JP2523876B2 - Video coding device - Google Patents

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a moving picture coding device used in a videophone, a video conference system, or the like.

2. Description of the Related Art FIG. 5 shows a schematic configuration of a conventional moving picture coding device.

In FIG. 5, the quantizer 50 quantizes the orthogonal transform coefficient with the quantization step width 54 sent from the transmission buffer 52, and the encoder 51 encodes the quantized representative value of the quantizer 50, The encoded data is output to the transmission buffer 52.

From the transmission buffer 52, the encoded data is output to the line 53 while matching the transmission speed of the transmission path, and when the occupied amount of the encoded data accumulated in the transmission buffer 52 is large, the quantization of the quantizer 50 is performed. The quantization step width 54 is increased to roughen the quantization characteristic, while when the occupation amount is small, the quantization step width 54 of the quantizer 50 is reduced to increase the quantization characteristic.

Therefore, the quantization step width of the quantizer 50 is changed according to the occupied amount of the encoded data accumulated in the transmission buffer 52.
Since 54 is controlled, it is possible to prevent the encoded data accumulated in the transmission buffer 52 from overflowing or underflowing.

However, in the above-mentioned conventional moving image encoding device, in order to prevent the encoded data accumulated in the transmission buffer 52 from overflowing or underflowing,
Since the quantization step width of the quantizer 50 is controlled, there is a problem that the reproduced image quality cannot be improved by this control.

The present invention has been made in view of the above-mentioned conventional problems, and an object of the present invention is to provide a moving picture coding apparatus capable of improving image quality.

Means for Solving the Problems In order to achieve the above object, the present invention quantizes a coefficient in which a prediction error of a moving image signal is orthogonally transformed for each block, and a current quantized by the quantizing means. A means for controlling the quantization step width of the quantization means is provided according to the average generated information amount that straddles the block of the processing frame and the block of the preprocessing frame, and the quantization step width control means is When the average generated information amount is large, the quantization step width of the quantizing means is increased, and when the average generated information amount is small, the quantization step width of the quantizing means is decreased.

Operation According to the present invention, the quantization step width control means increases the quantization step width of the quantization means when the average generated information amount is large, and the quantization step width of the quantization means when the average generated information amount is small. Since the conversion step width is reduced, the amount of information generated in the past block depends on the property of the image, so that the image quality can be improved.

Embodiments Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of a moving picture coding apparatus according to the present invention, FIG. 2 is an explanatory diagram showing a method of calculating a past generated information amount, and FIG. 3 is a past generated information. Explanatory drawing which shows the characteristic at the time of determining a quantization step width by quantity, 4th
The figure is an explanatory view showing the characteristics of the quantizer of FIG.

In FIG. 1, 10 is a quantizer that quantizes the orthogonal transform coefficient with a step width g as shown in FIG. 4, and the orthogonal transform coefficient uses predictive coding that utilizes the correlation between frames of an image. It is obtained by orthogonally transforming the obtained prediction error signal for each block of M pixels × N lines.

Reference numeral 11 denotes an encoder that encodes the quantized representative value of the quantizer 10, and 12 temporarily stores the encoded data from the encoder 11, and encodes the encoded data while matching the transmission speed of the transmission path. A transmission buffer that outputs to the line 13 and adaptively controls the target value of the quantization step width of the quantizer 10, 14 is
This is an information amount calculation circuit that calculates the amount of information generated in the past based on the encoded data from the encoder 11 and controls the quantization step width of the quantizer 10 according to this amount of information.

Next, the operation of the above embodiment will be described with reference to FIGS.

FIG. 2 shows a case where one frame is divided into J blocks GOB and the quantization step width is controlled.

In FIG. 2, 20 indicates the current processing frame, and 21 indicates
A pre-processing frame is shown. Further, 22 indicates a block GOB of the current processing frame 20 to be encoded, and 23 indicates a block for one frame processed in the past from the block 22.
Indicates GOB.

Therefore, the information amount calculation circuit 14 uses a block for one frame that straddles the current processing frame 20 and the preprocessing frame 21.
An average generated information amount of 23 is calculated to control the quantization step width of the quantizer 10.

Here, FIG. 3 shows the generated information amount and the quantization step width g.
B shows the expected value of the amount of information generated per block, and A shows the target value of the quantization step width.

In FIG. 3, assuming that the transmission rate of the line is K (bit / sec) and the frame rate is L (Hz), the expected value B can be expressed by B = K / (L × J), which is a linear characteristic. is there.

Further, as indicated by the broken line in FIG. 3, the transmission buffer 12 increases the quantization step width target value of the quantizer 10 when the occupied amount of the coded data accumulated in the transmission buffer 12 is large (A → A1) to roughen the quantization characteristic, while when the occupation amount is small, the quantization step width target value is reduced (A → A2) to make the quantization characteristic dense.

Therefore, according to the above embodiment, the current processing frame 20
Since the quantization step width is controlled by using the image characteristic of the average generated information amount of the block 23 that straddles the pre-processing frame 21, the image quality can be improved, and the data is stored in the transmission buffer 12. Since the quantization step width of the quantizer 10 is adaptively controlled according to the occupied amount of the coded data being stored, it is possible to prevent the coded data stored in the transmission buffer 12 from overflowing or underflowing. Can be prevented.

In the above embodiment, as the amount of information generated in the past,
1 across the current processing frame 20 and pre-processing frame 21
Although the average amount of generated information of the block 23 for the frame is used,
The present invention is not limited to this information, and the quantization step width is determined by the linear characteristic shown in FIG. 3, but the present invention is not limited to this characteristic.

As described above, the present invention increases the quantization step width of the quantization means when the quantization step width control means has a large average generated information amount, and increases the quantization step width of the quantization means when the average generated information amount is small. Since the quantization step width of the quantizing means is made small, the amount of information generated in the past block depends on the property of the image, so that the image quality can be improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a moving picture coding apparatus according to the present invention, FIG. 2 is an explanatory diagram showing a method of calculating a past generated information amount, and FIG. 3 is a past generated information. FIG. 4 is an explanatory view showing the characteristics when the quantization step width is determined by the amount, FIG. 4 is an explanatory view showing the characteristics of the quantizer of FIG. 1, and FIG. 5 is a conventional moving picture coding apparatus. It is a block diagram. 10 ... Quantizer, 11 ... Encoder, 12 ... Transmission buffer, 13 ... Line, 14 ... Information amount calculation circuit.

Claims (1)

(57) [Claims] 1. A means for quantizing a coefficient obtained by orthogonally transforming a prediction error of a moving image signal for each block, and a block for a current processing frame and a block for a preprocessing frame quantized by the quantizing means. Means for controlling the quantization step width of the quantizing means according to the average generated information amount, and the quantization step width control means sets the quantization step width of the quantizing means when the average generated information amount is large. A moving picture coding apparatus, wherein the moving picture coding apparatus is increased and the quantization step width of the quantization means is reduced when the average generated information amount is small.