JPH0479670A - Encoding device - Google Patents

Abstract

PURPOSE:To reduce the data quantity, with the picture quality hardly spoiled, by setting the quantizing step of a requantizing device larger so as to increase the compression ratio when the dynamic range is large. CONSTITUTION:A dynamic range calculator 6 calculates the dynamic range 7 of an input signal 1. When the dynamic range 7 is large, a re-quantizing device 8 sets the quantizing step larger so as to reduce the data quantity of an output signal 5 encoded by means of an encoder 4. The encoder 4 is a variable-length encoder and, when the encoder 4 is used with an orthogonal transducer 2, the data quantity can be reduced by hardly spoiling the picture quality. Therefore, by providing the dynamic range calculator and re-quantizing device 8 and changing the quantizing step on the basis of the dynamic range, the data quantity can be reduced by hardly spoiling the picture quality.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention is a means for reducing the amount of data required to transmit and store image signals in order to reduce communication costs and extend recording time. The present invention relates to an encoding device.

Conventional technology A block diagram of a conventional encoding device is shown in FIG.

In the figure, 1 is an input signal, 2 is an orthogonal transformer that orthogonally transforms input signal 1, 3 is a signal that has been orthogonally transformed by orthogonal transformer 2, 4 is an encoder that encodes signal 3, and 5 is a code Transformer 4
This is the output signal encoded by .

In the conventional encoding device configured as described above,
Input signal 1 is orthogonally transformed by orthogonal transformer 2, and encoder 4
encoded in By choosing an appropriate function for this orthogonal transformation, it is possible to make certain transformation components smaller than other components, and as proven in information theory, signals with strong correlations such as image signals can be significantly reduced. It is possible to reduce the amount of data. Furthermore, when DCT (discrete cosine transform) is used as an orthogonal transform, distortion for high frequency components is less visually noticeable than distortion for low frequency components, and this can be used to further reduce the amount of data without compromising image quality. can be reduced.

Problems to be Solved by the Invention However, in the configuration described above, the visual characteristics of the image are not used, and there is still room for reducing the amount of data.

In view of this, an object of the present invention is to provide an encoding device that further reduces the amount of data compared to conventional encoding devices without substantially degrading image quality.

Means for Solving the Problems The present invention provides a dynamic range calculating means for calculating a dynamic range or a value corresponding to the dynamic range for each block of a digitized input signal, and a dynamic range calculated by the dynamic range calculating means. quantization means for selecting one from several types of requantizers and quantizing the input signal with the requantizer; and encoding means for encoding the quantized value quantized by the quantization means. This is an encoding device characterized by comprising:

According to the above-described configuration, the present invention calculates the dynamic range of the input signal using the dynamic range calculation means, and when the dynamic range is large, the quantization step of the requantizer is increased to increase the compression ratio. (reduce the amount of data).

In general, when the dynamic range is large, there is little visual deterioration in image quality even if encoding distortion increases slightly, so by using the above configuration, the same level of playback image quality can be achieved with less data than a conventional encoding device. Obtainable.

Embodiment FIG. 1 is a block diagram of an encoding apparatus in a first embodiment of the present invention. In the same figure, 1 is an input signal, 2 is an orthogonal transformer that outputs an orthogonal transformed signal 3 obtained by orthogonally transforming the input signal 1, 6 is a dynamic range calculator that calculates the dynamic range 7 of the input signal, and 8 is an orthogonal transformer. A requantizer 4 requantizes the signal 3 with a quantization step corresponding to a dynamic range 7, and a coder 4 encodes the requantized value 9 and outputs an output signal 5.

In the encoding device of this embodiment configured as described above,
The operation will be explained below. A dynamic range 7 is calculated for the input signal 1 by a dynamic range calculator 6. This dynamic range 7 may be calculated using the same block as the orthogonal transform, or may be selected for several orthogonal transform blocks or independently of the orthogonal transform blocks. When the dynamic range 7 is large, the requantizer 8 increases the quantization step, and the encoder 4 encodes the output signal 5.
reduce the amount of data. This encoder 4 is a variable length encoder, and it is known that when used in conjunction with orthogonal transformation, the amount of data can be reduced with almost no loss in image quality.

Next, control of the quantization step using the dynamic range will be explained. Figure 2 is an example of an image, and the dynamic range inside the circle is small. It is assumed that the inside of the rectangle has a large dynamic range. The solid line in FIG. 3(a) represents the original signal inside the circle, and the solid line in FIG. 3(b) represents the original signal inside the rectangle. The horizontal axis indicates pixel position, and the vertical axis indicates pixel value. For each, the dotted line indicates the encoded data.

Generally, sign distortion is additively added to the signal, so the signal shown in Fig. 3(b) has a larger dynamic range than the signal shown in Fig. 3(b).
The SN ratio of the signal a) deteriorates. Visually, grainy noise is noticeable in areas with a small dynamic range, while it is hardly noticeable in areas with a large dynamic range. Therefore, by increasing the compression rate when the dynamic range is large, it is possible to reduce the amount of data with almost no loss in image quality.

As described above, according to this embodiment, by providing the dynamic range calculator 6 and the requantizer 8 and changing the quantization step depending on the dynamic range, it is possible to reduce the amount of data with almost no deterioration in image quality. .

FIG. 4 is a block diagram showing in more detail the structure of the dynamic range calculator and requantizer of the present invention. In the figure, 10 is a digitized input signal, 11 is a maximum value calculator that calculates the maximum value 12, 13 is a minimum value calculator that calculates the minimum value 14, 15 is a subtracter that calculates the difference 16, and 17 2o is a ROM (read only memory) that outputs the divisor 18, and 2o is a divider that calculates the quotient 21 obtained by dividing the orthogonally transformed input signal 19 by the divisor 18.

The operation of the second embodiment configured as above will be described below. Maximum value calculator 11, minimum value calculator 1
3 and the subtractor 5 constitute the dynamic range calculator 6 of the first embodiment, and the first
This constitutes the requantizer 8 of the embodiment. input signal 10
The maximum value 12 and minimum value 14 are calculated from the maximum value 1
The minimum value 14 is subtracted from 2, and the difference 16 becomes the dynamic range. A quantization step corresponding to this dynamic range is stored in the ROM 17, and this ROM
Requantization is performed by dividing the input signal 19 by the divisor 18, which is the output of the input signal 17, and rounding the result to the nearest whole number (or rounding to the nearest 5 digits). When the dynamic range is large, the divisor 18 is made large, and when the dynamic range is small, the divisor 18 is made small, so that the amount of data can be reduced with almost no loss in image quality.

As explained above, according to this embodiment, the maximum value calculator 11, the minimum value calculator 13, and the subtracter 15 calculate the dynamic range, read out the divisor from the calculated value in the ROM, and divide by the divisor. The requantizer of the first embodiment can be constructed depending on the situation.

Note that although the encoding means using orthogonal transformation has been described in the first and second embodiments, encoding means other than orthogonal transformation such as predictive encoding may be used. Furthermore, the requantizer may be controlled using other parameters representing amplitude fluctuations without using the dynamic range. Furthermore, the quantization step of the requantization of the present invention may be encoded together with the quantization value,
If this quantization step can be obtained by some method in the decoding device, it is not necessary to encode it.

Further, in the second embodiment, a linear quantizer was used as the requantizer, but a nonlinear quantizer may also be used.

As described in detail, according to the present invention, it is possible to reduce the amount of data with almost no loss in image quality, and its practical effects are great.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of an encoding device according to the present invention, FIGS. 2 and 3 are explanatory diagrams of the encoding method of the present invention, and FIG. 4 is an encoding device according to a second embodiment of the present invention. FIG. 5 is a block diagram of a conventional encoding device. 6...Dynamic range calculator, 8...Song...
Requantizer. Name of agent: Patent attorney Shigetaka Awano (1 person) Figure 1 Figure 2 Figure 2 Figure Figure 2

Claims (2)

[Claims] (1) A dynamic range calculation means for calculating a dynamic range for each block of a digitized input signal or a value corresponding to the dynamic range, and several types of requantizers based on the dynamic range calculated by the dynamic range calculation means. quantization means for selecting one and quantizing the input signal with the requantizer; and encoding means for encoding the quantized value quantized by the quantization means. An encoding device that does (2) The encoding device according to claim (1), further comprising quantization means that increases the quantization step width when the dynamic range is large.