KR100283980B1 - Variable output video bit rate control method - Google Patents

Abstract

1. TECHNICAL FIELD OF THE INVENTION

The present invention relates to a variable output video bit rate control method.

2. Technical problem to be solved by the invention

The present invention minimizes overflow or deficiency of a buffer in an encoder while satisfying a variable target output bit rate during video encoding and transmission, and at the same time, makes image quality between consecutive images as uniform as possible.

3. Summary of Solution to Invention

Calculating the bit amount allocated to the encoding of the next image, the allowable range of the quantization step size, and the quantization step size for generating the minimum bit amount among the quantization step sizes for generating a bit amount larger than the calculated allocated bit amount. Selecting a bit amount and a quantization step size and a bit amount for generating a maximum bit amount among the quantization step sizes for generating a bit amount smaller than the calculated allocated bit amount, wherein the respective quantization step sizes and corresponding bits are selected. And selecting a quantization step size close to the quantization step size used for encoding the previous image.

4. Important uses of the invention

Applicable to the bit rate controller.

Description

Variable output video bit rate control method

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an output bit rate control method in video encoding, and more particularly to a variable output video bit rate control method.

In general, the bit rate control apparatus (see FIG. 1) is characterized in that when a video frame is input from the video frame input unit 1 as described in US Patent No. 5,617,150 (1997.4.1), which is granted by the US Patent Office, the encoder 2 ) Encodes the input frame, stores the encoding result encoded by the encoder 2 in the buffer unit 3, and transmits the encoded result value to the outside at a constant transmission bit rate. At this time, in order to adjust the bit rate output from the encoder 2, the bit rate controller 4 adjusts the encoder 2 so that there is no buffer overflow or buffer shortage according to the buffer fullness of the buffer 3.

In the related art, since the state of the buffer in the encoder has not been sufficiently considered when calculating the bit rate allocated to the encoding of the next video using the above-described bit rate controller, the possibility of overflow or underflow of the buffer is very high. When encoding the next image, there is a problem in that the change in image quality between encoded consecutive images is large by using a quantization step size corresponding to the predicted bit rate that is closest to the target output bit rate calculated.

Therefore, in order to minimize the overflow or underflow of the buffer in the encoder while satisfying the variable target output bit rate that may be caused by problems in the transmission path in image encoding and transmission, When calculating the target bit amount to be allocated, sufficient consideration of the current buffer state is necessary, and in order to keep the image quality between consecutive images as uniform as possible, use the quantization step size as close as possible to the quantization step size used for encoding the previous image. It is necessary to encode the next image.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and overflows or deficiency of a buffer in an encoder while satisfying a variable target output bit rate that may be caused by a problem in a transmission path in video encoding and transmission. It is an object of the present invention to provide a variable output video bit rate control method that minimizes (underflow) and maintains the image quality between consecutive images as uniformly as possible.

1 is a block diagram of a bit rate control apparatus to which the present invention is applied.

2 is an overall flowchart of an image encoding process according to the present invention;

3 is a detailed flowchart of the quantization step size calculation process of FIG. 2;

* Description of the symbols for the main parts of the drawings

1: video frame input unit 2: encoder

3: buffer section 4: bit rate control section

According to the present invention for achieving the above object, in the variable output video bit rate control method, the allocation of the next video by using the number of bits generated in the encoding result of the previous video, the current transmission path state, and the current buffer state. A first step of calculating the amount of bits to be made; A second step of calculating an allowable range of the quantization step sizes usable in the next image encoding according to the quantization step sizes among the encoding results of the previous image; A quantization step size and a corresponding bit amount for generating a minimum bit amount among quantization step sizes for generating a bit amount larger than the calculated allocation bit amount within an allowable range of the quantization step size, and the calculated allocation bit amount A third step of selecting a quantization step size for generating a maximum bit amount among the quantization step sizes for generating a smaller bit amount and a corresponding bit amount; And a fourth step of selecting a quantization step size close to the quantization step size used for encoding a previous image, even if the error between the calculated quantization step size and the corresponding bit amount is large to some extent. Is done.

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

2 is a flowchart of the entire process of video encoding according to the present invention.

First, after the target output bit rate Rt and the target frame rate Nf set by initialization are determined (21), the bit amount and quantization step size allocated for encoding the first image are calculated (22). After encoding the image using the quantization step size obtained in the bit quantity and quantization step size calculation process 22, and then returning to the process 22 to calculate the bit amount allocated for encoding of the next image, The quantization step size calculation process 22 and the next video encoding process 23 are repeatedly performed until the encoding of the image is completed 24.

3 is a process flow diagram for the quantization step size calculation method 22 used to encode the next image of FIG.

First, in the case of encoding the first image, the bit rate allocated to the next frame is calculated using the target output bit rate R0 and the target frame rate F0 obtained in the initialization, but in the case of encoding the subsequent nth image, The target output bit rate Rn of Equation 1 and the frame rate Fn of Equation 2 are used.

Rn = Wr * R0

Fn = Wf * F0

In Equation 1, Wr is a weight that varies depending on a state of a transmission path through which coded data is output, and '1' when a data transmission capacity of a transmission path is equal to R0, and a value greater than 1 when R0 is greater than R0, If less than R0, it is given a value less than 1.

Similarly, in [Equation 2], Wf is a weight that varies depending on the state of a transmission path on which coded data is output, and a value greater than 1 when the frame transmission capacity of the transmission path is equal to F0 and larger than F0 when If F0 is less than F0, it is given a value less than 1.

The bit amount Rt allocated to the next frame by using Equations 1 and 2 is calculated as shown in Equation 3 below (31).

Rt = Wt * (Ws * Bsize-Bs) / Fn} + Rn / Fn

(Wt is a weight for adding or decreasing the amount of bits allocated to encoding of the next image, Ws is the ratio of the average bit amount remaining in the buffer to the buffer size, Bsize is the size of the buffer, Bs is remaining in the buffer Current bit amount.)

Next, the minimum and maximum values of the quantization step size that can be used to encode the next image are calculated using the quantization step size (Qp) used when encoding the previous image. In this case, the calculated minimum value (Qm) and maximum value (QM) of the quantization step size may have a different value according to the value of the quantization step size (Qp), an example of a function that can be used is shown in Equation 4 below. (32).

[Equation 4] If Qp is less than or equal to 2

Qm = 1;

If Qp is less than or equal to 2 and is 3

Qm = 2;

Qp is less than or equal to 2, not 3, and less than or equal to 5

Qm = 3;

Qp is less than or equal to 2, not 3, less than or equal to 5, and less than or equal to 6

Qm = 4;

Qp is less than or equal to 2, not 3, less than or equal to 5, less than or equal to 6, less than or equal to 7

Qm = 5;

Qp is less than or equal to 2, not 3, less than or equal to 5, less than or equal to 6, less than or equal to 7 and less than or equal to 9

Qm = 6;

And Qm = maximum value ((Qp-5), 7)

If Qp is 1, QM = 2;

If Qp is not 1 and 2, then QM = 3;

QM = 5 if Qp is not 1 or 2 and 3;

QM = 7 if Qp is not 1 or 2 or 3, and 4;

QM = 10 if Qp is not 1 or 2 or 3 or 4, and 5;

And if Qp is not one of 1,2,3,4,5,

QM = minimum value ((Qp + 7), 31);

Generates the smallest bit amount among the quantization step sizes that generates a bit amount larger than Rt of [Equation 3] between the minimum and maximum values of the quantization step size calculated in the allowable range calculation process 32 of the quantization step size. A quantization step that generates the largest bit amount among the quantization step sizes to be generated and the corresponding bit amounts Q (1, R (1)) and quantization step sizes that generate a bit amount smaller than the allocated bit amount obtained in the above process. The size and the corresponding bit amounts Q (2) and R (2) are selected.

In this case, the bit amount generated by the arbitrary quantization step size may be represented by an arbitrary function as shown in Equation 5, and an example of a function that may be used is illustrated in Equation 6 (33).

R (i) = f (Q (i))

R (i) = p1 / sqrt (Q (i)) + p2 / Q (i) + p3

For reference, the parameters used in [Equation 5] or [Equation 6] are used to calculate the quantization step size (Q (i)) obtained from the encoding of the previous image and the corresponding bit amount (R (i)). Calculated using In this case, the parameters may be obtained by using less data than the number of parameters among the previous quantization step size and the corresponding bit amounts. In this case, the estimated quantization step size and Corresponding bit amount values are used.

This method provides the advantage that it is possible to quickly adapt to the rapid change of video encoding complexity such as scene transition by applying to the next video encoding considering only the encoding characteristic of the most recent video.

One of (Q (1), R (1)) and (Q (2), R (2)) obtained in the step 33 of selecting the candidate quantization step size and the expected number of bits is selected to display the next image. It is used as a quantization step size Q for encoding, in which Equation 7 is used (34).

(ABS (Q (1) -Qp) is less than ABS (Q (2) -Qp)),

((R (1)-Rt) is less than or equal to Delta * (Rt-R (2)),

Q = Q (1);

(ABS (Q (1) -Qp) is less than ABS (Q (2) -Qp)),

((R (1)-Rt) is less than or equal to Delta * (Rt-R (2)),

Q = Q (2);

(ABS (Q (1) -Qp) is not less than ABS (Q (2) -Qp))

If ((Rt-R (2)) is less than or equal to Delta * (R (1)-Rt),

Q = Q (2);

(ABS (Q (1) -Qp) is not less than ABS (Q (2) -Qp)),

If ((Rt-R (2)) is less than or equal to Delta * (R (1)-Rt),

Q = Q (1);

In this case, since the image quality of the image and the quantization step size are directly related to each other, when the [Equation 7] is used to determine the quantization step size for encoding the next image, the change in the image quality that may exist between consecutive images is minimized. There is an advantage.

The above-described present invention divides the entire interval of the buffer in the encoder into N partial intervals and multiplies different weights according to the intervals to which the current buffer state belongs to calculate the target bit amount allocated for encoding the next frame. In this case, the multiplied weight may be increased as the interval to which the current buffer state belongs becomes farther from any reference position, thereby minimizing overflow or underflow of the buffer.

And, in determining the quantization step size that satisfies the target output bit amount, the largest value among the values smaller than the target bit amount (RT) allocated for encoding the next image in order to maintain the image quality between consecutive images as uniformly as possible. After obtaining the smallest prediction bit amount R2 and the quantization step size Q2 used at this time among the prediction bit amount R1 and the quantization step size Q1 and the values larger than RT, Q2 is immediately before Q1. If it is close to the quantization step size used for encoding the image, the next image is encoded using Q2, even if (R2-RT) is somewhat larger than (RT-R1), and Q1 is used for encoding the previous image before Q2. If it is close to the quantization step size used, even if (RT-R1) is somewhat larger than (R2-RT), the next image is encoded using Q1. This minimizes the change in image quality between successive images.

The present invention described above is not limited to the above-described embodiments and the accompanying drawings, and various substitutions, modifications, and changes can be made in the art without departing from the technical spirit of the present invention. It will be apparent to those of ordinary knowledge.

The present invention as described above, in order to minimize the overflow or underflow of the buffer in the encoder while satisfying the variable target output bit rate that can be changed from time to time due to the characteristics on the transmission path in video encoding and transmission At the same time, the image quality between successive images can be maintained as uniformly as possible, thereby eliminating the need for additional calculations or scene calls required for scene change required in the prior art.

Claims (5)

In the variable output video bit rate control method, A first step of calculating the amount of bits allocated for encoding the next image using the number of bits generated in the encoding result of the previous image, the state of the current transmission path, and the state of the current buffer; A second step of calculating an allowable range of the quantization step sizes usable in the next image encoding according to the quantization step sizes among the encoding results of the previous image; A quantization step size and a corresponding bit amount for generating a minimum bit amount among quantization step sizes for generating a bit amount larger than the calculated allocation bit amount within an allowable range of the quantization step size, and the calculated allocation bit amount A third step of selecting a quantization step size for generating a maximum bit amount among the quantization step sizes for generating a smaller bit amount and a corresponding bit amount; And And a fourth step of selecting a quantization step size close to the quantization step size used for encoding a previous image regardless of the magnitude of the error from the calculated allocation bit amount among the respective quantization step sizes and corresponding bit amounts. Variable output video bit rate control method. The method of claim 1, The first step is, And a bit amount Rt allocated to the encoding of the next image is calculated by Equation 1, Equation 2 and Equation 3 below. Rn = Wr * R0 Fn = Wf * F0 (Wr is a weight that varies depending on the state of the transmission path on which the encoded data is output, Wf is a weight that varies depending on the state of the transmission path on which the encoded data is output, Rn is a target output bit rate, and Fn is a frame rate.) Rt = Wt * (Ws * Bsize-Bs) / Fn} + Rn / Fn (Wt is a weight for adding or decreasing the amount of bits allocated to the encoding of the next image, Ws is the ratio of the average bit amount remaining in the buffer to the buffer size, Bsize is the size of the buffer, Bs is remaining in the buffer Current bit amount.) The method of claim 1, The second step, When calculating the allowable range of the quantization step size that can be used for encoding the next image, the allowable range is small when the value of the quantization step size used for the encoding of the previous image is small, and when the large value is large, the allowable range is large. Variable output video bit rate control method. The method of claim 1, The third step, And a bit rate Ri generated when an image block is encoded using an arbitrary value Q according to Equation 6 below. R (i) = p1 / sqrt (Q (i)) + p2 / Q (i) + p3 (Where, Q (i) is the quantization step size obtained from the result of encoding the previous image, R (i) is the generated bit amount corresponding to Q (i), and p1, p2, and p3 are the quantization steps of the previously encoded image). Are parameters derived from the amount of occurrence bits corresponding to the sizes.) The method of claim 1, The fourth step, And calculating by Equation (7) to select one of a plurality of quantization step sizes and a corresponding amount of generated bits to encode a next image block. (ABS (Q (1) -Qp) is less than ABS (Q (2) -Qp)), ((R (1)-Rt) is less than or equal to Delta * (Rt-R (2)), Q = Q (1); (ABS (Q (1) -Qp) is less than ABS (Q (2) -Qp)), ((R (1)-Rt) is less than or equal to Delta * (Rt-R (2)), Q = Q (2); (ABS (Q (1) -Qp) is not less than ABS (Q (2) -Qp)) If ((Rt-R (2)) is less than or equal to Delta * (R (1)-Rt), Q = Q (2); (ABS (Q (1) -Qp) is not less than ABS (Q (2) -Qp)), If ((Rt-R (2)) is less than or equal to Delta * (R (1)-Rt), Q = Q (1); (Where, Q (i) is the quantization step size obtained from the result of encoding the previous image, R (i) is the generated bit amount corresponding to Q (i), and p1, p2, and p3 are the quantization steps of the previously encoded image). A parameter obtained from the amount of generated bits corresponding to the sizes.