JP3083595B2 - Code amount control device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a code amount control device for controlling a code amount when a moving picture is encoded and transmitted using an ATM switch or the like.

[0002]

2. Description of the Related Art In general, in moving image transmission, an input image is divided into image blocks of, for example, 8 × 8 pixels, and each image block is subjected to an orthogonal transformation such as DCT, and then quantized to a predetermined step size. This is encoded and transmitted to the partner station.

At this time, since the data amount of the input image is not constant, if the quantization step size is determined to be constant, the code amount fluctuates according to the data amount. Further, if the code amount is too large, there is a drawback that image frames are dropped, and if it is too small, there is a drawback that transmission efficiency is reduced. Therefore, it is necessary to keep the code amount as constant as possible. Is done.

Therefore, conventionally, a code amount control device that appropriately changes the quantization step size while monitoring the code amount has been used. For example, the document “Rate control method in packet video coding” (PCSJ90: Picture Codi)
ng Symposium Japan 1990, 9-5, Takishima / Wada) is known as a known technique.

According to this document, a reference code amount is determined in advance, an excess code amount with respect to the reference code amount in each image frame is obtained, and the accumulated code amount up to the current frame becomes larger than a predetermined limit value. In this case, a method is described in which the code size is reduced by increasing the step size.

[0006]

However, in such a conventional code amount control apparatus, when the accumulated excess code amount is larger than the limit value, if the current code amount is small and this state is maintained, the accumulated excess code amount is maintained. Can be processed, the operation is performed in the direction of increasing the step size and decreasing the code amount. For this reason, there is a disadvantage that the image quality is unavoidably deteriorated. In addition, the step size frequently fluctuates due to a temporary increase or decrease in the code amount, so that there is a problem that the variation in image quality between frames becomes large.

SUMMARY OF THE INVENTION The present invention has been made to solve such a conventional problem. It is an object of the present invention to control code amount for preventing image quality variation and image quality deterioration due to unnecessary step size change. It is to provide a device.

[0008]

In order to achieve the above object, the present invention provides an apparatus for quantizing an input signal for each predetermined frame and outputting encoded code data by appropriately changing a quantization step size. A code amount control device for controlling the amount of data to be encoded, wherein the code amount monitoring means monitors the amount of the code data generated for each frame; An excess code amount calculating means for calculating an excess code amount in the current frame from a difference from the code data amount; a cumulative excess code amount calculating means for sequentially adding the excess code amount up to the current frame to obtain a cumulative excess code amount; Correction means for correcting the reference code amount of the next frame based on the reference code amount of the frame and the excess code amount of the current frame, wherein the accumulated excess code amount is greater than a first predetermined value. Increasing the quantization step size when large, when the second smaller than a predetermined value is characterized in that a means for reducing the quantization step size.

[0009]

According to the present invention having the above-described configuration, when it is determined that the input signal cannot be processed with the current quantization step size when the input signal is quantized for each predetermined frame and coded data is output. That is, since the step size is changed only when the accumulated excess code amount is not within the range defined by the first and second predetermined values, unnecessary increase and decrease of the step size can be prevented. In particular, in the present invention, the reference code amount for obtaining the accumulated excess code amount of the next frame is corrected based on the excess code amount of the current frame. The accumulated excess code amount also changes, and the quantization step size can be changed promptly.

[0010]

FIG. 1 is a block diagram showing the configuration of an embodiment of a moving picture transmission apparatus including a code amount control apparatus to which the present invention is applied.

As shown in the figure, the moving picture transmission apparatus includes a block dividing circuit 1 for dividing an input image into image blocks of, for example, 8 × 8 pixels, and an orthogonal transform for performing an orthogonal transform such as a discrete cosine transform on each block. A circuit 2, a quantization circuit 3 for quantizing the output to a predetermined step size,
A variable length coding circuit 4 for performing variable length coding on the quantized data; a code for monitoring the current code amount and the accumulated excess code amount up to the present, and instructing an increase or decrease of the step size based on a predetermined condition. It comprises an amount monitoring circuit 6 and a step size determination circuit 5 for determining a quantization step size based on the instruction to increase or decrease the step size.

Next, the operation of the code amount monitoring circuit 6, which is a main part of the present invention, will be described with reference to the flowchart shown in FIG. In the figure, n is a frame number, G
_n is the step size for quantizing the n-th frame, D
The cumulative abnormal code amount, r is the reference code amount of one frame, s _n
Is the code amount when the n-th frame is coded, α is the limit value of the accumulated excess code amount, and β is the limit value of the excess code amount in one frame.

First, as an initial setting, a frame number n =
0, the accumulated excess code amount D = 0, and the step size of the first frame is set as an initial value (step ST1). When the first frame of image data is given (step ST2), reads the code amount s ₁ of the first frame from the variable length coding circuit 4 shown in FIG. 1 (step ST3), the reference code amount from the s ₁ The excess code amount d of one frame is obtained by subtracting r (however, d may be negative, and in this case, indicates an insufficient code amount). Further, the accumulated excess code amount D (however, at this time, zero) is added to the excess code amount d to obtain a new accumulated excess code amount D (step ST4).

Thereafter, it is determined whether or not the code amount D is larger than a limit value α (step ST5). When D <α (NO in step ST5), it is determined whether or not the code amount D is larger than -α. A determination is made (step ST9). When D> -α (NO in step ST9), the current accumulated excess code amount D is within the range of the limit value ± α, so the step size is left as it is. That is, the step size G ₂ of the second frame is a next image frame, and G ₂ = G ₁ (step ST13). Then, it outputs to the step size determining circuit 5 indicating the step size G ₂ in FIG. 1 (step ST14).

On the other hand, if code amount D is larger than limit value α (YES in step ST5), it is determined whether excess code amount d is larger than limit value β (step ST5).
6). When d> β (YE in step ST6)
In S), the accumulated excess code amount D is large, and the excess code amount d of the current frame is also large.
Increase the step size to reduce. That is, the step size G ₂ in the next second frame is set to G
₁ + ΔG (ΔG is an increment) (step ST8).
When d> β (NO in step ST6), the accumulated excess code amount D is large but the current excess code amount d is small, so it is determined that the accumulated amount can be processed without changing the step size. , G ₂ = G ₁ (step ST
7).

D <α (N in step ST5)
O), when it is determined that D <−α (step ST9)
YES), the excess code amount d is compared with the limit value -β (step ST10). When d <-β (YES in step ST10), the accumulated excess code amount D is large in the negative direction (that is, strictly not excess but insufficient), and the code amount of the current frame is small. Since the shortage is large, the step size is reduced. That is, G ₂ = G ₁
−ΔG is set (step ST12). On the other hand, when d> −β (NO in step ST10), although the accumulated excess code amount D is large in the negative direction, the current excess code amount d is large, so that the accumulated excess code amount can be obtained without changing the step size. D decides to proceed in the increasing direction and does not change the step size. That is, G ₂ = G ₁ (step ST1)
1).

Then, the above operation is repeated to always obtain a suitable step size.

As described above, in this embodiment, the change of the step size is determined based on the accumulated excess code amount D up to the present and the excess code amount d of the current frame. Therefore, quantization with the optimal step size according to the code amount can be performed, and unnecessary change in the step size can be avoided.

FIG. 3 is a flowchart showing the operation of another embodiment. In the figure, n is a frame number,
Gn is a step size when quantizing the n-th frame,
r is a reference code amount of one frame, Rn is a virtual reference code amount in the n-th frame, Sn is a code amount when the n-th frame is coded, and β is a predetermined limit value.

First, initialization is performed with n = 0, R ₁ = r, G ₁ = initial value (step ST21). And
When the first frame of the image data is provided (step S
T22), reads the code amount S ₁ of the first frame from the variable length coding circuit 4 shown in FIG. 1 (step ST23).

Then, the accumulated excess code amount h is calculated by the following equation (1).
Is obtained (step ST24).

[0022]

(Equation 1)

Here, m is the number of past frames referring to the excess code amount. That is, the excess code amount h indicates a cumulative value of the excess code amount from m frames before.

In order to apply the expression (1), it is necessary that at least n> m + 1. Accordingly, the frames up to the m-th frame are quantized by the initially set constant step size G ₁ , and hereinafter, the frames where n> m + 1 will be described.

The accumulated excess code amount h obtained by the equation (1)
Is first compared with a limit value β (step ST25),
When h <β (NO in step ST25), h and −β
Are compared (step ST27). And h>-
If β (NO in step ST27), it is determined that the accumulated value is within the range of the limit value, and the step size is not changed. That is, G _{n + 1} = G _n is set (step ST29).

On the other hand, when h> β (YES in step ST25), since the accumulated value is large, the step size is increased to process this. That is, G _{n + 1} = G
_n + △ G (step ST26).

When h <-β (step ST27)
In (YES), the step size is reduced because the shortage of the code amount is large. _{Soku, G n + 1 = G n} - and △ G (step ST28). Then, G _{n + 1} determined in each step is output as the step size of the next frame (step ST30).

Further, a virtual reference code amount R _{n + 1} of the next frame is obtained by the following equation (2) (step ST31).

R _{n + 1} = R _n − (s _n −r) / γ (2) The second term on the right side of the equation (2) is a numerical value having a magnitude proportional to the excess code amount of the current frame. The larger this is, the smaller the virtual reference code amount R _{n + 1} of the next frame is. Therefore,
For example, when the amount of excess code with respect to the reference code amount r is the same in the current frame and the next frame, the amount of excess code with respect to the virtual reference code amount is larger in the next frame.

That is, as the excess code amount of the current frame increases, the cumulative value h increases at an accelerated rate.
The condition h> β of T25 is more likely to be satisfied. On the other hand, the cumulative value h increases in the negative direction at an accelerated rate as the amount of insufficient code of the current frame increases.
That is, the condition h <−β of T27 is easily satisfied.

Thus, in other embodiments,
The step size can be changed based on the excess code amount of the current frame and the excess code amount accumulated up to the present, and a suitable step size can be obtained.

[0032]

As described above, in the present invention, the quantization step size is determined based on the excess code amount of the current frame and the accumulated excess code amount up to the present. Therefore, the step size is not changed unnecessarily unlike the conventional case, and the effect of preventing image quality deterioration and image quality variation can be obtained.

[Brief description of the drawings]

FIG. 1 is a configuration diagram showing one embodiment of the present invention.

FIG. 2 is a flowchart illustrating the operation of the present embodiment.

FIG. 3 is a flowchart showing an operation of another embodiment.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Block division circuit 2 Orthogonal transformation circuit 3 Quantization circuit 4 Variable length encoding circuit 5 Step size determination circuit 6 Code amount monitoring circuit

Continuation of front page (56) References JP-A-2-194734 (JP, A) JP-A-2-270490 (JP, A) JP-A-56-48737 (JP, A) JP-A-3-10486 (JP, A) JP-A-3-140074 (JP, A) JP-A-3-49475 (JP, A) JP-A-3-66281 (JP, A) JP-A-3-113979 (JP, A) 3-263927 (JP, A) (58) Field surveyed (Int. Cl. ⁷ , DB name) H03M 7/30

Claims (1)

(57) [Claims] An input signal is quantized every predetermined frame.
In a device for outputting encoded code data, the quantization step size is appropriately changed to obtain the encoded data.
A code amount control device for controlling the amount of data, monitoring the amount of the code data generated for each frame
Code amount monitoring means, and whether the difference between the reference code amount and the code data amount for each frame is
Excess code amount to calculate the excess code amount in the current frame from
The arithmetic means and the excess code amount up to the current frame are sequentially added to obtain the cumulative excess code.
Means for calculating the amount of accumulated excess code for determining the amount of code
The reference code amount of the next frame is corrected based on the over code amount.
A correction unit that, the when the accumulated excess code amount is greater than a first predetermined value to increase the quantization step size, and means when the second is smaller than the predetermined value to reduce the quantization step size A code amount control device, comprising: