JPH07143479A - Video compression coder - Google Patents

Abstract

PURPOSE:To obtain a video compression coder in which video data are compressed while suppressing deterioration in quality of a video image when a large margin of a code quantity is taken. CONSTITUTION:After input video data are transformed into a DCT coefficient by a DCT section 11, a quantization section 12 applies quantization at a quantization level designated by quantization control data and the quantization data are inputted to a variable length coding section 13, in which compression coding is made in the unit of a predetermined block. A code quantity detection section 14 detects the code quantity of coding data and the result of detection is given to a quantization control section 15. A code quantity - quantization level control characteristic representing the quantization level and the code quantity to obtain an average code quantity of input picture data is given in advance in the quantization control section 15 and a quantization level of a succeeding block is decided based on the characteristic and the code quantity information from the code quantity detection section 14 to generate and output corresponding quantization control data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to transmission of video data,
The present invention relates to a video compression encoding device used for recording.

[0002]

2. Description of the Related Art Conventionally, in the transmission of video data, there are many cases in which the video data is compressed and encoded for transmission due to the limit of the data transmission capacity of the transmission line or the cost. Further, even in the video recording apparatus, the video data may be compressed and encoded and recorded due to the cost of the recording medium.

In a conventional video compression coding apparatus for performing such compression coding processing of video data, control is performed so that the code amount per unit time becomes constant according to the capability of the transmission line or the recording medium. ing.

Specifically, as shown in FIG. 7, when the code amount dc to be controlled is constant, when the code amount is large, the quality of the image, that is, the quantization level is lowered to reduce the code amount. When the amount is small, the image quality, that is, the quantization level is increased to increase the code amount.

At this time, the code amount changes according to the control of the quantization level. However, since the average code amount falls within the target code amount dc, it is possible to absorb the variation in the code amount by using the buffer memory in the output stage. This prevents the code amount after compression from changing depending on the picture pattern,
The code amount is kept constant.

However, in the video compression coding apparatus which makes the code amount constant as described above, the quality of the video varies greatly depending on the design. FIG. 8 shows a range in which the quantization level Q changes when the code amount D of the video data shown in FIG. 8 is controlled to be constant. From the figure, it can be seen that the image quality greatly changes depending on the design.

On the other hand, for example, some video recording apparatuses using a semiconductor memory as a recording medium have a certain allowable range of code amount. In such a case, it is not always necessary to make the code amount constant, and it is desirable to maintain the quality of video as much as possible.

[0008]

As described above, the conventional video compression coding apparatus employs the constant code amount method, and therefore has the drawback that the quality of the video varies greatly depending on the design. It was

The present invention has been made to solve the above problems, and provides a video compression encoding apparatus capable of compressing video data while suppressing deterioration of video quality when a large margin of change in code amount can be taken. The purpose is to do.

[0010]

In order to solve the above problems, a video compression coding apparatus according to the present invention is a quantizer for quantizing input video data at a quantization level designated by quantization control data. A variable length coding unit for compressing and coding the quantized data output from the quantizing unit in predetermined block units, and a code amount for detecting the code amount of the coded data output from the variable length coding unit. A detection unit, and a quantization control unit that generates quantization control data so that the quantization level of the quantization unit becomes an appropriate level based on the detection result of the code amount detection unit and outputs the quantization control data to the quantization unit. Wherein the quantization control unit has a code amount indicating a relationship between a code amount and a quantization level for obtaining an average code amount of input image data in advance−
A quantization level control characteristic is given, the quantization level of the next block is determined based on the characteristic and the code amount information from the code amount detection unit, and the corresponding quantization control data is generated and output. Is characterized by.

[0011]

In the video compression coding apparatus having the above configuration, the input video data is quantized at the quantization level specified by the quantization control data in the quantizing unit, and the quantized data is input to the variable length coding unit and predetermined. Compress and encode in block units. The code amount of this encoded data is detected by the code amount detecting unit, and the detection result is input to the quantization control unit. This quantization control unit is previously given a code amount-quantization level control characteristic indicating the relationship between the code amount and the quantization level for obtaining the average code amount of the input image data. The quantization level of the next block is determined based on the code amount information from the detection unit, and the corresponding quantization control data is generated and output.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 shows the configuration of a video recording / reproducing apparatus using the video compression encoding apparatus according to the present invention.

In FIG. 1, input video data is converted into DCT coefficients by a DCT unit (discrete cosine transform unit) 11,
After being quantized by the quantizer 12 based on the quantized control data, the variable length encoder 13 encodes each frame and sends it to the video data recording / reproducing device 21.

The code amount detector 14 takes in the encoded data output from the variable length encoder 13 and detects the code amount, and the information is sent to the quantization controller 15. The quantization control unit 15 determines the appropriate quantization level from the input code amount information, and sends the quantization control data corresponding to the determination result to the quantization unit 12.

The video data recording / reproducing apparatus 21 uses, for example, a semiconductor memory as a recording medium, and in the recording mode, encoded data output from the variable length encoding unit 13 and quantization control data output from the quantization control unit 15. Are read together and sequentially written into the semiconductor memory, and in the reproduction mode, arbitrary encoded data is read out from the semiconductor memory and corresponding quantization control data is also read out.

The encoded data reproduced by the video data recording / reproducing apparatus 21 is sent to the variable length decoding unit 31, and the quantization control data is sent to the dequantization unit 32. The encoded data input to the variable length decoding unit 31 is restored to the original quantized data, and the dequantization unit 32 dequantizes the dequantized data based on the quantized control data. Then, the inverse DCT unit 33 performs an inverse DCT conversion process, whereby the original video data is demodulated.

In the above configuration, the algorithm of the quantization controller 15 which is a feature of the present invention will be described below with reference to FIGS. 2 to 4. Now, the image recording / playback unit 21
As a recording condition of, the code amount D of each frame is at least dL =
From 0 kilobytes to a maximum dU = 300 kilobytes, the quantization level Q to be ensured as video quality is at least qL = 90% (the quantization level Q of input video data is represented by 1 to 100%). Shall be

First, the code amount D of each frame is accumulated for a long time and divided by the number of accumulated frames to obtain the average code amount dav. Further, a DQ (code amount-quantization level) control characteristic for controlling the quantization level so that the output data of the variable length coding unit 13 has the average code amount dav is obtained. Since this characteristic depends on the design, it is necessary to encode images of various designs and investigate the tendency.
An example is shown in FIG. In FIG. 3, the quantization level Q is set to qH = 96% at the maximum so that a slight compression effect can be obtained.

FIG. 2 shows a control flow when the DQ control characteristic shown in FIG. 3 is set in the quantization controller 15. First, in step S1, it is determined whether or not the current quantization level qi is 90% or more. If it is less than 90% (NO), it is further determined in step S2 whether or not the current code amount di is 300 kilobytes or less. If it is less than 300 (YES), the quantization level is set to 1 in step S3.
Generate quantized control data for increasing the number of steps, and
If it exceeds the kilobyte (NO), the quantization control data for reducing the quantization level by one step is generated in step S4, and the process returns to step S1 to process the next frame.

On the other hand, when the quantization level is 90% or more in step S1 (YES), D in step S5.
Referring to the -Q control characteristic, the quantization level qa on the characteristic corresponding to the same code amount di is obtained. Then, step S6
Where q0 = k (qi + qa) (where k is 0.5 as standard)
To determine the quantization level q0 for the next frame, generate corresponding quantization control data in step S7, and return to step S1 to process the next frame.

When the control flow is quantized, as shown in FIG. 4, the code amount D is a pattern (quantization level 90% or more, code amount 300) within the allowable code amount dU.
(Kilobytes or less) will vary greatly. However, averaging over a long period of time approaches the average code amount previously obtained.

Since the average code amount obtained in advance is a predicted value, it may happen that the average code amount does not match the average code amount of the actually input video data. In such a case, it is possible to deal with the input video data by calculating the average code amount in the past specified number of frames and correcting the value of k based on the average code amount.

Further, even if the value of k is not changed and the DQ control characteristic itself is adjusted from the actually measured average code amount, the same effect can be obtained. In this case, automatic adjustment is possible by feedback control. FIG. 5 shows the control flow, and FIG. 6 shows how the DQ control characteristic changes due to the control.

First, in step S11, the code amount of the current frame is fetched, and in step S12, the average code amount of the past 1000 frames including the code amount of the current frame is calculated. Then, in step S13, the average code amount which is the basis of the current DQ control characteristic is compared.

Here, when the difference between the two is within ± 10%, D
The -Q control characteristic is left unchanged as indicated by A in FIG. 5 (step S14). Further, when it is less than -10%,
As shown in the middle B, the slope of the DQ control characteristic is increased by one step (step S15). On the contrary, when it is + 10% or more, the slope of the DQ control characteristic is lowered by one step as shown by C in FIG. 5 (step S16).

Finally, in step S17, the quantization level Q of the next frame is calculated by approximation calculation based on the newly obtained DQ control characteristic, and corresponding quantization control data is generated. It is sent to the quantizer 12. As a result, the optimum DQ control characteristic can always be obtained,
More effective compression encoding can be realized.

In the above embodiment, the case where the present invention is applied to the video recording / reproducing apparatus in which the margin of the code amount change is large has been described. However, the present invention is not limited to this, and a transmission rate such as ATM transmission can be used. It can also be applied when there is elasticity. The encoding may be performed in arbitrary block units such as field units instead of frame units. Needless to say, the present invention can also be implemented in various ways without departing from the scope of the invention.

[0028]

As described above, according to the present invention, it is possible to provide a video compression coding apparatus capable of compressing video data while suppressing deterioration of video quality when a large margin of change in code amount can be taken. it can.

[Brief description of drawings]

FIG. 1 is a block circuit diagram showing a configuration of a video recording / reproducing apparatus using the video compression encoding apparatus according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a basic control algorithm of the quantization controller of the same embodiment.

FIG. 3 is a characteristic diagram showing a code amount / quantization level control characteristic preset in the quantization control unit.

FIG. 4 is a waveform diagram showing a change in code amount under the control of the quantization controller.

FIG. 5 is a flowchart showing a control flow for automatically adjusting the code amount-quantization level control characteristic of the quantization controller.

FIG. 6 is a characteristic diagram showing how the code amount-quantization level control characteristic is adjusted by the control flow of FIG.

FIG. 7 is a waveform diagram showing a change in code amount by a conventional video compression encoding device.

FIG. 8 is a diagram showing a control width of a quantization level of a conventional video compression encoding device.

[Explanation of symbols]

11 ... DCT unit, 12 ... Quantization unit, 13 ... Variable length coding unit, 14 ... Code amount detection unit, 15 ... Quantization control unit, 21 ...
Video data recording / reproducing device, 31 ... Variable length decoding unit, 32
... inverse quantization section, 33 ... inverse DCT section.

Continuation of front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical indication location H04N 1/41 B (72) Inventor Hideyuki Shimada 1 Komukai Toshiba-cho, Saiwai-ku, Kawasaki-shi, Kanagawa Toshiba Corporation Komukai Factory

Claims (4)

[Claims] 1. A quantizer that quantizes input video data at a quantization level specified by quantization control data, and a variable that compresses and encodes the quantized data output from the quantizer in predetermined block units. A long encoding unit, a code amount detecting unit for detecting the code amount of encoded data output from the variable length encoding unit, and a quantization level of the quantizing unit based on the detection result of the code amount detecting unit. , And a quantization control unit that generates quantization control data so as to be an appropriate level and outputs the quantization control data to the quantization unit, wherein the quantization control unit is for obtaining the average code amount of the input image data in advance. A code amount-quantization level control characteristic indicating the relationship between the code amount and the quantization level is given, and the quantization level of the next block is determined based on the characteristic and the code amount information from the code amount detection unit. The corresponding quantization control data Video compression encoding apparatus is characterized in that so as to generate output data. 2. The quantization control unit, when the current block is compressed at a quantization level qi, sets the code amount after compression to di, and a quantization level corresponding to di from the code amount-quantization level control characteristic. 2. The method according to claim 1, wherein qa is obtained, k (qi + qa) is approximated using a constant of 0 <k <1, and the result of the calculation is determined as the quantization level of the next block. Video compression encoding device. 3. The quantization controller sets the value of k to 0 <
3. The video compression encoding apparatus according to claim 2, which has a function of arbitrarily correcting in the range of k <1. 4. The quantization control unit obtains an average code amount of past specified number of blocks every time the code amount information from the code amount detection unit is input, and the average code amount and the current code amount-
2. The video compression coding apparatus according to claim 1, wherein the average code amount which is the basis of the quantization level control characteristic is compared, and the control characteristic is adjusted according to the comparison result.