KR100207417B1 - Method and apparatus for controlling generation of bit rate in video encoding - Google Patents

Abstract

The present invention relates to a bit rate controller method that can respond strongly to a transmission bit rate by adaptively controlling a quantization step size in response to a channel request in a transmission system on the transmitting side. The final proportional constant of the bit rate function is calculated based on the proportional constant of the i-th macroblock obtained based on the bit rate and the other proportional constant obtained from the delayed previous macroblock, and the quantization parameter calculating means refers to the final proportional constant. By calculating the quantization parameter, the bit rate can be adaptively controlled at a bit rate corresponding to the target required by the channel, regardless of the size of the image, thereby preventing deterioration of image quality of the image reconstructed in the decoding system on the receiving side. will be.

Description

Coded bit rate control method and control device

1 is a schematic block diagram of a typical conventional video encoding system.

2 is a conceptual block diagram of an apparatus for controlling a coded bit generation rate according to a preferred embodiment of the present invention.

3 is a detailed block diagram of the proportional constant calculation unit shown in FIG.

4 is a graph showing the relationship between the quantization parameter and the bit rate function for explaining the operation of the bit rate control apparatus according to the present invention.

5 is a schematic block diagram of a typical coded bit rate control apparatus in the related art.

* Explanation of symbols for main parts of the drawings

82: proportional constant calculation unit 86: quantization parameter calculation unit

88: activity calculation unit 822: multiplier

824: retarder 826: constant calculating unit

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for controlling the generation rate of coded bits employed in an image encoder, and more particularly, to a method for controlling a bit generation rate to efficiently meet a channel request by controlling a quantization step size in accordance with a change amount of a macroblock. It relates to a control device.

As is well known in the art, the transmission of discrete video signals can maintain better image quality than analog signals. When a video signal composed of a series of image frames is represented in a digital form, a considerable amount of data must be transmitted, especially in the case of high-definition television (HDTV). However, since the usable frequency range of the conventional transmission channel is limited, in order to transmit a large amount of digital data, it is necessary to compress the data to be transmitted and reduce the amount of transmission. In addition, the compressed video signal and the audio signal are encoded through different coding techniques due to the characteristics of those signals. In this encoding, a video signal compression technique that generates a greater amount of digital data than the audio signal is particularly important. It can be said to take part.

Therefore, when transmitting a video signal, the transmitting side compresses and encodes the video signal using the spatial and temporal correlation of the video signal and then transmits the compressed and encoded video signal to the receiving side through a transmission channel.

Meanwhile, among various compression techniques mainly used for encoding an image signal, a hybrid encoding technique combining a stochastic encoding technique and a temporal and spatial compression technique is known to be the most efficient, and the present invention is a hybrid combining a temporal and spatial compression technique. Related to the coding scheme.

Most of hybrid coding schemes, which are one of the efficient coding schemes described above, use motion compensated DPCM (differential pulse code modulation), two-dimensional discrete cosine transform (DCT), quantization of DCT coefficients, VLC (variable modulation coding), and the like. Here, the motion compensation DPCM determines a motion of the object between the current frame and the previous frame, and predicts the current frame according to the motion of the object to generate a differential signal representing the difference between the current frame and the predicted value. This can be done for example by Staffan Ericsson's Fixed and Adaptive Predictors for Hybrid Predictive / Transform Coding, IEEE Transactions on Communication, COM-33, NO.12 (1985, December), or A motion Compensated Interframe from Ninomiy and Ohtsuka. Coding Scheme for Television Pictures, IEEE Transactions on Communication, COM-30, NO.1 (January, 1982).

More specifically, the motion compensation DPCM predicts the current frame from the previous frame according to the motion of the object estimated between the current frame and the previous frame. Here, the estimated motion may be represented by a two-dimensional motion vector representing the displacement between the previous frame and the current frame.

Typically, there are several approaches to estimating the pixel displacement of an object, which are generally classified into two types, one of which is a motion estimation method in blocks and the other is a motion estimation method in pixels. In the block-based motion estimation according to the present invention, a block of the current frame is compared with blocks of a previous frame to determine an optimal matching block, and thereafter, an interframe displacement vector (interframe to frame) for the entire block for the current frame to be transmitted. The extent to which the block has moved) is estimated.

Therefore, when transmitting a video signal, the transmitting side compresses and encodes the image signal in the buffer of the output side in order by taking into account the spatial and temporal correlation of the image signal in block units or pixel units through the above-described encoding technique. The video data is transmitted to the decoding system on the receiving side through the transmission channel at a desired bit rate in response to the channel request.

More specifically, the coding system on the transmitting side removes spatial redundancy of the video signal by using transform coding such as discrete cosine transform (DCT), and further uses temporal encoding of the video signal by using differential coding through motion estimation and prediction. By eliminating redundant redundancy, the video signal can be efficiently compressed.

As described above, an example of a typical encoding system using a hybrid encoding scheme that compresses and encodes a video signal using temporal and spatial correlation for compression transmission to a receiver may be of the type shown in FIG. . As shown in the figure, a typical coding system includes a subtractor 10, a DCT unit 20, a quantizer 30, an inverse quantizer 40, a local decoder 50, a motion prediction block 50, variable A variable length coding unit (hereinafter referred to as VLC) 60, a buffer 70, and a bit rate controller 80 are included.

In addition, as indicated by a dotted line in FIG. 1, the local decoder 40 includes an inverse quantizer 42, an IDCT unit 44, and an adder 46, and the motion prediction block 50 includes a frame memory 52. ), A motion estimation unit 54 and a motion compensation unit 56.

First, in the subtractor 10, the predicted previous frame signal provided from the motion compensator 56 in the motion prediction block 50 for motion compensation differential coding is subtracted from the current frame signal input from the input side. That is, the difference signal representing the difference pixel value is converted into a series of quantized DCT conversion coefficients through the DCT unit 20 and the quantization unit 30. Then, the quantized DCT transform coefficients are simultaneously provided to the dequantization unit 42 constituting the VLC unit 60 and the local decoder 40.

More specifically, the DCT unit 20 uses the cosine function to time-domain video signals (pixel data) for the differential signals based on the input motion estimation and prediction. Convert to In addition, the quantization unit 30 quantizes the DCT transform coefficients from the DCT unit 20 to a finite number of values through a nonlinear operation. The difference signal between the frame to be encoded and the predicted frame is quantized. Quantize. In such quantization, the quantization unit 30 adjusts the quantization step size by the quantization parameter QP calculated by the bit rate controller 80 based on the fullness of the data provided from the output side buffer 70. In the coding system as described above, the present invention is an improvement of a bit rate control scheme that can strongly meet the transmission bit rate by efficiently controlling the quantization step size according to the channel requirements, that is, the amount of change of each macroblock. Is related.

Accordingly, the VLC unit 60 encodes the quantized differentially coded image data (quantized DCT transform coefficients) into runs and coefficients through zigzag scanning as described above. More specifically, the VLC unit 60 is variable according to the frequency of occurrence of each code using a code table, that is, a code having a large frequency of occurrence of a code is a short length code, and a code of a long length has a small frequency of code generation. After encoding, the signal is provided to the output side buffer 70 for transmission to the receiving side. Here, the reason for allocating different lengths to all codes through the VLC unit 60 is to increase the coding efficiency by substantially reducing the average value of the code lengths.

On the other hand, the inverse quantization unit 42 and IDCT unit 44 which are employed in a typical encoder as a means for performing the motion prediction differential coding as described above to form a local decoder 40 are quantized with the DCT unit 10 described above. The video signal (quantized DCT transform coefficient), which has been compressed and coded through the unit 30, is restored to the original signal before being encoded for motion estimation and compensation, and provided to the adder 46, after which the adder 46 is added. Adds the restored current frame signal (differential signal) provided from the IDCT unit 44 and the predicted previous frame signal provided from the motion compensator 56 to the frame memory 52, thereby providing the frame memory 52. Is stored as the immediately preceding frame of the restored current frame signal, that is, the current frame signal input for current encoding.

Therefore, through this process, the previous frame signal stored in the frame memory 52 is continuously updated with the image data immediately before the currently encoded input image data.

Next, the motion estimation unit 54 estimates the motion of the current input frame to be encoded in 16 × 16 units within the predetermined search range from the previous frame stored in the frame memory 52, that is, the current frame and the most. A motion vector, which is a displacement value between candidate blocks of a similar previous frame, is determined and provided to the motion compensator 56. The motion compensator 56 performs a frame based on the output information (motion vector) from the motion estimator 54. The corresponding block of the previous frame is read from the memory 52 to generate a predictive frame, and then provided to the subtractor 10 and the adder 46, respectively. Although not shown in FIG. 1, the motion vector determined by the motion estimation unit 54 is encoded after a predetermined encoding process for transmission to the receiving side decoding system, and then sent to a transmitter (not shown).

Accordingly, the subtractor 10 obtains the difference signal (differential pixel value) by subtracting the current frame signal from the input side with the predicted previous frame signal provided from the motion compensator 56, and the difference signal thus obtained is The DCT and quantization of the difference signal as described above are performed by being provided to the DCT section 20 of the next stage.

As described above, a conventional typical bit rate controller (corresponding to reference numeral 80 in FIG. 1) employed in the encoding system and controlling the quantization step size based on the data fullness of the output side buffer is shown in FIG. Similarly, it is composed of a fullness calculator 84, a quantization parameter calculator 86, and an activity calculator 88.

First, the fullness calculating section 84 of the output side buffer 70 shown in FIG. 1 is based on the number of bits generated up to the previous macro block and the target number of bits corresponding to the image obtained from the bit allocation block (not shown). The fullness d is obtained by the following equation.

Where i is the index of the macroblock, B _i-1 is the amount of bits occurring immediately before the i-th macroblock, d _o is the last buffer of the previous image, T is the target number of bits in the picture, and MB Is the number of macro blocks per picture.

Then, as described above, when the fullness d of the buffer 70 is calculated by the fullness calculator 84, the quantization parameter calculator 86 calculates the quantization parameter Q based on the fullness of the buffer. Obtained by

Here, r is 2 bit rate / picture rate.

Accordingly, the activity calculator 88 considers the activity according to the quantization parameter Q calculated as described above, that is, calculates how much activity the macroblock currently being processed has a scale in the frame and rescales the value mquant. Is calculated, and the scaled value is a value that substantially determines the quantization step size in the quantization unit 30 shown in FIG.

As a result, in the conventional bit rate control method as described above, the bit rate depends on the value mquant to be rescaled according to its activity, and mquant depends on the reference quantization parameter Q. In addition, the quantization parameter Q depends on the fullness d of the buffer, since the fullness d of the buffer is not sensitive to the amount of change per actual macroblock (e.g., simple, complex, etc.) The fullness d of the buffer up to the frame is strongly dependent on the result.

In addition, according to the conventional bit rate control method, the effect of the above equation (2), which is a relation for obtaining the quantization parameter Q from the fullness d of the buffer, varies depending on the size of the image. The problem is that it is very difficult to fit.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems of the prior art, and in the coding system on the transmitting side, the bit generation rate control method capable of strongly meeting the transmission bit rate by adaptively controlling the quantization step size according to the channel request, and its Its purpose is to provide a control device.

According to the present invention, the quantization parameter is calculated based on the target number of bits of the currently processed picture and the bit generation amount of the quantized and variable-length coded macroblocks up to now. A coded bit rate control method for determining a quantization step size for a macroblock to be quantized based on a rescaled value calculated according to activity in a frame of a block, the i-th macroblock coded with the quantized and variable length coded bits. A proportionality constant of the quantization parameter bit rate function is obtained based on the bit generation amount of each macroblock, and the proportional constant value of the i-th macroblock obtained based on the quantization parameter and the calculated bit rate function is delayed. Another ratio obtained from the previous macro block On the basis of the constant value and calculating a proportional constant of the final bit rate function, and calculating the quantization parameter with reference to the final proportional constant, rate function for the quantization parameter, is calculated as follows,

Here, if substituted with α = (B _i-1 -T '(i-1-N) / N, the quantization parameter is calculated as follows:

The proportionality constant of the incidence function is calculated as

In the above equation, B _i-1 is the amount of bits generated so far, T 'is the target number of bits of the macroblock, and N is a constant bit rate control method for controlling the convergence speed of the controller.

According to another aspect of the present invention, there is provided an apparatus comprising: means for calculating a quantization parameter based on a target number of bits of a picture currently being processed and a bit generation amount of a quantized and variable length coded macroblock so far; An encoding bit generation rate control device having means for determining a quantization step size for a macroblock to be quantized based on a rescaled value calculated according to activity in a frame of the macroblock, wherein the control device comprises: i Means for generating a proportionality constant of the quantization parameter incidence function by multiplying the quantization parameter value of the first macro block by the bit generation amount up to the i-th macro block generated by the quantization parameter; Buffer means for delaying the proportional constant for the processing time of the macro block and outputting a proportional constant for the previous macro block; And means for generating one new proportional constant based on the proportional constant of the occurrence function provided by the multiplication means and the proportional constant of the previous macro block provided by the buffer means, wherein the quantization parameter calculating means comprises: The quantization parameter for the corresponding macroblock to be quantized is calculated with reference to the new proportional constant, and the new proportional constant is calculated as follows.

In the above equation, K ₁ ′ is a proportional constant of the current macroblock, K ₂ ′ is a proportional constant of the previous macro block, and β is a value for adjusting the weight of K ₁ ′ and K ₂ ′. To provide.

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

First, in implementing the bit rate control scheme according to the present invention, in order to realize this, if T / MB in Equation (1) is replaced with T ', Equation (1) is expressed as follows.

In the above Equation (3), d is an amount representing the degree of control of the bit rate as a data full state of the virtual buffer. Therefore, in the present invention, in order to propose an encoding system that does not perform bit rate control by the virtual buffer, the fullness d of the buffer is removed. That is, in the above formula (3), if d _o on the right side is turned to the left side, it is expressed as follows.

In Equation (4), since T '(i-1) is the target bit generation amount of the macro block to date and B _i-1 is the bit generation amount to the present, as a result, the difference between the two values indicates the degree of control. . Therefore, in the ideal bit rate control, the left and right sides of Equation (4) should both be zero. In the above Equation (4), the left and right terms equal to 0 means that the target bit amount and the actually generated bit amount up to the corresponding macroblock are the same.

On the other hand, when the bit generation amount due to the quantization parameter Q is represented by the function F (Q), the relationship between the bit generation amount B (ie, the information amount) and the function can be written as follows.

If the left and right sides of Eq. (4) are both zero, the following equation will be used. That is, when the fullness of the buffer is converted into a function F (Q) for the amount of bits generated by the quantization parameter Q, the above equation (4) is converted as follows.

Where N is a constant parameter for adjusting the convergence speed of the control device.

Here, rearranging the above expression (6),

Becomes Therefore, if α = (B _i-1 -T '(i-1-N) / N, the above equation (7) is converted as follows.

Equation (8) can be converted as follows.

As a result, the quantization parameter Q can be easily obtained from the above expression (9).

On the other hand, the method of returning to the above-described formula (5) and obtaining a function F (Q), which represents the amount of bits generated by the quantization parameter Q, is described.

First, according to the rate_distortion theory, the rate has a logarithmic relationship to distortion with respect to an input having a Gaussian model distribution. The quantization parameter Q also has a linear relationship to distortion. Therefore, since the relationship between the quantization parameter Q and the incidence rate can be modeled in inverse proportion, it will be a graph in the form as shown in FIG.

Therefore, the function F (Q), which means the bit rate by the quantization parameter Q, can be modeled as follows.

Therefore, using the above equations (9) and (10), an algorithm for the reference quantization parameter computing device of the new bit rate control system can be obtained.

However, when (alpha) becomes smaller than 0 in said Formula (8), it becomes impossible to apply said Formula (10). Therefore, at this time, since B is larger than T '(i-1-N) at α = (B _i-1 -T' (i-1-N) / N0, the generation amount may be too large. Set parameter Q to its maximum value.

2 shows a conceptual block diagram of an apparatus for controlling a coded bit rate rate proposed according to the present invention for implementing a new control system as described above. As shown in the figure, the bit rate control apparatus of the present invention is composed of a proportional constant calculator 82, a quantization parameter calculator 86, and an activity calculator 88.

In FIG. 2, the quantization parameter calculator 86 and the activity calculator 88 perform functions similar to those corresponding to those in the conventional bit rate control apparatus shown in FIG. Since 86) performs substantially the same function except calculating the reference quantization parameter Q using Equation (9) above, the detailed description herein is omitted to avoid overlapping.

First, B _i representing the amount of bits generated in the i-th macroblock is input to the quantization parameter calculator 86 and to the proportional constant calculator 82. In addition, the quantization parameter Q calculated by the quantization parameter calculator 86 is simultaneously input to the activity calculator 88 and the proportional constant calculator 82.

On the other hand, the proportional constant calculation unit 82 inputs the quantization parameter Q output from the quantization parameter calculation unit 86 and the bit generation amount Fi (Qi) of each macro block by the quantization parameter Q as input, and performs a predetermined calculation process. The proportional constant K of the quantization parameter Q_occurrence function is calculated and provided to the quantization parameter calculator 86.

Next, the operation of the proportional constant calculation unit 82, which constitutes the most characteristic part of the bit rate control apparatus of the present invention as described above, will be described in detail with reference to FIG.

3 shows a block diagram of the proportional constant calculation unit 82 for obtaining the proportional constant K of the quantization parameter Q_occurrence function as described above, and as shown in the same figure, the proportional constant calculation according to the present invention. The unit 82 may include a multiplier 822, a delayer 824, and a constant calculator 826.

In FIG. 3, the multiplier 822 uses the quantization parameter Qi of the i-th macroblock output from the quantization parameter calculator 86 shown in FIG. 2 and the amount of bits generated in the i-th macroblock according to the quantization parameter Qi. The proportional constant K ₁ ′ is calculated by multiplying by the function Fi (Qi), and the calculated proportional constant K ₁ ′ is input to the delay unit 824 to be delayed for a predetermined time, and then the constant calculating unit 826 of the next stage. Is provided. The constant K ₁ 'value from multiplier 822 is also provided to constant calculation unit 826.

Accordingly, the constant calculating unit 826 obtains the following equation based on the proportional constant K ₂ ′ of the previous macro block input from the delay unit 824 and the proportional constant K ₁ ′ input from the multiplier 822. The final proportional constant K is calculated.

Here, β is a value for adjusting the weights of the proportional constant K ₁ 'obtained from the quantization parameter Qi and the occurrence rate Fi (Qi) of the i-th macroblock and the proportional constant K ₂ ′ of the previous macroblock.

Accordingly, the constant calculating unit 826 obtains the proportional constant K value of the quantization parameter Q_occurrence function of the macroblock that is finally applicable, and the proportional constant K value thus obtained is the quantization parameter calculator shown in FIG. Provided at 86.

Therefore, the quantization parameter calculation unit 86 calculates and provides the quantization parameter Q for the corresponding macroblock to the activity calculation unit 88 using equations (9) and (10) described above, and activity calculation unit 88. In consideration of the activity according to the quantization parameter Q calculated as described above, i.e., calculating how much activity the macroblock currently being encoded has in the frame, and the scaled value mquant is calculated again. This substantially determines the quantization step size in the quantization unit 30 shown in FIG.

Therefore, by appropriately adjusting the quantization step size of the macroblock to be encoded (quantized) through the operation processing as described above, the bit generation rate of the final output image data is controlled to meet the channel requirements regardless of the image size. can do.

As described above, according to the method for controlling the bit rate of the present invention, regardless of the size of an image (not affected), the bit rate is adaptively controlled at a bit rate corresponding to a target required by a channel, thereby deciphering the receiving side. It is possible to prevent deterioration of image quality of the image restored in the system.

Claims (2)

The quantization parameter is calculated based on the target number of bits of the currently processed picture and the bit generation amount of the quantized and variable length coded macroblocks up to now, and the rescaled calculated according to the activity in the frame of the calculated macroblock. A coded bit rate control method for determining a quantization step size for a macroblock to be quantized based on a value, the method comprising: generating a bit generation amount of each macroblock based on a bit generation amount of the i-th macroblock having been quantized and variable length coded. A proportional constant of a quantization parameter bit rate function is obtained, and the bit is based on the proportional constant value of the i-th macroblock obtained based on the quantization parameter and the calculated bit rate function and the other proportional constant value obtained from a delayed previous macroblock. Final ratio of incidence functions Calculating a constant, calculating the quantization parameter with reference to the final proportional constant, and an occurrence rate function for the quantization parameter is calculated as follows: Here, if substituted with α = (B _i-1 -T '(i-1-N) / N, the quantization parameter is calculated as follows: The proportionality constant of the incidence function is calculated as Wherein B _i-1 is the amount of bits generated so far, T 'is the target number of bits of the macroblock, and N is a constant for adjusting the convergence speed of the controller. Means for calculating a quantization parameter based on the target number of bits of the currently processed picture and the bit generation amount of the quantized and variable-length coded macroblocks so far, and the recalculation calculated according to the activity in the frame of the calculated macroblock. A coded bit rate control apparatus having means for determining a quantization step size for a macroblock to be quantized based on a scaled value, wherein the control apparatus is generated by the quantization parameter value of the i-th macroblock and the quantization parameter. Means for generating a proportional constant of the quantization parameter incidence function by multiplying the bit generation amount up to an i-th macroblock; Buffer means for delaying the proportional constant for the processing time of the macro block and outputting a proportional constant for the previous macro block; And means for generating one new proportional constant based on the proportional constant of the occurrence function provided by the multiplication means and the proportional constant of the previous macro block provided by the buffer means, wherein the quantization parameter calculating means comprises: The quantization parameter for the corresponding macroblock to be quantized is calculated with reference to the new proportional constant, and the new proportional constant is calculated as follows. In the above equation, K ₁ ′ is a proportional constant of the current macroblock, K ₂ ′ is a proportional constant of the previous macro block, and β is a value for adjusting the weight of K ₁ ′ and K ₂ ′. .