JPH06311497A - Video signal encoding device - Google Patents

Abstract

PURPOSE:To suppress the quantity of generated codes in a constant period of time to be below a certain quantity by predicting the code quantity at every quantization factor from the code quantity of an encode finished block, and selecting the quantization factor on the basis of this predicted value. CONSTITUTION:This video signal encoding device is constituted of a signal converting part 102 to execute the blocking and the orthogonal transformation of an inputted video signal, a quantizing part 103 in which the quantization factor can be selected, a variable length encoding part 104, a code quantity calculating part 106, a code quantity predicting part 108, and a quantization factor selecting part 109. The quantity of the generated codes is controlled by predicting the code quantity at every quantization factor before quantization and by selecting the appropriate quantization factor, and thus, any kind of a video can be encoded so as to be always below a certain code quantity.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video signal coding apparatus for performing high efficiency coding which is a means for reducing the code amount for the purpose of storing or transmitting a video signal.

[0002]

2. Description of the Related Art When digitally transmitting or recording a video signal,
Compressing the video signal by high-efficiency coding is effective in reducing the amount of transmission and storage. As a high-efficiency encoding method, a video signal is divided into blocks of about 8 × 8 to 16 × 16 pixels, and orthogonal transform such as discrete cosine transform (DCT transform) or wavelet transform is performed for each block to determine visual characteristics. In many cases, a method of performing variable length coding by performing quantization taking into consideration is used. Since this method includes variable-length coding, the generated code amount per unit time of the video signal is not necessarily constant. Therefore, various control methods have been proposed and put to practical use in order to make the output code amount constant in a certain fixed period.

This control method is roughly classified into a feedback control method and a feedforward control method. As an example of the former, in a video conference system or the like, a buffer storage device is provided in the latter stage of variable-length coding to adjust the communication speed, and the storage capacity of the buffer storage device allows dynamic change of the quantization characteristics, Methods such as pixel thinning out of video signals and frame thinning are used (reference: Takahiko Fukibuki, "TV
Multidimensional signal processing of images ”, Nihon Kogyo Shimbun).

The latter is a method of repeating the operation of actually performing the encoding for a fixed period and changing the quantization characteristic and performing the encoding again when the predetermined code amount is not obtained, or the output code amount for each quantization characteristic in advance. Is used to select an appropriate quantization characteristic.

FIG. 2 is a block diagram of an example of a video signal coding apparatus for controlling the code amount by the conventional feedback control method.

In FIG. 2, reference numeral 201 is an input terminal for inputting a video signal, 210 is a blocking unit for collecting the video signal in a predetermined rectangular unit, and 211 is a signal for orthogonally transforming the video signal in block units to obtain transform coefficients. Converter, 20
3 is a quantizer for quantizing the transform coefficient with the selected quantizer, 204 is a variable length encoder for variable length coding the quantized transform coefficient, and 220 is a temporary storage for this variable length code. A buffer storage unit, 205 is an output terminal for outputting the code accumulated in the buffer storage unit 220, and 208 is the buffer storage unit 2.
This is a quantization factor selection unit that detects the code amount accumulated in 20 and selects the quantization factor used in the quantization unit 203 according to the code amount.

This operation will be described below. The digitized video signal is input from the input terminal 201, is combined into a block unit of 8 × 8 pixels in the blocking unit 210, and is subjected to DCT conversion in the signal conversion unit 211 to obtain a conversion coefficient.
The quantization unit 203 can change the quantization characteristic, and the quantization characteristic is changed by the quantization factor designated by the quantization factor selection unit 208. The variable length coding unit 204 converts the quantized transform coefficient into a variable length code and stores it in the buffer storage unit 220. Here, since the code length differs depending on the code, the code amount per constant time is generally not constant.

On the other hand, the code is taken out from the buffer memory 220 at a constant speed and is output from the output terminal 205 as the output of the video signal coding apparatus. The quantization factors have a magnitude relation, and the same transform coefficient is quantized by the quantization unit 203 with a larger quantization factor, and the variable length coding unit 20.
When the variable length coding is performed with 4, the code amount is reduced, and when the variable length coding is performed with a smaller quantization factor, the code amount is increased, and the quantization unit 203 and the variable length coding unit 204 are configured. And

Input and output to the buffer storage unit 220 are performed asynchronously, but the quantization factor selection unit 208 detects the code amount stored in the buffer storage unit 220, and the quantization unit 203 according to the code amount. Select the quantization factor to use. In fact, when the code amount accumulated in the buffer storage unit 220 is large, a large quantization factor, that is, a quantization factor in which the output code amount of the variable length encoding unit 204 becomes small, is selected, and conversely stored in the buffer storage unit 220. If the code amount is small, a small quantization factor, that is, a quantization factor that the output code amount of the variable length coding unit 204 increases will be selected.

FIG. 3 is a block diagram of an example of a video signal coding apparatus for controlling the code amount by the conventional feedforward control method. In FIG. 3, 301 is an input terminal for inputting a video signal, 310 is a blocking unit that collects the video signal in a predetermined rectangular unit, 311 is a signal conversion unit that performs orthogonal transformation on the video signal in block units, and obtains transform coefficients, 32
Reference numeral 1 is a transform coefficient storage unit that temporarily stores transform coefficients corresponding to the determined number of block units, 303 is a quantization unit that quantizes the transform coefficient with a selected quantization factor, and 304 is a quantized transform coefficient. A variable length coding unit for variable length coding, 305
Is an output terminal for outputting a sequence of variable-length codes, 306 is a code amount calculation unit for calculating the code amount in the case where variable coefficients are quantized by transform coefficients, and 308 is used by the quantization unit 303 according to the calculated code amount. This is a quantization factor selection unit that selects a quantization factor to be used.

This operation will be described below. The digitized video signal is input from the input terminal 301, and the blocking unit 310 collects the video signal into blocks of 8 × 8 pixels,
The signal conversion unit 311 performs DCT conversion to obtain a conversion coefficient. This conversion coefficient is stored in the conversion coefficient storage unit 321 for a certain number of block units. Hereinafter, this conversion coefficient storage unit 3
It is assumed that the fixed number of block units that can be stored in 21 is n, and the number of block units when one video signal frame is divided into blocks is an integer m times n.

The quantizing unit 303 can change the quantizing characteristic, and the quantizing characteristic is changed by the quantizing factor designated by the quantizing factor selecting unit 308. Variable length coding unit 304
Converts the quantized transform coefficient into a variable length code and outputs it from the output terminal 305. The code amount calculation unit 313 quantizes the transform coefficient output from the signal conversion unit 311 by the quantization unit 303 for each quantization factor selectable by the quantization unit 305, and performs variable length coding by the variable length coding unit 304. In this case, the code length is calculated.

After the conversion coefficients for n blocks are stored in the conversion coefficient storage unit 321, all the conversion coefficients stored therein are quantized, and the total code amount for variable-length coding is assigned to one frame. The quantization factor of each block is selected to be 1 / m or less, and the transform coefficients stored in the transform coefficient storage unit 321 are sequentially quantized and variable length coding is performed. By repeating this operation m times per frame, the code amount per frame is controlled.

[0014]

However, in the feedback code amount control, in the above-mentioned conventional example, the code amount detection for the coding parameter control using the quantization factor is delayed and the control becomes unsatisfactory. In addition, the correlation between the quantization factor and the code amount after variable-length coding differs depending on the image, and it is difficult to select an appropriate quantization factor from the code amount accumulated in the buffer storage unit. This cannot be done, and the code amount for a certain period overflows or excessive underflow occurs. Furthermore, the buffer storage section overflows,
You should also consider the case of underflow.

In the feedforward code amount control, it is possible to avoid code amount overflow and excessive underflow, but if the integer n is made small, that is, if the capacity of the transform coefficient storage unit is small, the information contained in the image is reduced. Due to the deviation of the amount within the frame, the variation of the quantization factor for each block unit becomes large, and the image quality of the reproduced video becomes poor visually. On the other hand, if the integer n is set large, the scale of the conversion coefficient storage unit becomes large. In particular, the conversion coefficient storage unit needs about 10 bits for the conversion coefficient when using the DCT conversion of 8 × 8 pixels for the resolution accuracy of one pixel of 8 bits, and the conversion coefficient storage unit stores in the frame memory for storing the video signal before encoding. It's also big compared. In addition, the quantization factor selection unit also requires a code amount memory proportional to n, which increases the circuit scale of the device.

In view of the above points, the present invention has an object to provide a video signal encoding apparatus which effectively controls the code amount, prevents the deterioration of the image quality in appearance, and generates an appropriate amount of codes.

[0017]

In order to achieve the above object, the present invention relates to a signal conversion unit that collects video signals in a predetermined fine unit, performs signal conversion to obtain a conversion coefficient, and selects a conversion coefficient in the fine unit. A quantizer that quantizes with a possible quantization factor, a variable-length encoder that variable-length-encodes the quantized signal, and a conversion coefficient of the fine unit for each of the selectable quantization factors. A code amount calculation unit that calculates a code amount when quantization and the variable length coding are performed, and the code amount calculation is performed for each of the selectable quantization factors of the video signal in the fine unit. A code amount prediction unit that predicts from the code amount calculated by the unit, and a quantization factor selection unit that selects the selectable quantization factor based on the predicted code amount, and the variable length coded signal A configuration for outputting a column and the selected quantization factor .

Further, in the above configuration, in the quantization of the transform coefficients of a plurality of predetermined fine units in the quantizing unit, the quantizing factor selecting unit selects a quantizing factor, and other than the quantizing. This is a configuration in which the same quantization factor as the immediately preceding one is used for quantization.

Further, in the above structure, a code amount storage unit for storing the code amount calculated by the code amount calculation unit for a predetermined period is provided, and the code amount prediction unit is capable of selecting the video signal in the fine unit. A configuration for predicting the code amount for each quantization factor from the code amount stored in the code amount storage unit near the image represented by the video signal of the fine unit and the code amount calculated by the code amount calculation unit. Is.

[0020]

According to the present invention, with the above-described configuration, the code length in the case of performing variable length coding by quantizing for each quantization factor is predicted, and by selecting an appropriate quantization factor, the code amount for a predetermined fixed period can be obtained. It is possible to control the amount to be equal to or less than a separately determined fixed amount.

[0021]

EXAMPLE A first example of the present invention will be described. Figure 1
FIG. 1 is a block diagram showing a configuration of a first embodiment of a video signal encoding device. This will be explained below.

In FIG. 1, 101 is a video input terminal for inputting a video signal for each frame, 102 is a signal conversion unit that collects video signals in block units and performs orthogonal transformation to obtain transformation coefficients, and 103 selects orthogonal transformation coefficients. A quantizer that quantizes with a quantization factor, and 104 is a variable-length encoder.
5 is a terminal for outputting a video signal encoded by this apparatus, 10
6 is a code amount calculation unit that obtains a code amount when variable length coding is performed by quantizing for each quantization factor, 108 is a code amount prediction unit that predicts the code amount, and 109 is the present encoding apparatus per frame of a video signal. Based on the target value of the code amount to be output, the amount of the code output from the output terminal 105 of the frame being encoded, and the code amount predicted by the code amount predicting unit 108, the quantization factor used in the quantizing unit is calculated. It is a quantization factor selection unit to be selected.

In order to simplify the description of the parts not directly related to the present invention, the video signal coding apparatus of the present embodiment will be described below.
It is assumed that the input video signal is only a luminance signal and is composed of 640 pixels in the horizontal direction and 480 pixels in the vertical direction per frame.

The signal conversion unit 102 comprises a blocking unit 110 and an orthogonal conversion unit 111. The blocking unit 110 collects the input video signal into blocks of 8 pixels in the horizontal direction and 8 pixels in the vertical direction for each frame, and the video signal of one frame is 80 blocks in the horizontal direction and 60 blocks in the vertical direction, for a total of 48 blocks.
Block into 00 blocks.

The blocks in the frame are as shown in FIG.
The block at the upper left corner is set to 1, and the blocks to the right are sequentially 2, 3 ,. ． 8
0 and the block immediately below the block 80 is 81, and 82, 83 ,. ． 160, the block immediately below the block 160 is numbered 161, and the block at the lower left end is numbered 4800 in the zigzag order, and the first block, the second block ,. ． , Block 4800.

The blocking unit 110 outputs the blocks of the video signal in the order of the numbers. Next, the orthogonal transform unit 111 performs a discrete cosine transform (DCT transform) for each block and outputs a transform coefficient. The quantization unit 103 can select four types of quantization factors, and these quantization factors are a quantization factor 1, a quantization factor 2, a quantization factor 3, and a quantization factor 4. The quantization unit 103 quantizes the transform coefficient with the quantization factor selected by the quantization factor selection unit 109. The variable length coding unit 104 variable length codes the quantized transform coefficient.

The quantizing unit 103 collectively quantizes the transform coefficients of 10 blocks with the same quantization factor.
The video signal of 10 blocks to be processed at one time will be referred to as a segment hereinafter, the segment consisting of the first block to the tenth block is the first segment, the segment consisting of the eleventh block to the twentieth block is the second segment, generally the A segment consisting of the (× n-9) th block to the 10th × n block is called the nth segment.

The output terminal 105 collectively outputs the output of the variable length coding unit 104 and the coded quantization factors selected by the quantization unit 103 for each segment. The code amount calculation unit 106 has four quantizers 1061, 1062 corresponding to the quantization factors selectable by the quantization unit 103 shown in FIG.
1063, 1064 and four variable length code amount calculator 106
5, 1066, 1067, 1068. The quantizer 1061 has the same function as the quantizer 103 when selecting the quantization factor 1 without fail. Similarly, the quantizer 106
2, 1063 and 1064 are quantization units 10 for selecting the quantization factor 2, the quantization factor 3 and the quantization factor 4, respectively.
It has the same function as 3.

Variable length code amount calculators 1065, 1066,
All of 1067 and 1068 have the same function, and the function obtains the output code amount when encoding is performed by the variable length encoding unit 104 without performing variable length encoding, and sums and quantizes the segments of this code amount. The sum of the code amounts for encoding the quantization factors selected by the unit 103 is obtained as the sum r of the code amounts per segment.

The code amount prediction unit 108 is an adder 1 shown in FIG.
081, prediction value temporary memory 1082, multiplier / divider 1083
7 for each quantization factor, and k = 1, 2, 3, 4 in FIG. The beginning of the frame of the video signal,
That is, before inputting the conversion coefficient of the first segment, 0 is set in the predicted value temporary storage unit 1082. The sum r of the code amount per segment is input, and the prediction value temporary storage 108
The contents of 2 are added by the adder 1081 and stored again in the predicted value temporary storage 1082 and input to the multiplier / divider 1083.

In the processing of the nth segment, the multiplier / divider 108
3 divides the input by n, multiplies (480-n), and outputs this value as the predicted code amount p together with the code amount r per segment. The prediction code amount p for each quantization factor is such that the average of the code amount per segment of the unprocessed segment is equal to the average of the code amount per segment of the already processed segment, and the unprocessed segment is subjected to the same quantization. It is the code amount when quantizing with a factor.

The quantization factor selection unit 109 is provided with subtractors 1091, 1092, 1093, 1094 and 1 shown in FIG.
098, comparison selector 1095, quantization factor memory 109
6, code amount selector 1097, remaining target temporary storage 1099
It consists of.

Before the start of the frame of the video signal, that is, the sum r of the code amount per segment of the first segment is input, the target value of the code amount of the frame to be encoded is set in the remaining target temporary storage 1099. , The quantization factor storage unit 1096 stores the quantization factor of the predetermined first segment.
Set to. The code amount selector 1097 selects the sum r of the code amounts per segment corresponding to the quantization factors stored in the quantization factor storage 1096, and the remaining target temporary storage 109
The subtractor 1098 subtracts from the content of 9 and stores it again in the remaining target temporary storage 1099.

Further, the predictive code amount p for each quantization factor is subtracted from the contents t of the remaining target memory 1099 by subtractors 1091 and 109.
The values subtracted at 2, 1093, and 1094 are input to the comparison selector 1095. The comparator / selector 1095 outputs the quantization factor that gives the smallest non-negative input, and stores it in the quantization factor storage 1096.

A video signal divided into frames is input from a video signal input terminal 101, a signal conversion unit 102 forms a block, and DCT conversion is performed to obtain a conversion coefficient. Quantization unit 1 using the quantization factor selected by quantization factor selection unit 109
Quantization is performed at 03, variable length coding is performed at the variable length coding unit 104, and the quantization factor selected by the quantization factor selecting unit 109 is output from the output terminal 105 together with the coded one. Also, the transform coefficient obtained above is input to the code amount calculation unit 106, passes through the code amount prediction unit 108, and then the quantization factor selection unit 10
9 selects the quantization factor used in the next segment.

In the first embodiment, the blocking unit 110 outputs the blocks in the order in which the blocks belonging to the same segment are always adjacent to each other in the image. However, it is necessary that the blocks belonging to the same segment are adjacent to each other in the image. There is no. For example, the blocking unit 110 is set to the n-th (n-
1) It is also possible to output the block of the video signal with the number obtained by x481% 4800 + 1 (where a% b represents the remainder obtained by dividing the integer a by the integer b). In this case, 10 blocks are set as segments in the order of output by the blocking unit, for example, the first segment is the first,
482, 963, 1444, 1925, 24
06, 2887, 3368, 3849, 433
It shall consist of 0 blocks.

When the deviation of the code amount for each segment in the frame when the video signal is coded becomes severe, the code amount prediction as compared with the first embodiment by configuring the blocking unit 110 as described above. The value p tends to be closer to the code amount when the unprocessed segment is quantized with the same quantization factor.

Next, a second embodiment of the present invention will be described. The configuration of the second embodiment and the function of each unit are the same as those of the first embodiment except for a part of the code amount prediction unit 108 and the quantization factor selection unit 109. Further, the video signal input to the video signal encoding device is the same as that described in the first embodiment.

The code amount prediction unit 108 includes a weighted adder 2081, a predicted value temporary storage 1082, and a shifter 20 shown in FIG.
It is assumed that the 83 sets are configured by a total of 4 sets with k = 1, 2, 3, 4 in FIG. 9 for each quantization factor. The beginning of the frame of the video signal, i.e. before processing the first segment,
The prediction value temporary storage 1082 is set to 0.

For the first segment to the 32nd segment, the sum r of the code amount per segment is input, the contents of the prediction value temporary storage unit 1082 and the weighted adder 2081 are simply added, and the prediction value temporary storage is performed again. It is stored in the container 1082. For the segments after the 33rd segment, the sum r of the code amount per segment is input, the contents of the predicted value temporary storage unit 1082 multiplied by 31/32 are added by the weighted adder 2081, and the predicted value temporary It is stored in the storage device 1082.

The shifter 2083 weights the adder 208 only for the n-th n-th segment where 480-n is a power of 2.
The addition result of 1 is multiplied by 480-n and divided by 32, and this value is output as the predicted code amount p together with the sum r of the code amounts per segment. The code amount prediction unit 108 outputs r of the code amount per segment to all the segments, and the predicted code amount p is the segment obtained by the shifter 2083, that is, 48.
Output only for the n-th n-th segment where 0-n is a power of 2.

The quantization factor selector 109 has the same structure as that of the first embodiment shown in FIG. Before the start of the frame of the video signal, that is, the sum r of the code amount per segment of the first segment is input, the target value of the code amount of the frame to be encoded is set in the remaining target temporary storage unit 1097 and is determined in advance. The quantization factor of the first segment is set in the quantization factor storage 1096. Quantization factor memory 109
6, the sum r of the code amount per segment corresponding to the quantization factor stored in 6 is selected by the code amount selector 1097, subtracted from the content of the remaining target temporary storage 1099 by the subtractor 1098, and the remaining target temporary storage is again stored. Stored in 1099.

In the segment to which the predictive code amount is inputted from the predictive code amount unit 108, the predictive code amount p for each quantization factor is subtracted from the content t of the remaining target memory 1099 by the subtractor 109.
The values subtracted by 1, 1092, 1093, and 1094 are input to the comparison selector 1095. Comparison selector 109
5 outputs a quantization factor giving a positive and minimum input, and stores it in the quantization factor storage unit 1096.

A video signal divided into frames is input from a video signal input terminal 101, and a signal conversion unit 102 blocks the signal and performs DCT conversion to obtain a conversion coefficient. Using the quantization factor selected by the quantization factor selection unit 109, the quantization unit 10
3 is quantized, variable-length coding unit 104 performs variable-length coding, and the quantization factor selected by quantization factor selection unit 109 is coded and output together from output terminal 105.
Also, the transform coefficient obtained above is input to the code amount calculation unit 106, passes through the code amount prediction unit 108, and then the quantization factor selection unit 10
9 newly selects the quantization factor to be used from the segment next to the n segment where 480-n is a power of 2.

Compared with the first embodiment, this embodiment has a tendency that more nearly optimal code amount control can be performed when the deviation of the code amount for each segment in the frame when the video signal is encoded is severe. is there. In addition, the circuit scale of the code amount prediction unit 108 can be made small.

Next, a third embodiment of the present invention will be described. The video signal coding apparatus described in this embodiment does not code the first frame to be input.

FIG. 4 is a block diagram showing a third embodiment of the video signal encoding device. The configuration and function of this apparatus are the same as those of the first embodiment except for the code amount storage unit 107 and the code amount prediction unit 108. Further, the video signal input to the video signal encoding apparatus of this embodiment is the same as that described in the first embodiment.

Code amount storage unit 107 and code amount predictor 1
Reference numeral 08 denotes a first-in first-out storage device 30 shown in FIG.
71, cumulative adder 3081, pre-code amount residual storage unit 30
It is assumed that the set of 82 and the subtractor 308 is composed of a total of 4 sets for each quantization factor. First-in first-out memory (FIRST
IN FIRST OUT REGISTER) is a storage device having a capacity for storing the sum r of code amounts per segment for one frame, and has a function of inputting values in order and extracting them in the input order.

In the first frame, the sum r of code amounts per segment is input and stored in the first-in first-out storage unit 3071. In the second and subsequent frames, the sum r of the code amount per segment is input, and the first-in first-out storage unit 307
At the same time, the sum r of the code amount per segment of the same segment of the immediately preceding frame is taken out and output. The first frame is the beginning of the video signal frame,
That is, before processing the first segment, the cumulative adder 3
0 is set to 081.

The sum r of the code amounts per segment is input and stored in the first-in first-out storage unit 3071, and the cumulative adder 3081 calculates the total sum r of the code amounts r of all the segments of one frame.

For the second and subsequent frames, the start of the frame of the video signal, that is, the sum of the code amount sums r of the segments of all the segments of the previous frame calculated by the cumulative adder 3081 before processing the first segment is performed. , The previous code amount remaining storage unit 3082 and the cumulative adder 3081 are set to 0.

The sum r of the code amounts per segment is input and stored in the first-in first-out storage unit 3071, and the cumulative adder 3081 calculates the total sum r of the code amounts of the segments of all the segments of one frame. In the above, the sum r of the code amount per segment is stored in the first-in first-out memory 3071.
At the same time, the sum r of the code amount per segment of the same segment of the previous frame is extracted. The sum r of the code amounts extracted per segment is the previous code amount remaining storage unit 3
The subtractor 3083 subtracts from the content of 082 and stores again in the previous code amount remaining storage unit 3082, and this value is set as the predicted code amount p as the code amount r per segment of the current frame.
Output with.

The predictive code amount p for each quantization factor is the sum of the sum of the code amounts per segment of the segment of the immediately preceding frame after the currently processed segment.

A video signal divided into frames is input from a video signal input terminal 101, and a signal conversion unit 102 blocks the signal and performs DCT conversion to obtain a conversion coefficient. The code amount for each quantization factor that can be selected by the quantization unit 103 is calculated from the transform coefficient by the code amount calculation unit 106, and stored in the code amount storage unit 107. In the second and subsequent frames, in addition to the above, the transform coefficient is quantized by the quantization unit 103 using the quantization factor selected by the quantization factor selection unit 109, and the variable length coding unit 10
4, the variable length coding is performed, and the coding of the quantization factor selected by the quantization factor selecting unit 109 is combined and output terminal 10
Output from 5. In addition, the code amount prediction unit 108 calculates the sum of the code amount per segment obtained by the code amount calculation unit 106 and the sum of the code amount per segment of the previous frame stored in the code amount storage unit 107, and the code amount of the unprocessed segment. , And the quantization factor selection unit 109 selects the quantization factor to be used in the next segment.

This embodiment is different from the first and second embodiments in that
When encoding a video signal consisting of consecutive frames,
There is a tendency that more nearly optimal code amount control can be performed.

Although the first frame is not coded in the third embodiment, the first frame is coded by the first and second embodiments by combining with the first and second embodiments. It is also possible to perform the encoding and to encode the second and subsequent frames in the third embodiment. Further, it is possible to provide a means for judging a scene change of the input video signal, and also in the second and subsequent frames, the frame immediately after the scene change can be encoded by the first and second embodiments. .

Although the above embodiment has described the case where the frames of the video signal are coded in order, it can be applied to the case where the inter-frame difference is coded. For example, when the odd-numbered frames are coded as they are and the even-numbered frames are coded with the difference from the previous frame, the code amount storage unit 107 for odd frames and even frames is provided in the third embodiment. It is also possible to switch to and use.

In the above embodiment, the output terminal 105
It is also possible to provide a means for detecting the code amount to be output from, and to subtract the detected code amount from the contents of the remaining target temporary storage unit 1087 by the subtractor 1098 of the quantization factor selecting unit 109.

Further, in the above embodiment, the orthogonal transformation section 111 is used.
Although the DCT transform is used in the above, other orthogonal transform such as the wavelet transform, or a transform other than the orthogonal transform may be used.

In the above embodiment, the case where there are four kinds of quantization factors has been described, but the number of kinds can be changed.

[0061]

As described above, according to the present invention, it is possible to select an appropriate quantization factor by obtaining the predicted value of the code amount for each quantization factor, and control the code amount for a predetermined fixed period. However, the amount can be reduced to a predetermined amount or less.

[Brief description of drawings]

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

FIG. 2 is a block diagram showing the configuration of a conventional encoding device.

FIG. 3 is a block diagram showing another configuration of a conventional encoding device.

FIG. 4 is a block diagram of a video signal encoding apparatus according to a third embodiment of the present invention.

FIG. 5 is a diagram showing blocks of a video signal for explaining an embodiment of the present invention.

FIG. 6 is a block diagram of a code amount calculation unit according to the embodiment of this invention.

FIG. 7 is a block diagram of a code amount prediction unit according to the first embodiment of this invention.

FIG. 8 is a block diagram of a quantization factor selection unit according to the embodiment of the present invention.

FIG. 9 is a block diagram of a code amount prediction unit according to a second embodiment of the present invention.

FIG. 10 is a block diagram of a code amount storage unit and a code amount prediction unit according to a third embodiment of the present invention.

[Explanation of symbols]

 102 signal conversion unit 103 quantization unit 104 variable length coding unit 106 code amount calculation unit 107 code amount storage unit 108 code amount prediction unit 109 quantization factor selection unit

Claims (3)

[Claims] 1. A video signal is input, and encoding is performed such that a code amount for a predetermined fixed period is equal to or less than a predetermined fixed amount.
A video signal coding device for outputting the coded video signal, comprising a signal conversion unit for collecting video signals in a predetermined fine unit, performing signal conversion to obtain a conversion coefficient, and a conversion coefficient in the fine unit. A quantizer for quantizing with a selectable quantizer, a variable-length encoder for variable-length coding the quantized signal, the quantization and the variable-length coding for the conversion coefficient of the fine unit. A code amount calculation unit that calculates the code amount for each of the quantization factors, and the code amount of the video signal in the fine unit for each of the quantization factors, the code amount calculation unit calculates the code amount. A code amount prediction unit that predicts from a code amount, and a quantization factor selection unit that selects the selectable quantization factor based on the predicted code amount are provided, and the variable length encoded signal sequence and the selected Video characterized by outputting a quantization factor No. encoding device. 2. In the quantization of the transform coefficients of a plurality of predetermined fine units in the quantizer, the quantizer selection unit selects a quantizer, and the same quantizer as the previous one is used except for the quantizer. The video signal encoding device according to claim 1, wherein the video signal encoding device performs quantization with a factor. 3. A code amount storage unit for storing the code amount calculated by the code amount calculation unit for a predetermined period, and the code amount prediction unit includes:
The code amount for each selectable quantization factor of the video signal in the fine unit is stored in the code amount storage unit near the image represented by the video signal in the fine unit, and in the code amount calculation unit. The video signal encoding device according to claim 1, wherein the video signal encoding device predicts from the calculated code amount.