JPH04316294A - Variable-length coding circuit - Google Patents

Abstract

PURPOSE:To prevent the worsening of a convergence condition, to make fast the recovery from the condition and to suppress the picture quality deterioration by controlling the prepared code quantity in accordance with the convergence condition of a net to constitute a variable-length coding circuit when the multi-value time series signal to be inputted is converted to a variable-length code. CONSTITUTION:The video signal to be inputted is divided into picture element blocks, the two-dimensional discrete cosine conversion is performed and after the coefficient is quantized by a quantizing part 12, it is stored in a storing part 13 for 1 frame. By an animation and transition picture deciding part, the difference of the coefficient value is calculated with the data of the present frame and the data stored in the storing part 13, and this and the threshold set beforehand are compared. Thus, it is decided that when the difference value is large, an animation block is obtained, and when it is small, a still picture block is obtained, the result is transmitted to a preparing code quantity control part 16 and the number of each block is counted. Subsequently, when the net is not the convergence condition, the instruction is performed so as to perform the usual coding from the control part 16 to a variable- length coding part 15, and at the time of the convergence condition, the preparing code quantity is suppressed.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a variable length code circuit for variable length coding an input multilevel time series signal.

0002

[Background Art] Conventionally, when encoding a digital video signal, it is divided into a fixed number of pixel blocks, and by performing orthogonal transformation on each block, redundancy between adjacent pixels is removed, and transform coefficients are varied. Highly efficient compression encoding is achieved by long encoding. In addition, at this time, variable rate encoding is performed to change the transmission rate over time, taking advantage of the fact that the amount of information in the video signal temporarily increases.
Asynchronous transfer mode (ATM, Asynchronous transfer mode)
A technique is also being considered in which the data is divided into fixed-length cells and transmitted using the same transfer mode. FIG. 2 shows an example of encoding a video signal using this variable length code circuit.

In this figure, 21 is a discrete cosine transform (DCT) applied to a video signal.
22 is a quantization unit that quantizes the transform coefficients, 23 is a storage unit that stores one frame worth of transform coefficient data, and 24 is a A video/transition image determination unit that compares transform coefficients of blocks and determines whether the block is a video block or a transition image block;
Reference numeral 25 denotes a variable length encoding unit that variable length encodes the transform coefficients.

[0004] Next, its operation will be explained. Signal line 201
The video signal inputted through is divided into 8×8 pixel blocks, and the orthogonal transform unit 21 subjects each block to discrete cosine transform. The conversion coefficient is the signal line 202
The signal is transmitted to the quantization unit 22 through the quantization unit 22 and quantized. The quantized transform coefficients are sent to the storage unit 23 through the signal line 203.
is stored in

[0005] In the moving image/transition image determining section 24, the signal line 20
The difference value is calculated between the conversion coefficient of the current frame input through 3 and the conversion coefficient of the previous frame read from the storage unit 23, and if it is larger than a preset threshold, it is used as a moving image block, and if it is smaller than a preset threshold, it is used as a still image block. It is determined to be an image block. Further, among still image blocks, those whose previous frame was determined to be a moving image block are distinguished from still image blocks as transition image blocks. The transformation coefficient data stored in the storage unit 23 is used as the transformation coefficient data of the previous frame when determining whether the next frame is a moving image or a still image. The moving image/transition image determining unit 24 instructs the variable length encoding unit 25 to perform encoding via the signal line 204 only when it is determined that the block is a moving image block or a transition image block. That is, information on still image blocks other than transition image blocks is not transmitted. Still image blocks other than transition image blocks are reproduced by replacing them with pixels of the previous frame on the receiving side. The variable length encoding unit 25 follows instructions from the moving image/transition image determining unit 24.
Variable length encode the transform coefficients. The generated variable length code is
It is divided into cells of fixed length, a header indicating the destination etc. is added, and the data is transmitted via ATM.

[0006]

By the way, when variable rate encoding as shown in the conventional example is performed, the amount of code to be transmitted varies over time. Although this method enables efficient encoding, it is difficult to control the amount of generated code, and it is also difficult to control when the amount of generated code increases temporarily due to scene changes and the network becomes congested. Since the amount of code cannot be controlled using conventional methods, it is difficult to recover from congestion. Therefore, there is a problem in that the image quality deteriorates significantly due to cell discard.

[0007]

[Means for Solving the Problems] The present invention has been made to solve the above problems, and in the invention according to claim 1, an input signal is generated by dividing it into a fixed number of blocks. The variable-length code circuit that sequentially variable-length encodes the generated information has a generated code amount control unit that controls the amount of variable-length codes to be transmitted, and when the network is not congested, the generated variable-length code is used as is. When the network is congested during transmission, the total amount of codes generated per block is compared with a preset threshold for the amount of codes generated, and when the amount of codes generated per block is smaller than the threshold for the amount of codes generated. transmits the generated variable length code as is, and when the amount of generated codes per block exceeds the threshold of the amount of generated codes, outputs a code indicating that the encoding of the block has been completed, and the subsequent It is characterized by not transmitting any code.

[0008] In the invention according to claim 2, the input signal is encoded using a quantization unit having a plurality of types of quantization levels and a different encoding table depending on the quantization level of the quantization unit. When the network is not congested, linear quantization is performed and quantization is performed at a fine quantization level for encoding, and when the network is congested, non-linear quantization is performed. It is characterized by encoding by coarsening the quantization level.

In the invention according to claim 3, in the variable length code circuit which divides an input video signal into a fixed number of blocks and performs variable length coding for each block, the block of the previous frame and the block of the current frame are When the difference value is larger than a preset threshold value, it is determined that it is a video block, and when it is smaller, it is determined that it is a still image block, and among the still image blocks, the block whose previous frame is a video block is determined. a moving image/transition image determining unit that determines a “transition image block”; and a variable length encoding unit that variable length encodes only the signals of the moving image block and the transition image block; The method is characterized in that the threshold value is increased.

0010

[Operation] With the above configuration, in the invention according to claim 1, when the network becomes congested, it is confirmed by comparing the generated code amount with a threshold value, and then the signal transmission of a part of the block is not performed. That's it. In the invention according to claim 2, when the network becomes congested, nonlinear quantization is performed to coarsen the quantization level for encoding.

[0011] In the invention according to claim 3, the threshold value of the video block is changed according to the congestion state of the network, and then,
Only video blocks and transition image blocks are transmitted.

0012

DESCRIPTION OF THE PREFERRED EMBODIMENTS The invention according to claim 1 will be explained below based on the embodiments shown in FIGS. 1 and 3. FIG. 1 is a block diagram of the system. 11 is an orthogonal transform unit that orthogonally transforms a video signal for each 8×8 pixel block, 12 is a quantization unit that quantizes transform coefficients, and 13 is the number of transforms for one frame. a storage unit storing 1
Reference numeral 4 denotes a video transition image determination unit that determines whether it is a video block or a transition image block; 16 a generated code amount control unit that controls the generated code amount of the coefficient block currently being encoded; and 15, a variable conversion coefficient. This is a variable length encoding unit that performs long encoding.

Next, the operation of each component shown in FIG. 1 will be explained. The video signal passed through the signal line 101 is 8
It is divided into ×8 pixel blocks and subjected to two-dimensional discrete cosine transformation. The coefficients are quantized by the quantization unit 12, and the coefficients of one frame of discrete cosine transform are stored in the storage unit 13 via the signal line 104. In addition,
The coefficient data of the discrete cosine transform stored at this time is
It is read out as the previous frame data when determining whether the next frame is a moving image or a still image.

The moving image/transition image determining unit 14 connects the signal line 103
The data of the current frame is inputted through the signal line 104, and the data of the previous frame is read out from the storage unit 13 through the signal line 104. Then, the difference between the coefficient values of the previous frame and the current frame is calculated and compared with a preset threshold value. When the difference value is larger than the threshold value, the block is determined to be a moving image block, and when it is smaller than the threshold value, the block is determined to be a still image block. Furthermore, among still image blocks, those whose previous frame was determined to be a moving image block are distinguished from still image blocks as transition image blocks.

The determination result of the moving image/transition image is transmitted to the generated code amount control section 16 via the signal line 105. A video block number counter (not shown) in the generated code amount control unit 16 counts the number of video blocks. Similarly, a transition image block number counter (not shown) counts the number of transition image blocks. These accumulated count values are 1
When encoding of a frame is completed, the count is reset and starts counting from 0 again from the next frame. When the counting of the number of video blocks and transition picture blocks in one frame is completed, the generated code amount control unit 16 controls one block of video blocks and transition picture blocks so that the number of generated codes in one frame is equal to or less than a preset value. Determine the threshold value of the amount of code per. On the other hand, the generated code amount control unit 16 is notified of the congestion state of the network through the signal line 106. If the network is not congested, the generated code amount control section 16 instructs the variable length encoding section 15 to perform normal variable length encoding through the signal line 107. Then, the variable length encoding unit 15 inputs the determination result of the moving image/transition image through the signal line 108, and performs variable length encoding on only the moving image block and the transition image block. For example, the encoding method is shown in Figure 3.
As shown in , zigzag scanning is performed from the DC component of the coefficient of the discrete cosine transform toward the high frequency component, and the "0" of the coefficient is
Huffman (
Huffman) encode. Conversely, when the network is congested, the generated code amount control section 16 inputs the code length of the variable length code generated through the signal line 109, and calculates the total amount of generated codes for that block. Then, the threshold value of the video block or transition image block is compared with the total amount of generated codes. When it is smaller than the threshold, encoding continues as is. On the other hand, when the total code amount becomes larger than the threshold (
3) instructs the variable length encoding unit 15 to end encoding through the signal line 107. The variable length encoding unit 15 generates codes (EOB, E
nd of Block code). and,
Encoding is not performed for the portion after that (the part indicated by the dotted line in FIG. 3). This makes it possible to control the amount of generated codes when the network becomes congested. Therefore, it is possible to suppress the deterioration of the congestion state and hasten recovery from the congestion state.

Next, the invention according to claim 2 will be explained based on the configuration diagram of the embodiment shown in FIG. The video signal input through the signal line 401 is subjected to discrete cosine transformation by the orthogonal transform unit 41 and quantized by the quantizer 42 . The quantization level control unit 43 is notified through the signal line 402 that the network is in a congested state. When the network is not congested, the quantizer 42 is instructed through the signal line 403 to perform linear, ie, uniform quantization. on the other hand,
When the network becomes congested, the quantization level control unit 43 instructs to perform nonlinear quantization. In accordance with this instruction, the quantization unit 42 changes the density function to linearly and finely quantize, or conversely to non-linearly and coarsely quantize. The variable-length encoding unit 45 performs variable-length encoding according to the quantization level transmitted through the signal line 404 using the encoding table corresponding to each case. In this way, the amount of generated code can be reduced.

Finally, the invention according to claim 3 will be explained based on the embodiment shown in FIG. 1, as in the case of the invention according to claim 1. When the network becomes congested, the generated code amount control section 16 instructs the moving image/transition image determination section 14 to increase the threshold value through the signal line 110. This reduces the number of pixel blocks that are determined to be moving pictures or transition picture blocks, thereby reducing the amount of generated code and, in turn, making it possible to control the amount of generated code. Note that the operation of the parts other than the generated code amount control unit 16 is described in claim 1.
This is the same as the embodiment related to.

[0018]

[Effects of the Invention] As explained above, according to the present invention,
In order to control the amount of generated code according to the congestion state of the network,
It is possible to prevent deterioration of the congestion state and hasten recovery from the congestion state, and in turn, it is possible to prevent and suppress image quality deterioration due to cell discard.

[Brief explanation of drawings]

FIG. 1 is a configuration diagram of an embodiment of the invention according to claims 1 and 3.

FIG. 2 is a configuration diagram of a conventional variable length code circuit.

FIG. 3 is an 8×8 discrete cosine transform coefficient block to be encoded.

FIG. 4 is a configuration diagram of an embodiment of the invention according to claim 2.

[Explanation of symbols]

11, 21, 41 Orthogonal transform unit (DCT) 12, 22
, 42 Quantization units 13, 23 Storage units 14, 24 Video/transition image determination units 15, 25, 45 Variable length encoding unit 16 Generated code amount control unit 43
Quantization level control units 101 to 110 Signal lines 201 to 204 Signal lines 401 to 404 Signal lines

Claims (3)

[Claims] 1. A generated code amount control unit that controls the amount of variable length codes to be transmitted in a variable length code circuit that sequentially variable length encodes information generated by dividing an input signal into a certain number of blocks. When the network is not congested, the generated variable length code is transmitted as is, and when the network is congested, the total amount of generated codes per block is compared with a preset threshold for the amount of generated codes. , when the generated code amount per block is smaller than the generated code amount threshold, the generated variable length code is transmitted as is, and when the generated code amount per block exceeds the generated code amount threshold, A variable length code circuit characterized in that it outputs a code indicating that encoding of the block has been completed, and does not transmit subsequent codes. 2. A quantization unit that has a plurality of types of quantization levels for an input signal, and a variable length encoding unit that performs encoding using a different encoding table depending on the quantization level of the quantization unit. When the net is not crowded,
A variable length code that performs linear quantization and performs quantization at a fine quantization level for encoding, and when the network becomes congested, performs nonlinear quantization to coarsen the quantization level and performs encoding. circuit. 3. In a variable-length code circuit that divides an input video signal into a fixed number of blocks and performs variable-length coding on each block, a difference value between a block of a previous frame and a block of a current frame is calculated. If the difference value is larger than a preset threshold, it is determined to be a video block, and if it is smaller, it is determined to be a still image block, and among the still image blocks, a block whose previous frame is a video block is determined to be a "transition image block". and a variable length encoding unit that performs variable length encoding only on the signals of the video block and the transition image block, and increases the threshold when the network becomes congested. A variable length code circuit featuring: