JPH1051784A - Dynamic image coder - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize a stable compression coding with high image quality. SOLUTION: A scene change detection circuit 102 detects a frame having a scene change. An image replacement discrimination circuit 104 discriminates whether or not a frame causing a scene change is a B-picture (two inter-frame coding frames to conduct compression by referencing two past and future frames) and in the case of the B-picture, the circuit 104 gives a signal to a changeover circuit 106. Upon the receipt of the signal, the changeover circuit 106 makes switching to provide an output of an image replacement code in place of the code of the B-picture. The recording of the Bpicture requiring many codes is inhibited and codes to display an image of other frame in place of the image are recorded, then a stable compression coding is realized.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture coding apparatus, and more particularly to a moving picture coding apparatus such as a television signal which can be variably encoded using motion prediction combining intra-frame coding and inter-frame coding. The present invention relates to a moving image encoding device that performs compression encoding using a long code. In particular, the present invention relates to a moving picture coding apparatus that stabilizes image quality by preventing adverse effects when a scene change occurs.

[0002]

2. Description of the Related Art As an encoding method for compressing and encoding moving images, the MPEG method is MPEG-Video (ISO / I / O).
EC 11172-2) and MPEG2-Video
(ISO / IEC 13818-2). This method is widely used in broadcasting and communication fields such as digital broadcasting and digital transmission in addition to recording.

[0003] In the MPEG system, an intra-frame coded frame called I-Picture, which performs coding within one frame image, and P-Picture, which performs compression by referring to past frames, and past and future 2 frames. Frame data is composed of two inter-coded frames called B-Pictures which perform compression by referring to one frame. The P-Picture is the closest reference frame existing in the past in the order of the reproduction time (the frame to be referred to, ie, I-Picture or P-Picture)
, The B-Picture is coded and decoded by referring to the reference frame present in the closest past and the reference frame present in the closest future.

[0004] In the MPEG system, a GOP (Group of Pictures), which is a set of a plurality of frames, is formed in a frame sequence, and one frame is subjected to intra-frame compression within this GOP. In this method, inter-frame compression is performed by referring to other image data, and compression encoding is performed.

In intra-frame compression, DCT (Discrete Cos
A method of performing variable length coding of data obtained by using ine transform and quantization is employed.

First, frame data is divided into small areas called macroblocks, and a DCT transform is performed for each 8 × 8 pixel block in this area to obtain a frequency domain.
Since the human visual characteristics are insensitive to high frequencies, by reducing the code allocation to the high frequency region in the obtained frequency region, frame data that does not visually notice a great deterioration of humans is obtained. be able to. This allocation is performed by quantization.

[0007] Quantization is realized by dividing 8x8 frequency domain data obtained as a result of performing 8x8 pixel DCT transform by a certain matrix.

On the decoder side, the matrix obtained by performing the original inverse DCT operation can be obtained by multiplying (inversely quantizing) the data obtained from the variable length decoder by the matrix.

Here, by setting a large division value for high-order data, that is, data for a high-frequency region,
It is possible to reduce the encoding accuracy in the high frequency region. This also makes it possible to reduce the data corresponding to the high frequency range.

[0010] The quantized data thus obtained is
The result is one with many zero values. For this reason, a variable length encoding table is formed by a set of the number of zero values for this data and the actual value following the zero value. Variable-length coding is a type of entropy coding in which a short code is assigned to data with a high appearance frequency and a long code is assigned to data with a low appearance frequency. This combination results in efficient compression. In MPEG, this frame is I-Pict.
ure.

As a method of inter-frame compression, a technique called motion compensation using motion vector detection is used. As in the case of intra-frame compression, DCT, quantization, and variable-length coding are combined to perform coding. First, the frame data is divided into small areas called macroblocks as in the case of intra-frame compression. And in the inter-frame compression, not only the frame data,
Other frame data for reference is required. As this frame data, data of a frame existing at a position close in time is used. Then, for each macroblock, an area closest to the macroblock in terms of data is searched from the referenced frame data. The “region similar in data” refers to a region in which only the luminance of the normal pixel data is compared and the average of the square errors thereof is the smallest.

In the case of MPEG, this search is performed in half-pixel units. As a result, the area closest to the data with this search accuracy is searched from the area within the predetermined range of each macroblock. This is called motion prediction.

The amount of motion is called a motion vector, and this value is also subjected to variable length coding. Thus, on the decoding side, if there is frame data to be referenced, it is possible to cut out this “data-wise close” area.

The obtained frame data is, of course, not exactly the same as the data of the macroblock to be encoded. Therefore, it is necessary to compensate for this difference. In MPEG, in order to compensate for this difference, a difference is calculated for each pixel (luminance, color difference), and this difference area data is configured.

Then, a DCT operation is performed on the data for each area of 8 × 8 pixels, and quantization is performed in the same manner as in the case of intra-frame compression. However, since the quantization performed here is for the difference data, it is generally performed as if the accuracy is reduced as a whole.

For this reason, the difference value which does not usually distinguish between the high-frequency region and the low-frequency region is prepared as a default in MPEG. Of course, this quantization matrix can also be freely changed on the encoder side.

The data obtained in this manner is inversely quantized and inversely DCT-processed. If the frame data to be referred to exists in the decoder, the obtained value is used as the image area obtained from the motion amount. , The macroblock is decoded.

In MPEG, there are two types of image data compressed between frames. Frame data referred to only from a past frame is referred to as P-Pict.
ure. Frame data referred to from both the past and the future is called a B-Picture.

The P-Picture is predicted only from I or P-Picture located in the past in time.
The B-Picture is predicted only from the I or P-Picture located in the past and future in time.
Therefore, B-Picture is not used as a reference frame for motion estimation.

By combining these frame compression techniques, the obtained coded data is reduced to several hundredths to several hundredths compared with the original image data. Also, relatively high image quality is obtained.

However, since the method of referring to a frame is used, in order to realize the above method, a structure for holding the frame data of the reference side on the decoder side is required. In addition, on the decoder side, only the decoded frame data exists as the frame data to be referred to, so that the encoder must perform encoding under this condition. That is, during the encoding process, I-Picture and P-Pi
It is necessary to perform decoding under the same conditions as those of the decoder after encoding the cture, and hold the decoded frame data. Then, by referring to the frame data, P-Picture, BP
must be encoded.

FIG. 5 shows an example of a frame sequence of an MPEG moving image. In the figure, A indicates the display order of each frame, and indicates the order of the types of frames most commonly used in MPEG, but is not limited to this order. The arrow indicates that the frame data at the start point is referred to when encoding and decoding the frame at the end point.

Among them, 503 and 518 are I-Pict.
ure. 506, 509, 512 and 515 are P-Pictures. 501, 502, 504,
505, 507, 508, 510, 511, 513, 5
14, 516 and 517 are B-Pictures. For example, frames 501 and 502 are IP
The picture 503 is decoded, and if it is a decoder, it cannot be decoded unless it is held as a frame inside the decoder. In the case of the encoder, the encoding cannot be performed unless the encoding is completed and the I-Picture is decoded and stored as frame data inside the encoder.

Similarly, frames 504 and 505
Cannot be encoded or decoded unless the I-Picture 503 and the B-Picture 506 are decoded. In P-Picture, 5
If 503 is not decoded for 06 and 506 for 509, encoding and decoding are impossible.

For this reason, the recording order of these data is as shown in FIG. Decoding is performed in this recording order, and IP
By sequentially holding the decoded frame data of the picture and the P-picture, encoding and decoding of the respective frame data become possible. In addition, due to such a method, the I-Picture itself needs to maintain high image quality. Also, P-Picture is B-
In order to be used for Picture reference, high image quality must be maintained. For this reason, the normal I-Pictur
e is assigned the most codes, and then P-Pictur
By reducing the amount of data used in the order of e and B-Picture, high image quality as a whole is maintained.

FIG. 4 shows a specific example of a moving picture coding apparatus for compressing and coding a video signal of a moving picture according to the MPEG coding method.

The apparatus includes an input terminal 411, an image rearrangement circuit 412, a scan conversion circuit 413 for converting a raster scan into a macroblock, a subtraction circuit 414, a DCT circuit 416, a quantization circuit 418 for performing weighting, and a variable circuit. A long encoding circuit 420, a read buffer 422 from an encoder, and an output terminal 423 for outputting an encoded data sequence
, A rate (quantization width) control circuit 424, an inverse quantization circuit 426, an inverse DCT circuit 428, and an addition circuit 430.
, An image memory 432, data read buses 432 a and 432 b for the image memory, a motion compensation circuit 434, a mode determination circuit 436, and a motion detection circuit 438.

First, an image signal is input from the input terminal 411 to the image rearranging circuit 412. Image reordering circuit 4
Reference numeral 12 denotes a circuit for changing the frame order in the image recording order, that is, the image processing order, as shown in FIG. 5B. As a result, a frame to be compression-coded is input first as an I-Picture, and thereafter, the frames are rearranged in the order shown in FIG. 5B.

First, the image signal is input to the scan conversion circuit 413. The normal image signal is a digitized image data sequence, which is digital data obtained by raster scanning. This is a format in which data is output one pixel at a time from the upper left of the screen, a line of one pixel is formed, and then data of a line corresponding to the next lower pixel is output. MP
In the EG, in order to process data in units of small areas called macroblocks, it is necessary to convert data into this block. In MPEG, this macro block is composed of 16 × 16 pixels of luminance data and 8 × 8 pixels of chrominance data.
It is composed of pixels. In MPEG, the ratio of the color difference is changed depending on the image format. However, a range corresponding to 16 × 16 pixels in luminance is processed as a macroblock, and a search for a motion vector is also performed using this as a unit.

The encoding procedure will be described below for each compression method. First, scanning of I-Picture data is usually performed because a motion vector is not searched (except in the case of MPEG, where a motion vector can be introduced into I-Picture for error recovery). The data is directly input to the DCT circuit 416 via the conversion circuit 412. After the space-frequency conversion of the image data is performed, the quantization circuit 418 weights each frequency region.

Then, this data is transferred to the variable length coding circuit 4.
20 and is subjected to variable length coding. Since these data are subjected to variable length coding, they cannot be always output at a constant transfer rate. Therefore, the read buffer 422 buffers the rate at the time of the transfer. As a result, the compressed code string is output from the output terminal 423 at a constant transfer rate.

The data amount output from the variable length coding circuit 420 is monitored by the rate control circuit 424. If the difference between the expected encoded data amount of each frame data and the observed data amount is large, the quantization value is changed.

In the quantization, a scalar value called a quantization width that can be determined by the encoder side can be determined in the matrix described above. Since this value is used by multiplying each value of the quantization matrix as it is, the larger the value, that is, the coarser the quantization, the smaller the amount of data generated, and the smaller the value, that is, the finer the quantization, the smaller the amount of data generated Will increase.

The output of the quantization circuit 418 is input to the variable length encoding circuit 420 and also to the inverse quantization circuit 426. This is for forming the above-mentioned code frame data used for reference, and thereafter the inverse DCT is performed by the inverse DCT circuit 428, and the result is sequentially input to the image memory 432. As a result, an I-Picture decoded image serving as a reference reference is stored in the image memory.

Next, P-Picture encoding will be described. At the time of P-Picture encoding, a motion vector is searched for each macroblock. This corresponds to the I-Picture stored in the image memory 432.
Alternatively, the search is performed by referring to P-Picture data. However, in P-Picture data, not all macroblocks are encoded using motion vectors. In moving image data, for example, in the case of a fast-moving image, and in the case of an image that pans largely to the left, the data of the macroblock on the right end is the data that first appeared in the frame data. For such data, even if a similar part is searched for from other frames, the compression effect is not so much obtained.

For this reason, in this case, the I-Pictur
As in the case of e, a mode for performing processing without using a motion vector is used. In addition, there is a mode in which block data at the same position is cut out as it is in order to easily perform background processing, and the mode determination circuit 436 determines which of these modes is effective for encoding the macroblock.

When the mode determination circuit 436 determines that motion compensation is to be performed, the command is input to the motion compensation circuit 434. The motion vector obtained by the motion detection circuit 438 is also input to the motion compensation circuit 434. The motion compensation circuit detects data of a corresponding area from these data in the image memory 432 and outputs the data to the subtraction circuit 414. The difference is DCT, quantized, and then coded as in the case of I-Picture.

However, the P-Picture is used for reference from a P-Picture located in the future in time or a B-Picture located in the past and the future. For this reason, this data is I-Pictu
It is necessary to store re in the image memory 432 in the same manner as storing re. In this case, the frame is decoded.
P-Picture cannot perform decoding on the data itself. Therefore, in this case, the motion vector used for encoding the frame is used again, and the motion compensation circuit 434 adds the pixel data of the corresponding area to the addition circuit 430 to perform decoding. Then, the decoded P-Picture data is stored in the image memory 432.

Next, encoding of B-Picture will be described. In the case of B-Picture, the encoding method is basically the same as that of P-Picture. However, in the case of B-Picture, I-Picture or B-Picture located in the past and in the future
, It is necessary to search for a motion vector twice as large as the P-Picture.

Then, encoding is performed, and B-Pict
Since ure is not used for reference from any frame,
Only the encoded data is output as it is, and the decoded image need not be generated in the encoder.

In the case of B-Picture, the coding mode of a macroblock is predicted from the past and the future,
The prediction is selected from four types: prediction only from the past, prediction only from the future, and intra-unit coding.

In this way, the MPEG frame data is generated. However, the data amount allocation differs for each coding type due to the difference in the nature of each coding type. I-Pictur in descending order
e, P-Picture and B-Picture.

[0043]

However, in the prior art, the B-Pic with the smallest data amount allocation is used.
When a scene change occurs in the “ture”, it becomes impossible to refer to past and future bidirectional frames.
This makes it impossible to maintain a high compression rate in B-Picture, and the data amount temporarily increases. The rate control circuit 424 reduces the amount of data allocated to subsequently encoded frames to absorb the temporarily increased amount of data.

For this purpose, I-Picture or P-Pictu, which is a reference frame to be subsequently encoded, is used.
The assignment of the data amount to re also decreases, and the reproduction image quality of the reference frame decreases. Therefore, there is a problem that the image quality of a reproduced image is temporarily reduced during decoding.

The present invention has been made to solve such a problem, and an object of the present invention is to provide a moving picture coding apparatus capable of providing a stable reproduced picture.

[0046]

According to one aspect of the present invention, a moving picture encoding apparatus receives a moving picture signal composed of a plurality of frames, A coding circuit for performing inter-frame coding including a frame referred to by another frame and a frame not referred to by another frame, and a frame in which a scene change is performed from a moving image signal including a plurality of frames. An identification circuit for identifying, a determination circuit for determining characteristics of the identified frame, and an image for displaying another frame image instead of an image of a frame where a scene change is performed, based on a determination result of the determination circuit. A first output circuit for outputting information.

Further, based on the judgment result of the judgment circuit,
A second output circuit that outputs the encoded moving image signal;

Further, the plurality of frames include a referenced frame and a non-referenced frame, the determination circuit determines that the identified frame is an unreferenced frame, and the first output circuit determines the identified frame. When the frame is not referenced, a code for displaying an image of another referenced frame is output instead of the image of the frame where the scene change is performed.

Further, the plurality of frames include a first frame to be referenced and a plurality of second frames not to be referenced, each of the second frames referring to the first frame, and When the identified frame is one of the second frames, the output circuit outputs a code for displaying the image of the first frame instead of the image of each of the second frames.

According to another aspect of the present invention, a moving picture coding apparatus comprises: a coding circuit for compressing and coding a moving picture signal composed of a plurality of frames for each frame; And an output circuit for outputting information for displaying an image of another frame instead of an image of a specific frame, based on the state of compression encoding.

According to still another aspect of the present invention, a moving picture coding apparatus includes: a coding circuit for compression coding a moving picture signal composed of a plurality of frames for each frame;
A measuring circuit for measuring the data amount of the encoded frame, and an output circuit for outputting information for displaying an image of another frame instead of the image of the encoded frame based on the measured data amount And

Further, the moving picture coding apparatus performs coding according to the MPEG system. According to these inventions, it is possible to reduce the adverse effect of image coding when a scene change occurs in a moving image signal. This makes it possible to provide a video encoding device that can provide a stable decoded image.

[0053]

FIG. 1 is a block diagram showing a configuration of a moving picture coding apparatus according to a first embodiment of the present invention.

The blocks having the same functions as those of the conventional device in FIG. 4 are denoted by the same reference numerals, and description thereof will not be repeated.

The apparatus shown in FIG. 1 includes a scene change detection circuit 102, an image replacement determination circuit 104, a code switching circuit 106, and an image replacement code generation circuit 108 in addition to the configuration shown in FIG.

The scene change detection circuit 102 is a circuit that receives a moving image signal input from the input terminal 411, determines a scene change, and outputs a scene change detection signal. A scene change can be detected by, for example, a change in a high-frequency component of a moving image, a change in an edge component of the image, a change in a low-frequency component, and the like. Further, as a frame serving as a reference for the current input frame of the scene change detection circuit, a reference frame referred to by the current input frame in the encoding device can be used in addition to the frame located immediately before on the time axis. Although the detection method and the target frame in the scene change detection have been described above by way of example, the detection method and the target frame are not limited thereto.

The image replacement determination circuit 104 receives the scene change detection signal of the scene change detection circuit 102 as an input, and encodes each frame, that is, whether the image is I-Picture, P-Picture or BP.
It is determined which of the above is the case. Thereby, the encoding type of the frame in which the scene change is detected is determined. When a scene change is detected in a B-Picture that is a non-reference frame (a frame that is not referenced from another frame), the image replacement determination circuit 104 outputs the B-Picture output from the variable-length encoding circuit 420 to the B-Picture. The corresponding code is
I-Picture or P-Pi, which is the closest past reference frame referenced by Picture
or the nearest future I-Pic
A signal for replacing with an image replacement code having a small amount of data, which instructs to output the same image as that of the Ture or P-Picture as a decoded image at the time of decoding, is output to the code switching means 106.

When replacing a non-reference frame, the image replacement code generation circuit 108 performs B-Pic coding at the time of coding.
B-Picture that instructs to output the I-Picture or P-Picture, which is the reference frame closest to the past in the past, or the I-Picture or P-Picture located in the closest future, as a decoded image. An image replacement code which is a code having a data amount smaller than the data amount of the code to be generated is generated.

In accordance with an image replacement instruction from the image replacement determination circuit 104, the code switching circuit 106 replaces the code corresponding to the scene change detected B-Picture output from the variable length coding circuit 420 with the image replacement code. The image is replaced with an image replacement code output from the generation circuit 108.

For example, in the MPEG2 system, codes for performing image replacement are not defined in the standard. However, the macros of the first macroblock (the block located at the top left of the image) and the last macroblock (the block at the bottom right of the image) in the scanning order in macroblock units located at the top left of each image If only the block header is encoded, a code with a small amount of data is
-Any I-Picture or P-Picture that is referred to when decoding the Picture is
It can be instructed to decode as Picture.

First, an operation in a normal state where a scene change does not occur will be described. The moving image signal input from the input terminal 411 is input to the scene change detection circuit 102. The scene change detection circuit 102 does not detect a scene change. Therefore, the image replacement determination circuit 104 instructs to select a code output from the variable length coding circuit 420. Code switching circuit 106
Outputs the code output from the variable length encoding circuit 420 to the encoder read buffer 422 at the subsequent stage.

Next, the operation when a scene change occurs in a non-reference frame will be described.

It is assumed that a scene change has occurred in the B-Picture 508 in FIG. The moving image signal input from the input terminal 411 is applied to the scene change detection circuit 10.
2 is input. In the scene change detection circuit 102,
The B-Picture 508 detects a scene change and outputs a scene change detection signal. In the image replacement determination circuit 104, the scene change detection signal is BP
After confirming that the error occurred in the image 508, the code switching circuit 10 outputs the image replacement code.
Instruct 6 The image replacement code generation circuit 108 converts a P-Picture 509 contributing to the encoding of the B-Picture 508 into a B-Pic as an image to be encoded.
An image replacement code for instructing output as a decoded image instead of the cure 508 is generated.

The code switching circuit 106 receives the instruction to output the image replacement code from the image replacement determination circuit 104 and receives the output of the variable length coding circuit 420 as a B-Picture.
All codes obtained by encoding 508 are removed, and the image replacement code from the image replacement code generation circuit 108 is output to the encoder read buffer 422 at the subsequent stage.

Next, the operation when a scene change occurs in the reference frame will be described.

Assume that a scene change has occurred in the P-Picture 512 of FIG.

The moving image signal input from the input terminal 411 is input to the scene change detection circuit 102. In the scene change detection circuit 102, P-Picture5
At 12, a scene change is detected and a scene change detection signal is output. In the image replacement determination circuit 104,
After confirming that a scene change has occurred in the P-Picture 512, an instruction is issued to select a code obtained by encoding the P-Picture 512 output from the variable-length coding circuit 420, as in the normal operation.

The code switching circuit 106 outputs a code obtained by coding the P-Picture 512 output from the variable length coding circuit 420 to the subsequent encoder read buffer 422.

In this manner, a code for a B-Picture having a very large data amount, which is generated when a scene change occurs in the B-Picture, can be replaced with an image replacement code having a small data amount. As a result, it is possible to suppress the occurrence of a code having a data amount larger than expected in the B-Picture, and to increase the coding efficiency when a scene change occurs. As a result, stable compression coding with high image quality can be realized as a result.

Here, the image replacement determination circuit 104
Is a BP which receives a scene change detection result of the scene change detection circuit 102 as an input and detects a scene change.
Icture image replacement was instructed.
(1) All B-Pictures that refer to the same reference frame as the B-Picture where the scene change has occurred, and (2) All B-Pictures before the B-Picture where the scene change has occurred.
All of the B-Pictures after the B-Picture in which the scene change has occurred or (3) the scene change has occurred
It is also conceivable to instruct re to replace the frame.

The P-Pi located in the future of the code inputted in the image replacement code generation circuit 108 is located in the future.
Although the description has been made such that the image is replaced with a decoded image at the time of decoding in place of the frame instructed to replace the image, the reference frame positioned in the future in time is I-Picture.
If so, the I-Picture is used as a decoded frame.

Further, in the image replacement code generation circuit 108, the P-Picture or I-Picture located in the future in time is decoded as a decoded frame in place of the image for which image replacement is instructed at the time of decoding. To the past P-Picture or I-Picture
re, P-Picture located closest in time
Alternatively, when the I-Picture is decoded, the I-Picture may be a decoded frame in place of the frame instructed to replace the image.

FIG. 2 is a block diagram showing a configuration of a moving picture coding apparatus according to the second embodiment of the present invention. Blocks having the same functions as those of the circuit shown in FIG. 4 or FIG. 1 are denoted by the same reference numerals, and description thereof will not be repeated.

In this circuit, in addition to the circuit of FIG. 4, an image replacement determination circuit 202 and an image replacement code generation circuit 10
8 and a buffer memory 204.

The image replacement determination circuit 202 receives the macroblock type output from the mode determination circuit 436 and the motion vector output from the motion detection circuit 438 as inputs, and calculates the number of intra-frame coded macroblocks in the B-Picture, Image replacement is determined based on a parameter indicating the degree of motion, such as the number of blocks predicted from, and the number of motion vectors exceeding a predetermined threshold.

The buffer memory 204 stores a period until the result of the determination is output from the image replacement determining circuit 202,
This is a memory for delaying the code output from the variable length encoding circuit 420.

In the present embodiment, a series of determination operations performed by scene change detection circuit 102 and image replacement determination circuit 104 in the embodiment of FIG. 1 are performed by new image replacement determination circuit 202 having different inputs. The same effect can be obtained.

FIG. 3 is a block diagram showing a configuration of a moving picture coding apparatus according to the third embodiment of the present invention. Blocks having the same functions as the circuits shown in FIGS. 1 and 4 are denoted by the same reference numerals, and description thereof will not be repeated.

This circuit has a data amount counting circuit 302 and an image replacement determining circuit 3 in addition to the circuit shown in FIG.
04, the rate control circuit 306, and the buffer memory 30
8 and an image replacement code generation circuit 108.

The data amount counting circuit 302 is a counter that receives a code output from the variable length coding circuit 420 and is reset every frame. The data amount counting circuit 302 outputs the data amount from the beginning of each frame.

The image replacement determination circuit 304 receives the data amount from the beginning of the frame, which is the output of the data amount counting circuit 302, and, when the data amount exceeds a predetermined threshold value in the B-Picture, Is the reference frame located closest to the B-Picture in the past, which is I-Picture or P-Picture.
re, or I-Pictur located in the nearest future
In order to make the same frame (frame replacement) as e or P-Picture, it is determined whether or not a code corresponding to the B-Picture output from the buffer memory 308 is replaced with an image replacement code having a small data amount. . Then, it instructs the code switching circuit 106 to perform image replacement, and also notifies the rate control circuit 306 that image replacement has been performed.

The rate control circuit 306 receives the image replacement execution signal from the image replacement determination circuit 304 in addition to the processing function of the conventional rate control circuit 424 and
-Correct the data amount of the code generated in Picture downward to the data amount of the image replacement code. This allows
The rate control circuit has a function of correcting the distribution of the data amount when encoding the subsequent frame.

The buffer memory 308 is a code delay buffer memory capable of storing data equal to or larger than the threshold value of the data amount determined to be replaced by at least the image replacement determination circuit 304, and includes a variable length coding circuit. The code output from 420 is input.

First, the operation in the case where the data amount in the B-Picture does not exceed a predetermined threshold will be described.

The output of the variable length coding circuit 420 is input to the data amount counting circuit 302. The data amount counting circuit is cleared at the beginning of the frame and outputs the data amount from the beginning of the frame. The image replacement determination circuit 304 receives the input.

At this time, since the data amount in the B-Picture does not exceed the predetermined threshold value, the image replacement determining circuit 304 delays the output of the variable length coding circuit 420, which is the output of the buffer memory 308. The user is instructed to select the code. The code switching circuit 106 outputs the code output from the variable length coding circuit 420 to the encoder read buffer 422 at the subsequent stage.

In addition, I-Picture, P-Pict
Also in ure, the same determination is made if the threshold is not exceeded, and the output of the buffer memory 308 is selected.

Next, the operation when the data amount in the B-Picture exceeds a predetermined threshold will be described.

The output of the variable length coding circuit 420 is input to the data amount counting circuit 302. The data amount counting circuit is
Cleared at the beginning of the frame and outputs the amount of data from the beginning of the frame. The image replacement determination circuit 304 receives the input.

The image replacement determining circuit 304 observes the data amount output from the data amount counting circuit 302, and switches the code to output an image replacement code when the data amount exceeds a predetermined threshold value. The circuit 106 is instructed. That is, even if a code corresponding to B-Picture is being output from the variable length encoding circuit 420, the code switching circuit outputs an image replacement code when the data amount exceeds a predetermined threshold. It instructs 106.

If the image replacement is determined at the same time when the threshold value is exceeded, the buffer memory 308 may be a memory capable of storing a data amount corresponding to a predetermined threshold value. The sign switching circuit 106
In response to an instruction to output an image replacement code from the image replacement determination circuit 304, all codes obtained by coding the B-Picture output from the variable length coding circuit 420 from the buffer memory 308 are removed, and an image replacement code is generated. Circuit 1
The image replacement code 08 is output to the encoder reading buffer 422 at the subsequent stage.

Further, the rate control circuit 306 replaces the total data amount of the code generated in the B-Picture with the total data amount of the image replacement code for the B-Picture, and corrects the data amount allocated to the subsequent frame.

Thus, in the present embodiment, the same effects as in the embodiment of FIG. 1 can be obtained.

In this embodiment, the image replacement determination circuit 304 performs replacement at the same time as the data amount exceeds a predetermined threshold value. However, the determination is made after the data amount for each frame is counted. You may do so.

As described above, when a scene change occurs in a non-reference frame, by suppressing the occurrence of an increase in the data amount, it is possible to increase the coding efficiency in the scene change. As a result, stable compression encoding with high image quality can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram of a video encoding device according to a first embodiment of the present invention.

FIG. 2 is a block diagram of a video encoding device according to a second embodiment of the present invention.

FIG. 3 is a block diagram of a video encoding device according to a third embodiment.

FIG. 4 is a block diagram illustrating a configuration of a moving image encoding device according to a conventional technique.

FIG. 5 is a diagram for describing a reference relationship of motion prediction coding including an intra-coded frame.

[Explanation of symbols]

 Reference Signs List 102 Scene change detection circuit 104, 202, 304 Image replacement determination circuit 106 Switching circuit 108 Image replacement code generation circuit 302 Data amount counting circuit 306 Rate (quantization width) control circuit

Claims (7)

[Claims] 1. A video signal composed of a plurality of frames is input, and at the time of encoding, inter-frame encoding including a frame referred to by another frame and a frame not referenced by another frame is performed. Encoding means, from a video signal composed of the plurality of frames,
Identification means for identifying a frame in which a scene change is performed; determination means for determining characteristics of the identified frame; and, based on a determination result of the determination means, in place of the image of the frame in which the scene change is performed, A moving image encoding apparatus, comprising: first output means for outputting information for displaying an image of another frame. 2. The video encoding apparatus according to claim 1, further comprising a second output unit that outputs the encoded video signal based on a result of the determination by the determination unit. 3. The plurality of frames include a referenced frame and a non-referenced frame, wherein the determination unit determines that the identified frame is a non-referenced frame, and wherein the first output The means outputs, when the identified frame is an unreferenced frame, a code for displaying an image of another referenced frame in place of the image of the frame where the scene change is performed. Item 3. The video encoding device according to item 1 or 2. 4. The method according to claim 1, wherein the plurality of frames are referred to as a first frame.
And a plurality of second frames that are not referred to. Each of the second frames refers to the first frame, and the first output unit outputs the identified frame, The code according to claim 1, wherein a code for displaying an image of the first frame is output instead of an image of each of the second frames when the frame is one of the second frames. Video encoding device. 5. An encoding unit for compressing and encoding a moving image signal composed of a plurality of frames for each frame, and said specific image is compressed based on a state of compression encoding of a specific frame by said encoding unit. An output unit that outputs information for displaying an image of another frame instead of an image of a frame. 6. An encoding unit for compressing and encoding a moving image signal composed of a plurality of frames for each frame, a measuring unit for measuring a data amount of the encoded frame, and the measured data. Output means for outputting information for displaying an image of another frame instead of the image of the encoded frame, based on the amount. 7. The moving picture coding apparatus according to claim 1, wherein said moving picture coding apparatus performs coding by an MPEG system.