KR101337410B1 - bit-rate control method for a macro-block - Google Patents

Abstract

The present invention is a method for determining a bit rate control amount for a block by predicting bit generation probability by inter-block average and variance in an H.264 encoding system of a video encoder of an ISDB-T and a digital cinema mastering system.
In the video encoder, a method of controlling a bit rate for a large block includes, after initialization, predicting a motion vector of a 16x16 block, calculating an average and a variance of each 4x4 block in the motion vector prediction result of the 16x16 block, and Gaussian distribution. Calculating a bit generation prediction value using the average and the variance information used in the step of calculating a strength of a bit generation amount by multiplying the bit generation prediction value and the variance; Obtaining a quantization coefficient control value based on the intensity of the bit generation amount, and controlling a bit rate of an image according to the quantization coefficient control value.
In the present invention, since the bit rate control amount for the block mode except for the I blocks in the P and B frames can be determined at high speed by only the first 16x16 block operation in video encoding, the discrete cosine operation (DCT) and quantization operation or two- Since the pass bit rate control is not necessary, the computation amount is drastically reduced, and the computation time can be drastically reduced.

Description

Bit-rate control method for a macro-block}

The present invention is a method of controlling the bit rate of an image by adding or subtracting quantization coefficients for a block by predicting the possibility of bit generation in a video encoder of an ISDB-T and a digital cinema mastering system. A method of controlling bit rate by using this information by predicting the possibility of bit generation performed by variance. The present invention relates to a bit rate control method that does not undergo discrete cosine coding and quantization for separate bit rate calculation for real-time video encoding. .

Digital video data used in digital cinema, high-definition television, video-on-demand (VOD) receivers, personal computers that support moving picture experts group (MPEG) video, game consoles, terrestrial digital broadcast receivers, digital satellite broadcast receivers, and cable television (CATV) Compression is compressed by an efficient compression method rather than being used as it increases the amount of data in the process of digitizing the analog signal and the characteristics of the image itself.

The compression of digital image data uses three methods. A method of reducing temporal redundancy, a method of reducing spatial redundancy, and a method of reducing using statistical characteristics of generated codes are mainly used. Among them, a representative method of reducing temporal redundancy is motion estimation and compensation, which is used in most video compression standards such as MPEG and H.264.

Most video standards find the most similar part from the previous or subsequent reference screen for a specific part of the current screen, and transmit only the difference component of the two parts. In many cases, operations are performed in units of macroblocks. Therefore, block-based video compression is the most representative algorithm for video compression.

The block unit estimation method divides an image into blocks of a predetermined size and finds a block that best matches a block of the current image in a search region of a previous image. The difference between the found block and the current video block is called a motion vector, which is encoded and processed. Various matching functions can be used to calculate the matching between two blocks. The most commonly used is sum of absolute difference (SAD), which is a sum of absolute values of the difference between pixels between two blocks.

Unlike the existing video codec, the H.264 codec searches through the Cost function based on Rate Distortion Optimization (RDO) instead of the SAD-oriented search method. The Cost function used in H.264 searches using the Rate-Distortion Cost (RD-Cost), which is the sum of the number of encoded coefficients multiplied by the Lagrangian multiplier. Therefore, in order to obtain RD-Cost of H.264, SSD (Sum of Squared Difference) is often used instead of SAD. Therefore, RD-COST is different from that of video codec using SAD.

In addition, in the H.264 video encoding, in order to simultaneously obtain compression efficiency and high image quality, the eight video block modes are all different from the conventional video encoding, which is performed in units of 16 to 16 large blocks or 8 to 8 large blocks. It is configured to select the mode having the minimum value among each block with.

However, in order to determine eight different blocking modes, not only all the encoding operations are performed independently for each mode, but also when the block rate control is performed for the improvement of image quality and the bit reduction, the conventional block mode decision result is All must be re-determined by the newly determined quantization coefficients, resulting in a tremendous amount of computation.

Therefore, in order to achieve real-time video encoding, an efficient algorithm should be applied to block-level bit rate control to reduce image quality while minimizing operation increase by mode determination and block-level bit rate control.

Accordingly, an aspect of the present invention is to solve the above-described problems, and a technical problem to be solved by the present invention is to determine a block rate in a video encoder of an ISDB-T and a digital cinema mastering system. Its purpose is to provide a method to perform quickly.

Another object of the present invention is to provide a method for encoding a high quality image while performing bit rate control in parallel through a bit probability prediction method in a video encoder of an ISDB-T and a digital cinema mastering system. .

According to an aspect of the present invention for achieving the above object, a method of controlling a bit rate for a large block in a video encoder after the initialization step of predicting the motion vector of the 16x16 block, each 4x4 in the motion vector prediction result of the 16x16 block Calculating a mean and variance of the block, calculating a bit generation prediction value using the mean and variance information used in the Gaussian distribution, and calculating the strength of the bit generation amount by multiplying the bit generation prediction value and the variance. With the steps. Obtaining a quantization coefficient control value based on the intensity of the bit generation amount, and controlling a bit rate of an image according to the quantization coefficient control value.

According to the present invention, since the bit rate control amount for the block mode except for the I blocks in the P and B frames can be determined at high speed only by the first 16x16 block operation in the video encoding, the discrete cosine operation (DCT) and the quantization operation for the bit rate control or two Since there is no need for -pass bit rate control, the amount of computation is greatly reduced, and computation time can be drastically reduced.

1 is a block diagram of an H.264 encoding system of a video encoder of an ISDB-T and a digital cinema mastering system according to the present invention.
2 is a diagram illustrating a large block division mode used in an H.264 video codec, and is a diagram for explaining an H.264 block mode applied to the present invention.
3 is a conceptual diagram of the similarity between a DCT and a probability density function for predicting bit probability.
4 is a flowchart illustrating a method of controlling a bit rate for a block using bit probability prediction according to an actual implementation of the present invention.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It should be understood, however, that the invention is not intended to be limited to the particular embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. Like reference numerals are used for like elements in describing each drawing.

The terms first, second, A, B, etc. may be used to describe various elements, but the elements should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

When a component is referred to as being "connected" or "connected" to another component, it may be directly connected to or connected to that other component, but it may be understood that other components may be present in between. Should be. On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting of the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are to be interpreted as either ideal or overly formal in the sense of the present application Do not.

Hereinafter, embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

In order to achieve the above object, the present invention derives a bit probability prediction method by definition of bit probability in a video encoder of an ISDB-T and a digital cinema mastering system, thereby performing a block rate determination. Describes this.

Hereinafter, a method of predicting bit probability for an H.264 video encoder according to an embodiment of the present invention and a method of controlling a bit rate of a block unit to which the same is applied will be described in detail with reference to the accompanying drawings.

FIG. 1 is a block diagram of a video encoder (H.264 and MPEG-4 Video Compression: Video Coding for Next-Generation Multimedia ', Iain EG Richardson, Wiley, 2003) of a mastering system for a digital cinema applied to the present invention. Indicates.

The video encoder shown in FIG. 1 includes two dataflow paths, a "forward" path (outward to right) and a "reconstruction" path (right to left).

In the forward path, the input frame or field F _n 101-1 is processed in units of blocks. Each large block is encoded in an intra mode or an inter mode, and for each block in the large block, a predictive block PRED (denoted as 'P' in FIG. 1) 143 is formed based on reconstructed picture samples.

In intra mode 135, the PRED is taken from samples in the current slice that were previously coded, decoded, and reconstructed ( _denoted as uF'n 157 in FIG. 1, unfiltered samples are used to form the PRED). Are formed (132, 134).

In inter mode 123, PRED is a motion estimator (ME) 112 and a motion compensation: from one or two reference picture (s) selected from a set of list 0 and / or list 1 reference pictures. Formed by motion-compensated prediction by MC (122).

In Fig. 1, the reference picture is the previous coded picture F ' _{n -1} (102-1) 101-1, but the prediction criteria for each large block division (in inter mode) are already coded, reconstructed and filtered Can be selected from a selection of past or future pictures (in the display order). The predictive PRED is subtracted from the current block in subtractor 102 to produce a residual (difference) block D _n , where D _n is reordered (108) and entropy coded (110). Transformed at converter 104 and quantized at quantizer 106 to provide X 107. Entropy-encoded coefficients, along with auxiliary information (prediction modes, quantization coefficients, motion vector information, etc.) needed to decode each block in a large block, are added to a Network Abstraction Layer (NAL) for transmission or storage. Form a compressed bitstream to be delivered.

In addition to encoding and transmitting each block of the large block, the encoder decodes (reconstructs) each block to provide a reference for further predictions. The coefficients X are scaled (in inverse quantizer 152) and inverse transformed (in inverse transformer 154) to produce a difference block D ' _n . Prediction block PRED is added to D' _n to produce reconstructed block uF ' _n (decoded block of original block; u indicates unfiltered) 157. Filter 158 is applied to reduce the blocking distortion effect, and the reconstructed reference picture is generated from a series of blocks F ′ _n 161.

In particular, the motion estimator 112 includes a refiner estimator for estimating a motion vector of a refinery unit and a subpixel estimator for estimating an optimal half pixel and a motion vector of a quarter pixel unit based on the found motion vector of the refinery unit. do.

On the other hand, as described above, in order to simultaneously obtain compression efficiency and high image quality, in H.264 video encoding, unlike the conventional video encoding, all 8 different blocking modes and modes having the minimum value in each block are used. It is configured to choose.

2 is a diagram illustrating a large block division mode used in an H.264 video codec, and is a diagram for explaining an H.264 block mode applied to the present invention.

As shown in FIG. 2, in H.264 video encoding, 16 x 16 large blocks, 16 x 8 large blocks, 8 x 16 large blocks, 8 x 8 large blocks, 8 x 4 large blocks, 4 x 8 large blocks, and 4 Eight block modes of four to four blocks are included.

Hereinafter, a method of controlling a bit rate of a video encoder of a digital cinema mastering system using the predicted bit probability predictive method and the predicted bit probability will be described.

First, the difference between the prior art and the present invention will be described in order to help understanding of the present invention.

Conventionally, 4x4 DCT and quantization are performed to derive a coded block pattern (CBP) instead of the concept of bit likelihood, and the CBP is obtained through the result, and then the skipped (SKIP) mode or the B frame is used in the P frame. Was used for direct mode determination. For the remaining modes shown in FIG. 2, a refinement and a subpixel estimation are performed, and a rate distortion calculation is performed to select a mode having a minimum rate distortion value.

On the other hand, in the present invention, after the first 16x16 mode refiner and the subpixel estimation are finished, the average value and the variance of the block determined by the estimation are calculated without performing a separate DCT operation and quantization operation, and then, the bit probability value is obtained through this. The control value for bit rate control between blocks is determined by predicting and predicting bit probability.

A bit rate control method for a large block of the present invention will be described with reference to FIGS. 3 and 4.

Referring to FIG. 4, first, initialization for large block encoding is performed in step 402. Next, in step 404, a motion vector (MV) of a 16 × 16 block is predicted. Next, in step 406, the average and variance of each 4x4 block are calculated from the motion vector prediction result of the 16x16 block. At this time, the average and variance of each 4x4 block uses a CBP (Coded Block Pattern). The CBP indicates whether the coefficient values, which are the result of performing the DCT on one block, are all zero or not. When the large block type is Intra, the CBP sets the first value to IntraDC and determines the CBP based on the remaining coefficient values except for this value. If the counts are all zeros, the CBP is zero, otherwise it is 1.

In general, the CBP (Coded Block Pattern) is determined as follows.

Referring to FIG. 4, first, initialization for large block encoding is performed in step 402. Next, in step 404, a motion vector (MV) of a 16 × 16 block is predicted. Next, in step 406, the average and variance of each 4x4 block are calculated from the motion vector prediction result of the 16x16 block. At this time, the average and variance of each 4x4 block uses a CBP (Coded Block Pattern). The CBP indicates whether the coefficient values, which are the result of performing the DCT on one block, are all zero or not. When the large block type is Intra, the CBP sets the first value to intra DC and determines the CBP based on the remaining coefficient values except for this value. If the counts are all zeros, the CBP is zero, otherwise it is 1.

In general, the CBP (Coded Block Pattern) is determined as follows.

를 움직임 보상된 4x4 화소 데이터의 DCT 결과라고 하면, 이때 CBP는

의 한 요소를 수학식 1과 같은 양자화 과정을 통해 결정된다.first

Is the DCT result of the motion-compensated 4x4 pixel data, where CBP

One element of is determined through a quantization process such as Equation 1.

의 i 번째 열 요소, j 번째 행 요소를 나타내는 첨자이며

는 양자화된

의 i 번째 열 요소, j 번째 행 요소 이며

는 양자화 되기 전

의 i 번째 열 요소, j 번째 행 요소 이며 QP는 H.264의 양자화 계수이다.

는 QP, i, j의 함수로서 양자화 계수에 12을 더한 것을 6으로 나눈 나머지와 i, j의 요소에 따라 결정되는 양자화 함수이며 f는 양자화 레벨 오프셋 값이다. In Equation 1 i, j is

The ith column element of, the jth row element of the subscript

Is quantized

The i th column element, the j th row element, and

Before quantization

Is the i th column element, the j th row element, and QP is the quantization coefficient of H.264.

Is a function of QP, i, and j, which is a quantization function determined according to the remainder of quantization coefficient plus 12 divided by 6 and the elements of i and j, and f is a quantization level offset value.

In Equation 1, if any one of the quantization coefficients is not 0, the CBP value of the 4x4 block is 1, otherwise it is 0.

The bit likelihood prediction is to consider the result of the DCT operation as a result of the two-dimensional probability density function, and to predict the likelihood of the bit occurrence by the mean and the variance as shown in FIG.

FIG. 3 shows a comparison between the one-dimensional DCT output value and the Gaussian probability density function in the present invention, where the portion having the highest value at which the frequency / variance value (X) and the DCT output value (Y) intersect is the average or DC component of the DCT. Is a graph showing the similarity between the DCT and the probability density function for predicting bit probability.

As shown in Fig. 3, the comparison between the one-dimensional DCT output value and the Gaussian probability density function has the highest value where X and Y intersect the average or DC component of the DCT.

In this case, the average value is the same as the DC component of the 4x4 DCT operation. For the other AC components except the DC component, the CBP is predicted to be 1 if the variance exceeds an adaptively determined threshold value, and 0 otherwise. .

The average is calculated as in Equation 2 and the variance is calculated as in Equation 3.

는 단위 4x4 블록에서 시간 t-k, i 번째 열과 j 번째 행에서의 화소 값이며

는

를 통해 예측된 4x4 블록의 i 번째 열과 j 번째 행에서의 화소 값이다. 비트 발생 가능성 예측은 수학식 4와 같이 수행되며 예측된 비트 발생 가능성인

가 0이 아니면 1의 값을 가진다. In Equations 2 and 3

Is the pixel value in the time tk, the i th column and the j th row in the unit 4x4 block

The

Pixel values in the i th column and the j th row of the 4x4 block predicted through. Bit likelihood prediction is performed as shown in Equation 4, and the predicted bit likelihood

If is not 0, it has a value of 1.

After calculating the average and the variance of each 4x4 block in step 406, it is determined in step 408 whether the frame is a P frame or a B frame, and proceeds to step 410 for the P frame to estimate the bit generation as shown in Equation 4 below. For the B frame, the process proceeds to step 412 to obtain a bit generation prediction value as shown in Equation 5 below.

는

이면 1, 그렇지 않으면 0의 값을 갖는 단위 계단 함수이며 θ는 비트 발생 가능성 예측을 위해 도입된 임계값(Threshold)로서 최소 자승법에 의해 2.5에서 3.5 사이로 구해진다.In Equation 4,

The

Is a unit step function with a value of 1, otherwise 0, and [theta] is a threshold introduced for predicting the probability of occurrence of bits, and is obtained from 2.5 to 3.5 by the least square method.

In this case, the P-picture obtains a bit generation prediction value using Equation 4, but the B-Picture does not use Equation 4 but uses Equation 5, which is a simplified form of Equation 4.

For the B-picture, even using only Equation 5, it can be determined that a bit of a large block does not occur when predicting bit generation with a high hit ratio of 97% or more. In the case of the P-picture, the hit rate is only 50% when using only Equation 5, or 75% or more when using Equation 4. Determine if it is.

There are two cases in which it is determined that a bit of a large block does not occur when predicting bit generation, that is, a case in which it is determined that a bit occurs even though no bit occurs, and a case in which a bit does not occur even when a bit occurs. In the former case, there is no problem because the image is corrected in the SKIP / Direct mode in the encoding process, but in the latter case, the image quality is deteriorated.

As such, the probability of bit generation can be predicted using only the average and variance information used in the Gaussian distribution, and in this case, the probability of bit generation can be predicted by about 50% of the computation amount compared to the case of using DCT. It's faster.

When the likelihood of bit generation is predicted, it is necessary to predict how many bits are generated. In other words, after the bit generation prediction is calculated, the bit generation strength is calculated in step 414.

값과

값을 사용하여 비트 발생량의 강도를 계산한다. 이 때 수학식 5에서 계산된

값은 실제 이산 여현 변환(DCT) 및 양자화 과정에서 DC 성분에 대한 값이 된다. 이 값과 각 4x4 블록의

값을 합한 것을 곱한 값을 비트 발생량의 강도로 놓는다.Specifically, when the probability of bit occurrence is predicted, the equations shown in Equations 2 and 3

Value and

The value is used to calculate the strength of the bit generation amount. At this time,

The value becomes a value for the DC component during the actual discrete cosine transform (DCT) and quantization process. Of this value and each 4x4 block

Multiply the sum of the values by the strength of the bit generation amount.

는 비트 발생량의 강도이다. 일반적으로, DC 값이 높은 경우라 하더라도 4x4 블록에서의 분산 값인

값이 작을 경우 비트 발생량은 극히 작아 대체로 32bit 이하의 값을 가지게 된다. 그러나, 블록내 복잡도가 높아 수백 비트의 값을 나타내는 경우는 대체로

값과

값의 곱이 매우 크게 나타나는 경우이다. In Equation (6)

Is the strength of the bit generation amount. In general, even if the DC value is high,

If the value is small, the bit generation amount is extremely small and usually has a value of 32 bits or less. However, if the complexity in the block is high and represents hundreds of bits,

Value and

This is the case when the product of the values is very large.

는 양자화 계수의 QP의 반 비례 함수이므로 QP값이 높아지면 비트 발생량의 강도는 작아지며 QP 값이 높아지면 비트 발생량의 강도는 커진다. 그러므로, 비트 발생량의 강도에 비례하여 QP 값을 높여 주도록 하면, 블록 단위의 비트 발생량은 작아지게 된다. 이때, 블록 단위의 제어 값의 경우 급격한 양자화 계수의 상승은 전체적인 화질 저하의 원인이 되므로 최대 2가 되도록 하며 다음 수학식에 의해 결정한다.After calculating the bit generation intensity, a quantization coefficient control value is calculated in step 416. Strength of bit generation

Is an anti-proportional function of the QP of the quantization coefficients. Therefore, as the QP value increases, the intensity of the bit generation amount decreases, and when the QP value increases, the strength of the bit generation amount increases. Therefore, if the QP value is increased in proportion to the strength of the bit generation amount, the bit generation amount in units of blocks becomes small. In this case, in the case of the control value in units of blocks, the sudden increase in the quantization coefficient causes the overall deterioration of image quality, so that the maximum value is 2 and is determined by the following equation.

는 양자화 계수의 제어 값이다.In equation (7)

Is the control value of the quantization coefficient.

After the quantization coefficient control value is calculated, in step 418, the quantization coefficient of the large block is determined as in Equation 8 below.

In Equation 8, t represents an index value for a large block, t represents an index of the current block, and t-1 represents an index value of the previous block.

On the other hand, when the intensity of the bit generation amount is weak, the quantization coefficient is appropriate or has a value that is too large. In this case, the negative quantization coefficient control value is generated through the frequency of generating the control value for the quantization coefficient. That is, when is too many cases, 0 gives a condition for generating a negative quantization coefficient as shown in Equation (9).

과

는 임의의 양수로

의 관계를 만족하며

는 Delta-Dirac 함수로 x가 0이 아니면 0, 0이면 1인 값이다. 예를 들어 최근 10개의

가 연속으로 0의 값을

+1 개 가지게 되면

는 -1의 값을 가지게 된다. 그런데 그 다음 블록에서도

가 0이면 음의 값이 나올 조건을 계속 만족하게 되므로 연속으로

는 음의 값이 된다. In Equation (9)

and

Is a random positive number

Satisfying the relationship

Is a Delta-Dirac function, where x is 0 if it is nonzero, and 1 if it is 0. For example, the last 10

Successively returns a value of 0

When you have +1

Has a value of -1. But in the next block

Is 0, it will continue to satisfy the condition that negative value will come out.

Is negative.

가 0 혹은 음의 값을 가지게 될 조건들은 비트량 발생 확률 이외에도 전체 슬라이스 혹은 전체 프레임에 할당되어야 할 비트량의 조건과 관계가 있으므로 비트량 발생 예측치에 의한 제어 조건은 이것으로 충분하다.

The condition that will have a value of 0 or negative is related to the condition of the bit amount to be allocated to the whole slice or the entire frame in addition to the probability of generating the bit amount, so the control condition by the bit amount generation prediction value is sufficient.

가 0이 되고 이것이 연속으로 나타날 확률이 높아지므로

가 음의 값을 가질 확률이 높아진다. 또한

가 음의 값을 가지게 되면 자연히 대 블록의 처리에 따라,

가 내려가기 시작하여 높은 비트 발생 예측 강도가 나타날 때 까지 양자화 계수가 낮은 값을 가질 수 있도록 하여 전체적인 화질을 향상시키도록 한다.In other words, it is predicted that a lot of bits are generated in each block, so that the quantization coefficient value increases, and when it accumulates and has a high value, it is naturally a control value.

Becomes 0 and the probability of this appearing consecutively increases

Has a higher probability of having a negative value. Also

Has a negative value, naturally according to the processing of large blocks,

The quantization coefficient can have a low value until it starts to go down until a high bit generation prediction strength appears, thereby improving the overall picture quality.

Thereafter, in step 420, the skip / direct mode is determined according to whether the CBP is all zeros, motion prediction is performed in step 430, and motion compensation is performed in step 424.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the present invention as defined by the following claims It can be understood that

Claims (6)

In encoding for video,
Predicting a motion vector of the large block;
Calculating an average and a variance of the divided block obtained by dividing the large block based on a prediction result of the motion vector of the large block;
Calculating a bit generation prediction value by using different algorithms based on the mean and the variance of the divided blocks by distinguishing a case where the target frame including the large block is a P frame and a B frame;
Calculating an intensity of a bit generation amount using the bit generation prediction value and the variance of the partition block;
Calculating a quantization coefficient control value which is inversely proportional to the intensity of the bit generation amount and whose maximum value is 2; And
And calculating a quantization coefficient for the large block by using the quantization coefficient control value. The method according to claim 1,
Computing the bit generation prediction value
If the target frame is a P frame,
Equation

Calculating the bit generation prediction value by
If the target frame is a B frame,
Equation

Calculating the bit generation prediction value by

Is the average of the split blocks,

Is the variance of the split block, QP is the quantization coefficient,

Is a function of QP, i, and j is a quantization function determined by the remainder of quantization coefficient plus 12 divided by 6 and the elements of i and j, f is the quantization level offset value,

The

Is a unit step function with a value of 1, otherwise 0, and is a large block of a video encoder characterized by a threshold introduced for predicting the probability of bit occurrence, obtained from 2.5 to 3.5 by the least-squares method. Bit rate control method for. The method according to claim 2,
Computing the bit generation prediction value
And calculating the bit generation prediction value using the mean and the variance of the partition block based on a Gaussian distribution. The method according to claim 1,
Computing the quantization coefficient control value
Equation

Calculating the quantization coefficient control value by

Is a strength of the bit generation amount, and is a predetermined value). The method according to claim 1,
The calculating of the quantization coefficient
Equation

Calculating the quantization coefficient by

Quantization coefficient control value, t is the index of the current large block, t-1 is the index of the previous large block) bit rate control method for a large block of the video encoder. The method according to claim 1,
Computing the quantization coefficients,
And calculating the quantization coefficient based on a property in which the quantization coefficient control value is inversely proportional to the strength of the bit generation amount.

Images