided to the second input of the switch 1207. The second input of the switch 1207 is provided with the second output of the prediction mode selector 1206. The output 923 of the switch 1207 is output from the output terminal 1202.

Next, the operation will be described. As one of high-efficiency coding schemes for coding a video signal, there is a hybrid coding scheme in which inter-picture predictive coding using motion compensation prediction and intra-picture transform coding are combined. The conventional example described here also employs the hybrid coding method. First, the outline of the inter-picture predictive coding will be described.

FIG. 19 is a conceptual diagram showing motion compensation prediction in the conventional video signal coding system. 20 is a conceptual diagram showing the operation of the memory circuit 901 in the conventional video signal encoding system.

In FIG. 19, each image is an intra-picture coded image (hereinafter referred to as I picture), a unidirectional predictive coded image (hereinafter referred to as P picture), and a bidirectional predictive coded image (hereinafter referred to as B picture). It is divided into three types. For example, if one image per N sheets is an I picture and one image per M sheet is a P picture or an I picture, n,
When m is an integer and 1 ≦ m ≦ N / M, (N × n +
The (M) th image is an I picture, the (N × n + M × m) th image (m ≠ 1) is a P picture, and (N × n + M × m +)
The 1st to (N × n + M × m + M−1) th images are B pictures. At this time, the (N × n + 1) th image to the (N × n + N) th image are collected and
It is called P (Group of Pictures).

FIG. 19 shows the case where N = 15 and M = 3. In the figure, the I picture does not perform inter-picture prediction, but only intra-picture transform coding. The P picture is predicted from the immediately preceding I picture or P picture. For example, the 6th image in the figure is a P picture, but this is predicted from the 3rd I picture. The 9th P picture in the figure is predicted from the 6th P picture. The B picture is predicted from the immediately preceding and succeeding I or P pictures.
For example, in the figure, the 4th and 5th B pictures are predicted from both the 3rd I picture and the 6th P picture. Therefore, the fourth and fifth images are encoded after the sixth image is encoded.

FIG. 16 is a block circuit diagram of the hybrid coding system. The digital video signal input from the input terminal 1 is input to the memory circuit 901. The memory circuit 901 rearranges the images in the order of encoding and outputs the images. That is, as described above, in FIG. 19, for example, the 1st B picture is coded after the 3rd I picture, so that the images are rearranged here. FIG.
0 indicates this rearrangement operation. The image sequence input as shown in FIG. 20A is output in the order of FIG.

The video signal 921 output from the memory circuit 901 is subtracted by the subtractor 9 in order to reduce redundancy in the time axis direction.
In 02, the difference between the image and the predicted image 923 output from the motion compensation prediction circuit 910 is calculated, and then DCT is performed in the spatial axis direction. The converted coefficient is used in the quantization circuit 904.
Is quantized by, and variable-length coded, and then output via the transmission buffer 906.

In the transmission buffer 906, the code amount is controlled so that the code amount of the variable length coded image data is substantially equal to the preset coding rate. That is, when the amount of information currently encoded in the transmission buffer 906 is higher than the set rate, a large quantization parameter is selected so that the amount of code generated in the quantization circuit 904 is reduced. Control the quantization parameter. On the other hand, if the amount of information currently encoded is less than the set rate, the quantization circuit 9
The quantization parameter is controlled so that a small quantization parameter is selected so that the code amount generated for 04 increases.

In practice, the quantizing circuit 904 is provided with 31 stages of quantizing parameters Q (m) (m = 1 to 31). For example, in the case of linear quantizing, Q (m) = 2m.
As shown in, the quantization step width is 2 and the quantization parameter has 31 steps from 2 to 62, and the transmission buffer 9
The DCT coefficient output from the DCT circuit 903 is quantized according to the quantization number m input from 06. However, the smaller the quantization number m, the smaller the quantization step and finer quantization, and the larger the quantization number, the larger the quantization step and coarser quantization.

On the other hand, the quantized transform coefficient is inversely quantized and subjected to IDCT and then added to the predicted image 923 by the post-adder 902 to obtain a decoded image 922. The decoded image 922 is input to the motion compensation prediction circuit 910 for encoding the next image.

Next, the operation of the motion compensation prediction circuit 910 will be described.
It will be described with reference to FIG. The motion compensation prediction circuit 910
Frame memory 1204a and frame memory 1204b
The two reference images stored in the memory circuit 90 are used.
1 performs motion-compensated prediction of the video signal 921 output from No. 1 and outputs a predicted image 923.

First, when the image 922 encoded and decoded as described above is an I picture or a P picture,
This image 922 is stored in the frame memory 1204a or 1204b for encoding the next image. At this time, the switch 1203 is switched so as to select one of the frame memory 1204a and the frame memory 1204b that has been updated earlier in time. If the decoded image 922 is a B picture, the frame memory 1204a and the frame memory 1
No writing to 204b is performed.

By such switching, for example, FIG.
When the 1st and 2nd B pictures of 9 are encoded,
Frame memory 1204a and frame memory 1204b
The P-picture No. 0 and the I-picture No. 3 are stored in each of the frames, and when the P-picture No. 6 is encoded and decoded after that, the frame memory 1204a rewrites the decoded picture of the No. 6 P-picture. To be Therefore, the next 4
When the No. 5 and No. 5 B pictures are encoded, the No. 6 P picture and No. 3 I picture are stored in the frame memory, respectively. Further, when the 9th P picture is encoded and decoded, the frame memory 120
4b is rewritten with the decoded image of the 9th P picture.
Therefore, when the 7th and 8th B pictures are encoded, the 6th P picture and the 9th P picture are stored in the frame memory, respectively.

When the video signal 921 output from the memory circuit 901 is input to the motion compensation prediction circuit 910, 2
Two motion vector detection circuits 1205a and 1205b,
Based on the reference images stored in the frame memories 1204a and 1204b, respectively, a motion vector is detected and a motion compensation predicted image is output. That is, the video signal 921 is divided into a plurality of blocks, a block having the smallest prediction distortion in the reference image is selected for each block, the relative position of the block is output as a motion vector, and Is output as a motion compensation prediction image.

On the other hand, the prediction mode selector 1206 selects the two motion-compensated prediction images output from the motion vector detection circuits 1205a and 1205b and the average image thereof having the smallest prediction distortion, and the prediction image is selected. Output as. At this time, if the video signal 921 is not a B picture, the motion compensation prediction image corresponding to the reference image input earlier in terms of time is always selected and output. In addition, the prediction mode selector 1206 selects one of the intra-picture coding that does not perform prediction and the inter-picture predictive coding that uses the selected predicted image, whichever has better coding efficiency.

At this time, if the video signal 921 is an I picture, the intra picture coding is always selected. When intra-picture coding is selected, a signal indicating the intra-picture coding mode is output as the prediction mode, and when inter-picture predictive coding is selected, the signal indicating the selected prediction picture is selected as the prediction mode. Is output. The switch 1207 outputs a 0 signal if the prediction mode output from the prediction mode selector 1206 is an intra-picture coding mode, and otherwise outputs a prediction image output from the prediction mode selector 1206. To do.

From the above, when the video signal 921 output from the memory circuit 901 is an I picture, the motion compensation prediction circuit 910 always outputs a 0 signal as the predicted image 923, so that the I picture performs inter-picture prediction. Instead, the image is intra-coded and encoded. Further, when the video signal 921 output from the memory circuit 901 is, for example, the No. 6 P picture in FIG. 19, the motion compensation prediction circuit 910 performs the motion compensation prediction from the No. 3 I picture in FIG. Image 923
Is output. Further, when the video signal 921 output from the memory circuit 901 is, for example, the 4th B picture in FIG. 19, the motion compensation prediction circuit 910 moves from the 3rd I picture and the 6th P picture in FIG. Compensation prediction is performed and a prediction image 923 is output.

Next, the operation at the time of decoding will be described with reference to FIG. The image data encoded with high efficiency is input to the input terminal 2
The variable length data is input to 0 and input to the variable length decoding circuit 1002 via the reception buffer. The variable length data is dequantized into fixed length data, subjected to IDCT, and output to the adder 1006. On the other hand, the prediction data decoding circuit 1005 decodes the prediction image according to the motion vector output from the reception buffer 1001, and the adder 10
It outputs to 06.

In this case, the prediction data decoding circuit 1005 is provided with a frame memory for storing the I picture and P picture data decoded by the adder 1006, like the motion compensation prediction circuit 910. At the time of P picture and B picture, , According to the input motion vector,
The predicted image is reproduced from the target I picture and P picture and output to the adder 1006. Note that the method of updating the reference image data at the time of the I picture and P picture is the same as that at the time of encoding, and therefore description thereof will be omitted.

In the adder 1006, the prediction data decoding circuit 1
The output of 005 and the output of the IDCT circuit 1004 are added,
It is output to the memory circuit 1007. Here, at the time of encoding, the time-sequential video signals are rearranged according to the encoding order as shown in FIG.
Therefore, in the memory circuit 1007, the data input in the order shown in FIG. 20B is rearranged so that the image data is temporally continuous in the order shown in FIG.

[0025]

Since the conventional video signal coding system is configured as described above, when the coded data is recorded on the storage medium having a constant information storage capacity, it is always constant. In order to perform the code amount control so that the coding is performed at a rate, a small code amount allocation can allocate a large code amount to a sufficient GOP,
Since a sufficient code amount allocation is not performed for a GOP that requires a large amount of code amount allocation, an appropriate code amount allocation is not performed for each GOP, so that the image quality deterioration of a reproduced image is easy to understand. In addition, there is a problem in that an appropriate code amount allocation cannot be performed because neither a scene in which image quality deterioration is easy to understand or a scene in which image quality deterioration is hard to understand is not detected.

The present invention has been made to solve the above problems, and it is possible to record / reproduce by performing variable rate code amount control in GOP units on a storage medium having a constant information storage capacity. In addition, since it is possible to determine whether or not the image quality deterioration is easy to understand and assign the code amount, it is an object to obtain a digital video signal recording / reproducing apparatus and a digital video signal encoding method capable of obtaining a reproduced image without image quality deterioration. To do.

[0027]

In the digital video signal recording / reproducing apparatus according to the present invention, the amount of information generated by encoding the input digital video signal with a fixed quantization level in the first encoding. Is recorded in GOP units, and the target information amount is set in GOP units so that the digital video signal fits in the target information amount using the generated information amount recorded in GOP units, and the second encoding is performed. The code amount control is performed according to the target information amount obtained in the GOP unit.

Further, in the first encoding of the input digital video signal, encoding is performed at a fixed quantization level, the generated information amount and time code are recorded in GOP units, and recorded in GOP units. Using the generated information amount, the target information amount is set in GOP units so that the digital video signal fits in the target information amount, and the target information amount and time obtained in the GOP unit in the second encoding are set. The code amount is controlled by a code.

Further, in the first encoding, the input digital video signal is encoded with a fixed quantization level, the information amount generated is stored in GOP units, and the generated information amount stored in the GOP units is stored. Is used to set the target information amount in GOP units so that the digital video signal fits in the target information amount, and the code amount is controlled by the target information amount obtained in the GOP units in the second encoding. It is configured in.

Further, the input digital video signal is set to 1
In the second encoding, the amount of information generated by encoding with a fixed quantization level is recorded in GOP units, and it is determined whether or not the GOP is prominent in image quality deterioration and recorded in GOP units. Using the recorded generated information amount and the determination result, the target information amount is set in GOP units so that the digital video signal fits within the target information amount,
The code amount control is performed according to the target information amount obtained in the GOP unit in the second encoding.

Furthermore, the input digital video signal is encoded at a fixed quantization level in the first encoding, the amount of information generated is recorded in GOP units, and GO
It is determined whether a panning scene is included in P, and G
Recording is performed in OP units, and the target information amount is set in GOP units so that the digital video signal fits in the target information amount using the generated information amount recorded in the GOP units and the determination result. The code amount control is performed according to the target information amount obtained by the GOP unit in the conversion.

Furthermore, in the first encoding, the input digital video signal is encoded with a fixed quantization level and the generated information amount is recorded in GOP units.
The brightness level of the entire picture is detected, and it is determined whether or not the GOP includes many pictures with a low brightness level.
Recording is performed in units, and the target information amount is set in GOP units so that the digital video signal fits in the target information amount using the generated information amount recorded in the GOP units and the determination result, and the second encoding is performed. The code amount control is performed according to the target information amount obtained in the GOP unit.

Further, in the digital video signal coding method according to the present invention, the input digital video signal is coded at a fixed quantization level in the first coding, and the amount of information generated is stored in GOP units. At the same time, it is determined whether or not the GOP is prominent in image quality deterioration, the GOP is stored in GOP units, and the generated image amount and the determination result stored in the GOP units are used so that the digital video signal falls within the target information amount. The target information amount is set for each GOP, and the code amount is controlled according to the target information amount obtained for each GOP in the second encoding.

Further, the input digital video signal is encoded at a fixed quantization level in the first encoding, and the generated information amount is stored in GOP units.
It is determined whether or not a panning scene is included in the GOP, the GOP is stored in GOP units, and the generated image amount stored in the GOP units and the determination result are used so that the digital video signal falls within a target information amount. The target information amount is set for each GOP, and the code amount is controlled according to the target information amount obtained for each GOP in the second encoding.

In addition, the input digital video signal is set to 1
In the second encoding, encoding is performed with a fixed quantization level,
The generated information amount is stored in GOP units, the brightness level of the entire picture is detected, it is determined whether or not the GOP includes many pictures with low brightness levels, the GOP units are stored, and the GOP units are stored. Using the generated information amount and the determination result, the target information amount is set in GOP units so that the digital video signal fits in the target information amount,
The code amount control is performed according to the target information amount obtained in the GOP unit in the second encoding.

[0036]

BEST MODE FOR CARRYING OUT THE INVENTION In a digital video signal recording / reproducing apparatus according to an embodiment of the present invention, when an input digital video signal is encoded at a fixed quantization level in the first encoding, The generated information amount is recorded in GOP units, and the target information amount is set in GOP units so that the digital video signal fits within the target information amount using the generated information amount recorded in GOP units. Since the code amount control is performed based on the target information amount obtained in the GOP unit in the second encoding, it is possible to realize the efficient variable rate code amount control.

Further, the input digital video signal is set to 1
The amount of information and the time code generated when encoding is performed at a fixed quantization level in the second encoding are recorded in GOP units, and the generated image amount recorded in the GOP units is used to generate the digital video signal. Is set to a target information amount so that the target information amount is set in a GOP unit, and the code amount is controlled by the target information amount and the time code obtained in the GOP unit in the second encoding. Efficient variable rate code amount control can be realized without deviation between the GOP to which the amount allocation is applied and the GOP to be actually encoded.

Further, the amount of information generated when the input digital video signal is encoded at the fixed quantization level in the first encoding is stored in GOP units, and the GO is stored.
Using the generated information amount recorded in P units, the target information amount is set in GOP units so that the digital video signal fits in the target information amount, and the target obtained in GOP units in the second encoding. Since the code amount control is performed according to the information amount, efficient variable rate code amount control can be realized.

Further, the input digital video signal is set to 1
The amount of information that occurs when encoding is performed at a fixed quantization level in the second encoding and the determination result of whether or not the GOP is prominent in image quality deterioration are recorded in GOP units, and are recorded in the GOP units. Using the information amount and the determination result, the target information amount is set in GOP units so that the digital video signal falls within the target information amount, and the GOP is set in the second encoding.
In order to control the code amount according to the target information amount obtained in units, a large amount of target information amount is allocated to a GOP where image quality deterioration is noticeable, and a small target information amount is allocated to a GOP where image quality deterioration is not noticeable. It is possible to realize efficient variable rate code amount control with less noticeable deterioration.

Further, the amount of information generated when the input digital video signal is encoded at the fixed quantization level in the first encoding is recorded in GOP units, and
It is determined whether or not a panning scene is included in the GOP, recording is performed for each GOP, and the generated image information amount and the determination result recorded for each GOP unit are used so that the digital video signal falls within a target information amount. Since the target information amount is set for each GOP and the code amount is controlled according to the target information amount obtained for each GOP unit in the second encoding, the image quality deterioration is likely to occur even in the panning scene where the image quality deterioration is easily noticeable. It is possible to realize efficient variable rate code amount control without occurrence of error.

Furthermore, the amount of information generated when the input digital video signal is encoded at a fixed quantization level in the first encoding is recorded in GOP units, and the brightness level of the entire picture is recorded. The GOP is detected and it is determined whether or not the GOP includes many pictures with a low luminance level.
Recording is performed in OP units, and the target information amount is set in GOP units so that the digital video signal fits in the target information amount using the generated information amount recorded in the GOP units and the determination result. In order to control the code amount according to the target information amount obtained by the GOP unit in the conversion, the target code amount can be reduced in a scene where the luminance signal level where the image quality deterioration is difficult to understand is small.
Since the target code amount of other scenes can be increased, efficient code rate control of the variable rate can be realized without causing image quality deterioration.

In the digital video signal encoding method according to the present invention, the amount of information and the image quality generated when the input digital video signal is encoded at the fixed quantization level in the first encoding. GOP that is prone to deterioration
The determination result as to whether or not it is recorded is recorded in GOP units, and the GO is adjusted so that the digital video signal falls within the target information amount by using the generated information amount recorded in GOP units and the determination result.
The target information amount is set in P units, and the G
In order to control the code amount based on the target information amount obtained in OP units, a large amount of target information amount is allocated to a GOP where image quality deterioration is noticeable, and a small target information amount is allocated to a GOP where image quality deterioration is not noticeable. It is possible to realize efficient variable rate code amount control with less noticeable image quality deterioration.

Further, the amount of information generated when the input digital video signal is encoded at a fixed quantization level in the first encoding is recorded in GOP units, and
It is determined whether or not a panning scene is included in the GOP, recording is performed for each GOP, and the generated image information amount and the determination result recorded for each GOP unit are used so that the digital video signal falls within a target information amount. Since the target information amount is set for each GOP and the code amount is controlled according to the target information amount obtained for each GOP unit in the second encoding, the image quality deterioration is likely to occur even in the panning scene where the image quality deterioration is easily noticeable. It is possible to realize efficient variable rate code amount control without occurrence of error.

Further, the input digital video signal is set to 1
The amount of information generated when encoding is performed with a fixed quantization level in the second encoding is recorded in GOP units, and the luminance level of the entire picture is detected to determine whether the GOP includes many pictures with low luminance levels. It is determined whether or not it is recorded in GOP units, and the target information amount is set in GOP units so that the digital video signal fits in the target information amount using the generated information amount recorded in GOP units and the determination result. In order to configure the code amount control according to the target information amount obtained in the GOP unit in the second encoding,
Since it is possible to reduce the target code amount for a scene where the luminance signal level where image quality deterioration is difficult to understand is small, it is possible to increase the target code amount for other scenes.
Efficient variable rate code amount control can be realized without image quality deterioration.

Hereinafter, the present invention will be described in detail with reference to the drawings showing the embodiments. Embodiment 1. 1 is a block circuit diagram of a recording system of a digital video signal recording / reproducing apparatus according to a first embodiment of the present invention. In the figure, the digital video signal output from the video signal input circuit 1 is input to the memory circuit 2. The video signal output from the memory circuit 2 is applied to the first input to the subtractor 3 and the second input to the motion compensation prediction circuit 11. The output of the subtractor 3 is input to the first input of the quantization circuit 5 via the DCT circuit 4. The output of the quantization circuit 5 is given to the first input of the buffer memory 12 via the variable length coding circuit 6. On the other hand, the output of the quantization circuit 5 is also input to the IDCT circuit 8 via the inverse quantization circuit 7. The output of the IDCT circuit 8 is given to the first input of the adder 10.

The output of the adder 10 is the motion compensation prediction circuit 1
1 to the first input. The first output of the motion compensation prediction circuit 11 is the second input of the adder 10 and the subtractor 4
To the second input. Also, the motion compensation prediction circuit 11
The second output of is applied to the second input of the buffer memory 12. The second output of the buffer memory is input to the code amount allocation circuit 15, and the output of the code amount allocation circuit 15 is provided to the first input of the code amount control circuit 16. Also,
The third output of the buffer memory 12 is the code amount control circuit 16
Of the code amount control circuit 16 and the output of the code amount control circuit 16 is output to the second input of the quantization circuit. On the other hand, the first output of the buffer memory 12 is recorded in the storage medium 14 such as an optical disk.

FIG. 2 is a block circuit diagram showing a decoding unit in the video signal recording / reproducing apparatus according to the first embodiment of the present invention. In the figure, the video information read from the storage medium 20 such as an optical disk is input to the buffer memory 21. The first output from the buffer memory 21 is input to the variable length decoding circuit 22, dequantized in the dequantization circuit 23, subjected to the inverse discrete cosine transform in the IDCT circuit 24, and the first output of the adder 25. Given to the input of. On the other hand, the output of the buffer memory 21 is also input to the prediction data decoding circuit 26, and the prediction data decoding circuit 26
Is provided to the second input of adder 25. The output of the adder 25 is output from the video signal output circuit 28 via the memory circuit 27.

Next, the operation will be described. In the present application, in order to encode the video signal at a variable rate, the encoding is performed by the 2-pass method. That is, in the encoding of the first pass, quantization is performed at a fixed rate, the code amount is calculated in GOP units, and the target code amount is determined for each GOP. Further, in the second pass, the code amount is controlled at a variable rate according to the target code amount for each GOP obtained in the first pass, and encoding is performed.

In the first pass encoding, the digital video signal is input from the video signal input circuit 1 line by line and supplied to the memory circuit 2. However, in the motion-compensated prediction, as shown in FIG.
Frames, 1 I picture, 4 P pictures (P1 to P4), 10 B pictures (B1 to B10)
Predictive coding is performed as. In this case, similarly to the conventional example, the memory circuit 2 rearranges the sequentially input video data in the order shown in FIG.

Further, in the memory circuit 2, the data input in line units is rearranged in block units of 8 × 8 pixels, and macro blocks (adjacent 4 blocks) as shown in FIG.
One luminance signal Y block and two positions corresponding to these Y blocks
One color difference signal Cr, Cb block is formed in total (6 blocks), and data is output in macro block units. Here, a macroblock is a minimum unit of motion compensation prediction, and a motion vector for motion compensation prediction is obtained in macroblock units. Further, as shown in FIG. 6A, a band composed of 45 macroblocks (MB) in the horizontal direction is called a slice, and one picture is shown in FIG. 6B.
It is composed of 30 slices as shown in FIG.

The output of the memory circuit 2 is input to the subtractor 3 and the motion compensation prediction circuit 11, but the subtractor 3 and the DCT
Since the operation of the circuit 4 is the same as that of the conventional embodiment, its explanation is omitted. Here, since the code amount control is not performed in the first pass encoding, the quantization circuit 5 performs the quantization at a fixed quantization rate. For example, in the quantization circuit 5, 3
Quantization is performed at the quantization level of Q (16) which is in the middle of one level of quantization.

The output of the quantization circuit 5 is input to the inverse quantization circuit 7 and the variable length coding circuit 6, and the variable length coding 6,
The operations of the inverse quantization circuit 7, the IDCT circuit 8, the adder 10 and the motion compensation prediction circuit 11 are the same as those in the conventional embodiment, and therefore their explanations are omitted.

The video data output from the variable length coding circuit 6 and the motion vector output from the motion compensation prediction circuit 11 are input to the buffer memory 12. Here, in the first pass, the buffer memory 12 detects the amount of information of 1 GOP worth of video data and the motion vector and outputs it to the code amount allocation circuit 15 in GOP units.

The code amount allocation circuit 15 records the code amount of the first pass input in GOP units, and when the encoding of the first pass is completed, the target to be assigned to each GOP in the encoding of the second pass. Determine the amount of information. Here, the method of determining the target information amount for each GOP is as follows: Do is the information amount assigned to the entire video data, information amount D (m) for each GOP generated in the first pass (m: GOP number), 1 The sum Damt of the information amount generated at the pass and the target information amount D2 (m) to be assigned to each GOP are calculated as D2 (m) = Do × (D (m) / Damt).

Next, the encoding operation of the second pass will be described. In the second pass, DC from the video signal input circuit 1
The operation up to the T circuit 4 is the same as that of the first pass, and therefore its explanation is omitted. In the second pass, encoding is performed according to the target code amount for each GOP obtained in the first pass.
That is, the quantization circuit 5 performs quantization in accordance with the quantization parameter determined by the code amount control circuit 16.

Here, in the code amount control circuit 16, each GOP obtained by the code amount allocation circuit 15 in the first pass.
From the target code amount for each and the current remaining amount of the buffer memory output from the buffer memory 12, an appropriate parameter is determined from the quantization parameters of 31 stages and the quantization circuit 5
Output to

For example, the control unit of the code amount control is performed in slice units. That is, in the slice immediately after the start of encoding, encoding is performed using the quantization parameter that is initialized, but when the encoding for one slice is completed,
The generated code amount is measured and the quantization parameter is controlled according to the error from the target code amount in GOP units output from the code amount allocation circuit 15.

Actually, the target code amount assigned for each GOP is Ta, the slice number is m, and the code amount generated up to the m-1 slice which is being encoded is S (m-
1), if the error with the target code amount generated up to the (n-1) th GOP is Dp (n), the error D (n, m) with the target code amount in the current slice is D (n, m) = Dp (n) + S (m−1) −Ta × (m
-1) / 30.

Furthermore, the current quantization parameter is set to Q
(T), the threshold for changing the quantization parameter is set to td, and the quantization parameter is changed as td = D (n, m) / th t = t + td. However, since the quantization parameter has 31 levels, the value of t is changed from 1 to 31.

The quantization circuit 5 quantizes the DCT coefficient for each slice according to the quantization parameter input from the code amount control circuit 16. Here, the output of the quantizing circuit 5 is input to the inverse quantizing circuit 7, but since the operation from the inverse quantizing circuit 7 to the motion compensating circuit 11 is the same as the operation of the first pass, its explanation is omitted. .

The output of the quantization circuit 5 is also input to the buffer memory 12 via the variable length coding circuit 6. In the buffer memory 12, the motion vector output from the motion compensation circuit 11 and the data output from the variable length coding circuit 6 are multiplexed and the generated code amount for each slice is output to the code amount control circuit 16 and the storage medium 14 The video data encoded in is output and recorded. Note that the decoding unit is the same as that of the conventional example, and therefore its description is omitted.

The operation of the encoding method will be described below according to the flowchart shown in FIG. The video signal input in FIG. 7 is rearranged for each picture in the picture rearrangement unit 61 as shown in FIG. 3, and further blocked as shown in FIG. 5 and output to the motion compensation prediction unit 62. It The motion-compensated prediction unit 62 does not perform motion-compensated prediction on the I picture but performs orthogonal transformation by DCT, and performs motion-compensated prediction on the P and B pictures with respect to the reference image and performs DCT on the difference error between them to perform quantization on the quantization unit. Output to 63. Further, the data quantized by the quantization unit 63 is input to the motion compensation prediction unit 62 again, subjected to dequantization and IDCT, and used as reference data for motion compensation prediction.

Next, the quantizing unit 63 quantizes in a fixed quantizing step in the first pass and outputs it to the variable length coding unit 64. In the variable length coding unit 64, GOP is used in the first pass.
The generated code amount is calculated for each and output to the code amount allocation unit 66. The code amount allocation unit 66 stores the code amount generated in GOP units, and when the first pass is completed, the information amount (average bit rate) allocated to the entire image data is generated in GOP units in the first pass. The allocation code amount for each GOP is determined by dividing according to the information amount.

On the other hand, in the second pass, the quantization unit 63 performs quantization according to the quantization parameter input by the code amount control unit 67. Here, in the code amount control unit, each G
The code amount control is performed so that the OP code amount becomes equal to the allocated code amount for each GOP determined by the code amount allocation unit 66. However, the control of the code amount is performed by the variable length coding unit 6
The code amount control is performed according to the generated code amount in units of slices output from No. 4. Further, the variable-length coding unit 64 successively outputs the variable-length-coded video data that is code amount controlled in the second pass.

As described above, when variable rate coding is performed, the code amount is trial calculated using the fixed quantization parameter in the first pass, and the target code amount is assigned to each GOP, and the second pass is used. By performing code amount control by using the target code amount in encoding, it is possible to realize accurate variable rate encoding, and to realize a low-rate and high-quality digital video signal encoding device. it can.

In the above embodiment, the code amount control circuit 16 controls the code amount in slice units. However, it is not always necessary to control the code amount in slice units and may be performed in m macroblock units (m> 1). . Further, although the code amount allocation circuit 15 allocates the code amount for each GOP, it is not always necessary to allocate each GOP, and the target code amount may be allocated in nGOP units (n> 1).

As shown in FIG. 8, the video signal input circuit 1
The time code of the video data that is currently encoded is input to the code amount control circuit 19 and the code amount allocation circuit 18, and when the first pass is encoded, the code amount allocation circuit 18 sets the code amount for each GOP. At the same time, the time code is recorded, and at the time of the second pass encoding, the code amount control circuit 19 refers to the time code recorded in the code amount allocation circuit 18 and the currently encoded time code to encode the time code. By doing so, it is possible to accurately perform variable rate code amount control without deviation between the first pass encoding and the second pass encoding.

Embodiment 2 9 is a block diagram showing a recording system of a digital video signal recording / reproducing apparatus according to a second embodiment of the present invention. In the first embodiment, the code amount is allocated for each GOP by the code amount generated by using the fixed quantization step in the first pass, but as shown in FIG. 9, it is easy to understand the image quality deterioration for each GOP. It is determined whether or not the image quality deterioration is easily recognized, and a large amount of target code is assigned to the GOP.
The OP may be assigned a small target code amount. In FIG. 9, 30 is a memory circuit, 31 is an image determination circuit, 32 is a code amount control circuit, and 33 is a code amount allocation circuit.

Next, the operation will be described. The video data input from the video signal input circuit 1 is input to the memory circuit 30, and after the data is rearranged and blocked by the memory circuit 30, the subtractor 3 and the motion compensation circuit 1
Input to 1. The operation from the subtractor 3 to the buffer memory 12 is the same as that in the first embodiment, and therefore its explanation is omitted.

On the other hand, the output of the memory circuit 30 is also input to the image determination circuit 31. The image determination circuit 31 determines for each GOP in the first pass whether or not the GOP is prominent in image quality deterioration. Here, a panning scene is a scene in which deterioration in image quality is easily understood. A panning scene is a scene that occurs when the camera is panned at the time of shooting, for example, a scene where one person walks in a movie or the like to follow by panning the camera or a landscape outside from a moving car or train There are scenes and the like, and the entire background moves in a fixed direction.

FIG. 10 is a block diagram showing the image determination circuit 31 according to the second embodiment. In the figure, 40 is a blocking circuit, 41 is a motion vector detection circuit, 42 is a determination circuit, and 43 is a frame memory. Next, the operation of the image determination circuit 31 will be described with reference to FIG. The video data input from the memory circuit 30 is the blocking circuit 4
0 is input to the frame memory 43.

Here, in the panning scene, since the entire background moves in a fixed direction, the search is performed with a larger area size than the search of the motion vector with the macroblock size (16 pixels × 16 lines) as in the motion compensation prediction. It is more effective to do it. Therefore, in the blocking circuit 40, as shown in FIG. 11, blocking is performed in an area of 180 pixels × 160 lines that divides one picture into 12, and the result is output to the motion vector detection circuit 41.

On the other hand, the frame memory 43 stores image data of one frame before, and the blocking circuit 4
For the blocks output from 0, the reference pattern is output within a range of horizontal ± 15 pixels and vertical ± 15 lines, for example. Further, in the motion vector detection circuit 41, horizontal ± 15 pixels and vertical ± 15 pixels output from the frame memory 43 are output.
A vector that minimizes the difference between the reference pattern in the line range and the output of the blocking circuit 40 is selected as the motion vector, and the motion vector and its difference value are output to the determination circuit 42.

The determination circuit 42 detects panning based on the motion vector for one picture output from the motion vector detection circuit 41 and the difference value, and determines whether or not a panning scene is included for each GOP. Panning judgment is 1
The two motion vectors are mv (i) and the difference value dif (i)
(I = 1 to 12), 12 motion vectors mv
(I) / nmv and difference value dif (i) / ndi (nmv
And ndi are integers, for example, nmv = 2, ndi = 120
If 6 or more of these values are the same, this picture is determined to be a panning scene. Furthermore, 1 GOP
When it is determined that 7 or more consecutive pictures out of 15 pictures are panning, this GOP is a GO including a panning scene.
Judge as P.

The determination result of the image determining circuit 31 is output from the buffer memory 12 together with the amount of information when encoded in a fixed quantization step and the code amount allocating circuit 33.
Is input to The code amount allocation circuit 33 records the code amount of the first pass input in GOP units, but the GOP including the panning scene has 0.5 in the generated information amount.
The value added with Mbit is recorded as the generated code amount. Further, the total code amount generated in the same manner as in the first embodiment is obtained at the time when the encoding of the first pass is completed, and each G is encoded at the time of the encoding of the second pass.
Determine the target amount of information to allocate to the OP. Where each G
Since the method of determining the target information amount for each OP and the operation of the code amount control circuit 32 in the second pass are the same as those in the first embodiment, the description thereof will be omitted.

The operation of the encoding method according to the second embodiment will be described below with reference to the flowchart shown in FIG. The video signal input in FIG. 12 is rearranged for each picture in the picture rearrangement unit 61 as shown in FIG. 3, further blocked, and output to the motion compensation prediction unit 62. The motion-compensated prediction unit 62 does not perform motion-compensated prediction on the I picture but performs orthogonal transformation by DCT, and performs motion-compensated prediction on the P and B pictures with respect to the reference image and performs DCT on the difference error between them to perform quantization on the quantization unit. Output to 63. Further, the data quantized by the quantization unit 63 is input to the motion compensation prediction unit 62 again, subjected to dequantization and IDCT, and used as reference data for motion compensation prediction.

On the other hand, the output of the picture rearrangement unit 61 is input to the image determination unit 71, and it is determined for each GOP whether the image quality deterioration is conspicuous or not, and output to the code amount assignment unit 70. Here, the image determination unit 71 detects, for example, a panning scene, and is configured by a flowchart as shown in FIG. The image data input in FIG. 13 forms a block that divides one screen into 12 blocks as shown in FIG. 11 in the blocking unit 73, and outputs the block to the motion vector detection unit 74. On the other hand, the input video data is also input to the reference pattern generation unit 76 and delayed by one frame, and the reference pattern for each block output by the blocking unit 73 is output.

The motion detecting section 74 obtains the difference between the block input from the blocking section 73 and the reference pattern input by the reference pattern generating section 76, obtains the minimum difference value and the motion vector, and outputs it to the determining section 75. Judgment unit 7
In step 5, first, panning is determined for each picture based on the motion vector for one picture and its difference value. Furthermore, when it is determined that seven consecutive pictures are panning scenes, the input GOP is determined to be a panning scene and is output.

In the first pass, the quantizing unit 63 quantizes at a fixed quantizing step and outputs it to the variable length coding unit 64. In the variable length coding unit 64, GOP is used in the first pass.
The generated code amount is calculated for each and output to the code amount allocation unit 70. The code amount allocation unit 70 stores the code amount generated in GOP units. At this time, if the image determination unit 71 determines that the scene is a panning scene, the information amount of 0.5 Mbit is added to the generated code amount. , Control to increase allocation of information amount to panning scene. Further, when the first pass is completed, the information amount assigned to the entire image data is divided according to the information amount generated in the GOP unit in the first pass, so that each GOP is divided.
Determine the amount of code assigned to

On the other hand, in the second pass, the quantization unit 62 performs quantization according to the quantization parameter input by the code amount control unit 67. Here, the code amount control unit 67 controls the code amount so that the code amount of each GOP becomes equal to the assigned code amount of each GOP determined by the code amount assigning unit 70. However, the control of the code amount is performed according to the generated code amount for each slice output from the variable length coding unit 64. Further, the variable-length coding unit 64 successively outputs the variable-length-coded video data that is code amount controlled in the second pass.

As described above, in the second embodiment, the amount of generated information is 0.5 in the GOP including the panning scene.
Since the value obtained by adding Mbit is used as the generated code amount, the target image amount is allocated to the GOP including the panning scene in which the image quality deterioration is easy to understand, and the target image information amount is allocated more than in the case of the normal GOP, so that the image quality deterioration is difficult to understand.

In the above embodiment, the picture quality judgment circuit 31 divides one picture into 12 to obtain the motion vector, but it is not always necessary to divide the picture into 12. Further, the determination of panning is made when the motion vectors and the difference values of 6 or more match, but it is not always necessary that the number of motion vectors and the difference are 6 or more. Further, the GOP in the case where 7 consecutive pictures are panning scenes is determined to be panning, but it is not always the case that the 7 pictures are consecutive, and when there are discontinuous n pictures (n integers) or more, there are panning pictures. GOP may be determined as panning. Further, in the above embodiment, the generated code amount is a value obtained by adding 0.5 Mbit to the information amount generated for the GOP including the panning, but it is not necessarily 0.5 Mbit, and any information amount can be added. Good

Third Embodiment In the second embodiment, the image determination circuit 31 detects only the panning scene in which the image quality deterioration is easily recognized. On the contrary, the overall low brightness scene in which the image quality deterioration is hard to be detected is detected, and the target for these GOPs is detected. The control may be performed so as to reduce the code amount. FIG. 14 is a conceptual diagram for explaining the image determination according to the third embodiment. In the figure, 50 is a brightness level detection circuit, and 51 is a determination circuit.

Next, the operation of the image determination circuit 31 will be described. The image determination circuit 31 determines for each GOP in the first pass whether or not the GOP has no noticeable deterioration in image quality. The image data input from the memory circuit 30 is input to the brightness level detection circuit 50, and the average brightness level for one picture is detected. Here, in general, in a scene with low brightness, the image quality deterioration is not noticeable. Therefore, in the determination circuit 51, the brightness level Yp (i) (i = 1 to 15) output from the brightness level detection circuit 50 in units of pictures is Yp <Ypth.
When there are eight or more pictures (for example, ypth = 60), it is determined that this GOP is a GOP in which image quality deterioration is not noticeable.

The determination result of the image determination circuit 31 is input to the code amount allocation circuit 33 together with the amount of information output from the buffer memory 12 when encoded in a fixed quantization step. The code amount allocation circuit 33 records the code amount of the first pass input in units of GOP, but the image determination circuit 31 determines that the image quality deterioration is less noticeable.
For P, a value obtained by subtracting 0.5 Mbit from the generated information amount is recorded as the generated code amount. Further, the total code amount generated in the same way as in the first embodiment is obtained at the time when the encoding of the first pass is completed, and the target information amount to be assigned to each GOP is determined in the second pass. However, the method of determining the target information amount for each GOP and the operation of the code amount control circuit 32 in the second pass are the same as those in the first embodiment, and therefore their explanations are omitted.

The operation of the encoding method according to the third embodiment will be described below with reference to the flowcharts shown in FIGS. 12 and 15. The operation from the picture rearrangement unit 61 to the variable length coding unit 64 of the input video signal in FIG. 12 is the same as that in the second embodiment, and therefore the description thereof is omitted.

The output of the picture rearrangement unit 61 is input to the image determination unit 71, and it is determined for each GOP whether the image quality deterioration is conspicuous or not, and then output to the code amount assignment unit 70. Here, the image determination unit 71 determines, for example, whether or not the GOP is composed of a picture having a low luminance level as a whole whose image quality deterioration is difficult to understand, and is configured by a flowchart as shown in FIG.

In the image data input in FIG. 15, the brightness level detecting section detects the average brightness level of each picture and outputs it to the judging section 79. When the luminance level input for each picture is smaller than a certain value, the determination unit 79 determines that the picture quality deterioration is difficult to recognize, and further determines that the picture quality deterioration is difficult to recognize is 1 GOP.
When there are 8 or more pictures in the GOP, the GOP is determined to be a GOP whose image quality deterioration is difficult to recognize and is output.

In the first pass, the quantizing unit 63 quantizes in a fixed quantizing step and outputs it to the variable length coding unit 64. In the variable length coding unit 64, GOP is used in the first pass.
The generated code amount is calculated for each and output to the code amount allocation unit 70. The code amount allocation unit 70 stores the code amount generated in GOP units. At this time, if the image determination unit 71 determines that the image quality deterioration is easy to understand, the information amount of 0.5 Mbit should be subtracted from the generated code amount. Thus, the control is performed so as to reduce the information amount allocation to the GOP whose luminance level is low and whose deterioration in image quality is difficult to see. Further, when the first pass is completed, the amount of information assigned to the entire image data is divided according to the amount of information generated in GOP units in the first pass to determine the assigned code amount for each GOP.

On the other hand, in the second pass, the quantization unit 62 performs quantization according to the quantization parameter input by the code amount control unit 67. Here, in the code amount control unit, each G
The code amount control is performed so that the OP code amount becomes equal to the allocated code amount for each GOP determined by the code amount allocation unit 70. However, the control of the code amount is performed by the variable length coding unit 6
The code amount control is performed according to the generated code amount in units of slices output from No. 4. Further, the variable-length coding unit 64 successively outputs the variable-length-coded video data that is code amount controlled in the second pass.

As described above, in the third embodiment, the generated code amount is a value obtained by subtracting 0.5 Mbit from the generated information amount for the GOP for which it is determined that the image quality deterioration is difficult to understand. Although the target information amount is small, the image quality deterioration is difficult to understand. Further, since the target code amount reduced from the GOP which is determined to be difficult to understand the image quality is allocated to other GOPs, the image quality is not deteriorated as a whole image, and efficient variable rate code amount control is realized. it can.

In the above embodiment, the image quality judgment circuit 3
In No. 1, the GOP in the case where it is determined that 8 pictures or more is the picture in which the image quality deterioration is not noticeable is determined to be the GOP in which the image quality deterioration is difficult to understand. When there is a picture whose image quality deterioration is hard to see, the GOP may be determined as a GOP whose image quality deterioration is hard to see. Further, in the above-described embodiment, GO is determined to be difficult to understand the deterioration of image quality.
The generated code amount is a value obtained by subtracting 0.5 Mbit from the amount of information generated for P, but it is not always 0.5 Mbit.
Does not have to be, and any amount of information may be subtracted.

[0093]

Since the present invention is constructed as described above, it has the following effects.

The input digital video signal is coded by a once fixed quantization step, and the generated code amount is GOP.
The target information amount is recorded in units of GOP, the target information amount is set in units of GOP so that the digital video signal fits within the target information amount by referring to the generated information amount recorded in units of GOP,
Since the code amount control is performed according to the target information amount obtained in units, efficient variable rate encoding can be performed.

In addition, the input digital video signal is set to 1
The digital video signal is encoded into a target information amount by referring to the generated information amount recorded in GOP units and encoding the generated code amount and time code in a fixed quantization step. As described above, the target information amount is set in GOP units, and the code amount control is performed based on the target information amount and time code obtained in GOP units. It is possible to perform efficient variable rate encoding without shifting the GOP to be encoded.

Further, the input digital video signal is encoded by a once-fixed quantization step, the generated code amount is recorded in GOP units, and the generated information amount recorded in GOP units is referred to for the digital signal. Efficient variable-rate coding because the target information amount is set in GOP units so that the video signal fits in the target information amount, and the code amount control is performed according to the target information amount obtained in GOP units. It can be performed.

Further, the input digital video signal is first encoded at a fixed quantization level, the generated information amount is recorded in GOP units, and it is determined whether or not the GOP is prominent in image quality deterioration. Recording is performed in units, and the target information amount is set in GOP units so that the digital video signal fits in the target information amount using the generated information amount recorded in the GOP units and the determination result, and is calculated in the GOP units. Since the code amount control is performed according to the target information amount that is set, a larger amount of code amount is assigned to a GOP in which image quality deterioration is easy to understand, and a larger amount of code amount is assigned to a GOP in which image quality deterioration is hard to understand than a normal GOP. Since a small amount of code is assigned, it is possible to perform efficient variable rate encoding in which deterioration of image quality is not noticeable.

Further, the input digital video signal is first encoded with a fixed quantization level, the amount of information generated is recorded in GOP units, and it is determined whether or not a panning scene is included in the GOP. Recording is performed in GOP units, and the target information amount is set in GOP units so that the digital video signal falls within the target information amount using the generated information amount recorded in the GOP units and the determination result. Since the code amount control is performed according to the target information amount obtained in step 1, the GOP including the panning scene in which the image quality deterioration is easy to understand is assigned a larger code amount than the normal GOP, and the image quality deterioration is less noticeable and efficient. Variable rate encoding can be performed.

Furthermore, the input digital video signal is first encoded with a fixed quantization level, the amount of information generated is recorded in GOP units, and the luminance level of the entire picture is detected to reduce the luminance level. It is determined whether or not the GOP includes many pictures, the GOP is recorded in GOP units, and the generated image amount recorded in the GOP unit and the determination result are used so that the digital video signal fits in the target information amount. Set the target amount of information with
Since the code amount control is performed according to the target information amount obtained in OP units, the target code amount of the GOP whose image quality deterioration is difficult to see is reduced and assigned to the normal GOP.
It is possible to perform efficient variable-rate encoding in which image quality deterioration is less noticeable.

According to the digital video signal coding method of the present invention, the input digital video signal is first coded at a fixed quantization level, and the generated information amount is recorded in GOP units. It is judged whether or not the GOP is prominent in image quality deterioration and recorded in GOP units.
Using the generated information amount recorded in the GOP unit and the determination result, the target information amount is set in the GOP unit so that the digital video signal falls within the target information amount.
Since the code amount control is performed according to the target information amount obtained in units, a GOP in which image quality deterioration is easy to understand is assigned a larger code amount than a normal GOP, and a GOP in which image quality deterioration is difficult to understand is normal. Since a code amount smaller than that of GOP is allocated, it is possible to perform efficient variable rate coding in which image quality deterioration is less noticeable.

Further, the input digital video signal is first encoded with a fixed quantization level, the amount of information generated is recorded in GOP units, and it is determined whether or not a panning scene is included in the GOP. Recording is performed in GOP units, and the target information amount is set in GOP units so that the digital video signal falls within the target information amount using the generated information amount recorded in the GOP units and the determination result.
Since the code amount control is performed according to the target information amount obtained in OP units, a GOP including a panning scene in which the image quality deterioration is easy to understand is assigned a larger code amount than a normal GOP, and the image quality deterioration is less noticeable and efficient. It is possible to perform variable-rate encoding with good accuracy.

Further, the input digital video signal is first encoded with a fixed quantization level, the amount of information generated is recorded in GOP units, and the brightness level of the entire picture is detected, and the picture with a low brightness level is detected. Is recorded in GOP units, and by using the generated information amount recorded in the GOP unit and the determination result, the digital video signal is recorded in GOP units so that the digital image signal fits in the target information amount. Since the target information amount is set and the code amount is controlled according to the target information amount obtained in the unit of GOP, the target code amount of the GOP whose image quality deterioration is difficult to see is reduced and assigned to the normal GOP. It is possible to perform variable-rate coding that is inconspicuous and efficient.

[Brief description of drawings]

FIG. 1 is a block circuit diagram showing a recording system of a digital video signal recording / reproducing apparatus which is Embodiment 1 of the present invention.

FIG. 2 is a block circuit diagram showing a reproduction system of the digital video signal recording / reproducing apparatus according to the first embodiment.

FIG. 3 is a conceptual diagram for explaining predictive coding according to the first embodiment.

FIG. 4 is a conceptual diagram for explaining predictive coding according to the first embodiment.

FIG. 5 is a conceptual diagram for explaining a macroblock according to the first embodiment.

FIG. 6 is a conceptual diagram for explaining a slice according to the first embodiment.

FIG. 7 is a flowchart showing a digital video encoding method according to the first embodiment.

FIG. 8 is a block circuit diagram showing another configuration example of the digital video signal recording / reproducing apparatus according to the first embodiment.

FIG. 9 is a block circuit diagram showing a recording system of a digital video signal recording / reproducing apparatus which is Embodiment 2 of the present invention.

FIG. 10 is a block circuit diagram showing an image determination circuit according to a second embodiment.

FIG. 11 is a conceptual diagram for explaining image determination according to the second embodiment.

FIG. 12 is a flowchart showing a digital video encoding method according to the second embodiment.

FIG. 13 is a flowchart showing image determination according to the second embodiment.

FIG. 14 is a conceptual diagram for explaining image determination according to the third embodiment.

FIG. 15 is a flowchart showing a digital video encoding method according to the third embodiment.

FIG. 16 is a block circuit diagram showing a configuration example of a recording system in a conventional video signal encoding system.

FIG. 17 is a block circuit diagram showing a configuration example of a reproduction system in a conventional video signal encoding system.

FIG. 18 is a conceptual diagram for explaining the operation of the memory circuit in the conventional video signal encoding system.

FIG. 19 is a conceptual diagram showing motion compensation prediction in a conventional video signal encoding system.

FIG. 20 is a conceptual diagram showing an operation of a memory circuit in a conventional video signal encoding system.

[Explanation of symbols]

1 video input circuit 2 memory circuit 3 subtractor 4
DCT circuit, 5 quantization circuit, 6 variable length coding circuit,
7 inverse quantization circuit, 8 IDCT circuit, 10 adder, 1
1 motion compensation prediction circuit, 12 buffer memory, 14
Storage medium, 15 code amount allocation circuit, 16
Code amount control circuit.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI technical display location H04N 9/808

Claims (9)

[Claims] 1. A digital video signal recording / reproducing apparatus which encodes an input digital video signal with high efficiency and records it on a storage medium having a constant information storage capacity at a variable rate, wherein the digital video signal has a fixed quantization level. The amount of information generated by encoding
a device for recording in units of up of Picture),
By using the generated information amount recorded in the GOP unit, the G is adjusted so that the digital video signal fits within the target information amount.
A digital video signal recording / reproducing apparatus comprising: a device for setting a target information amount in OP units; and a device for controlling a code amount according to the target information amount obtained in GOP units. 2. An apparatus for recording the information amount and time code generated by encoding the digital video signal at a fixed quantization level in GOP units, and the generated information amount and time code recorded in the GOP units. 2. The digital video signal recording / reproducing apparatus according to claim 1, wherein the digital video signal recording / reproducing apparatus comprises a device for setting a target information amount in GOP units so that the digital video signal can be contained in a target information amount. 3. A digital video signal encoding system for performing high-efficiency encoding of an input digital video signal at a variable rate within a predetermined average rate to control a code amount, wherein the digital image is fixed at a fixed quantization level. Means for encoding the signal and storing the generated information amount in GOP units, and in GOP units so that the digital video signal fits in the target information amount using the generated information amount recorded in GOP units. Means for setting the target information amount, and the G
A digital video signal recording / reproducing apparatus comprising a means for controlling a code amount according to a target information amount obtained in OP units. 4. In a digital video signal recording / reproducing apparatus which encodes an input digital video signal with high efficiency and records it on a storage medium having a constant information storage capacity at a variable rate, whether the GOP is prominent in image quality deterioration or not. An apparatus for performing determination, an apparatus for encoding the digital video signal with a fixed quantization level, and recording the amount of information generated and the determination result in GOP units, and determining the amount of generated information recorded in the GOP unit A device for setting the target information amount in GOP units so that the digital video signal fits in the target information amount using the result, and a device for controlling the code amount in accordance with the target information amount obtained in the GOP units. A digital video signal recording / reproducing apparatus characterized by the above. 5. An apparatus for detecting a GOP including a panning scene in which the entire background moves in a fixed direction and recording the GOP in GOP units, and setting the GOP including the panning scene so as to increase the target information amount. 5. The digital video signal recording / reproducing apparatus according to claim 4, wherein 6. A luminance level of the entire picture is detected, it is determined whether or not the GOP includes many pictures with a small luminance level, the GOP is recorded in GOP units, and the GOP including many pictures with a small luminance level is a target. 5. The digital video signal recording / reproducing apparatus according to claim 4, wherein the apparatus is configured to perform setting so as to reduce the amount of information. 7. A digital video signal encoding system for performing high-efficiency encoding of an input digital video signal at a variable rate within a predetermined average rate to control a code amount, wherein the digital video is provided at a fixed quantization level. The signal is encoded, the amount of information generated is stored in GOP units,
GO determines whether the GOP is prominent in image quality deterioration.
The target information amount is stored in P units, and the target information amount is set in GOP units so that the digital video signal fits in the target information amount by using the generated information amount recorded in the GOP units and the determination result recorded in the GOP units. A method for encoding a digital video signal, which is configured by setting and controlling a code amount according to a target information amount obtained in the unit of GOP. 8. A GOP including a panning scene in which the entire background moves in a fixed direction is detected and stored in GOP units, and a target information amount is allocated to the GOP including the panning scene more than a normal GOP. 8. The digital video signal encoding method according to claim 7, wherein the method is configured by setting. 9. A luminance level of the entire picture is detected, it is determined whether or not the GOP includes many pictures with a small luminance level, and the GOP is stored in GOP units. 8. The digital video signal encoding method according to claim 7, wherein the setting is made so that the target information amount is allocated to be smaller than that of the GOP.