JPS633529A - Variable length quantization adaptive predicating dpcm transmission equipment - Google Patents

Abstract

PURPOSE:To minimize the generated information quantity even if the characteristic or property of an input signal is changed by providing a variable length encoder to minimize the generated information equantity depending the frequency of production of a predicating quantized error at every predication system. CONSTITUTION:A switch SW2 selects a predication device 18A by using a prediction system selecting signal at outputted from an optimum predicting selection controller 20 and when a switch SW3 selects a variable length coder 14A in such case, the variable length coding suited to frame prediction is executed and an picture moving rapidly is subjected to DPCM transmission with minimum information quantity. If the switch SW2 selects a prediction device 18B by the signal at outputted from the controller 20 conversely and the switch SW3 selects a variable length coder 14B, the coder 14B executes the variable length coding suited to the inter-frame prediction and a complicated picture with less moving is subjected to DPCM transmission with minimum information quantity.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a DPC~1 transmission device, and more specifically, a method in which a prediction method in DPCv1 encoding is adaptively controlled so as to become an optimal prediction method. The present invention relates to a variable-length quantization-applied predictive DPCM transmission device that variable-length encodes and transmits the quantization prediction error so as to minimize the amount of generated information.

BACKGROUND ART For example, when transmitting an image signal using DPCM, prediction methods include intra-frame prediction, inter-field prediction, and inter-frame prediction. There is an adaptive predictive DPCM transmission method that selects these prediction methods as needed so that the prediction error is minimized. - On the other hand, when DPCM quantization prediction errors are encoded and sent out through a transmission path, each quantization prediction error Q
Since the statistical probability of occurrence differs for each L (Q+, q2, ..., ql, ...q,), a short code is assigned to a quantum prediction error that occurs frequently, and a short code is assigned to a quantum prediction error that occurs less frequently. Long codes are assigned to each error, and variable-length coding is often performed to minimize the amount of information generated.

FIG. 2 shows the configuration of an adaptive predictive DPCMPCM transmission apparatus for image signals using such a variable length code.

Digital image signal Xt is applied to the plus input of adder 10. The adder 10 outputs the difference between the signals applied to the plus input and the minus input, that is, the prediction error e, to two quantizers 12A and 12B. The quantized prediction error (it) quantized by the quantizers 12A and 12B is input to the variable length encoder 14 via the 11 selection switch SW1. The quantization prediction error q is variable-length coded so that the amount of generated information is minimized depending on the frequency of occurrence of each residue.

The output end of the selection switch SW1 is further connected to the plus input of an adder 16, and the output of the adder 16 is input to two predictors 18A and 18B. The outputs of the two predictors are selected by a selection switch S W 2;
A negative input of adder 10 and another positive input of adder 16 are provided.

The selection switch SW2 selects the input digital image signal X,
It is controlled by the prediction method selection signal a from the optimal prediction selection controller 22 which receives the prediction method selection signal a. The selection of the predictors 18A and 18B is adaptively controlled so that the prediction error is minimized by the optimum prediction selection controller 22 determining the state of the input image.

Further, the output of the variable length encoder 14 is sent to a buffer memory 22.
is output as a variable length DPCM code. This buffer memory is caused by buffer memory overflow,
In order to prevent underflow, a quantizer selection signal S is output to the selection switch SWl to control selection of the quantizer.

Problems to be Solved by the Invention Conventional variable length quantization application prediction D shown in FIG. 2 above
In the PCM coding system, for example, assume that two types of prediction methods are employed: interframe prediction, which is effective for images with little motion, and intraframe prediction, which is effective for images with movement. For images with little movement, the inter-frame prediction error is small (the frequency of motion increases, but
When intra-frame prediction is selected for a moving and complex picture, large prediction errors occur more frequently. In this way, the probability of occurrence of a prediction error is different between intra-frame prediction and inter-frame prediction. -On the other hand, only one variable length encoder is provided in common to both prediction systems.

Now, for example, when an input image with little movement is input and an interframe prediction method is adopted accordingly, a short digital signal is assigned to a small prediction error to minimize the amount of information generated. Assume that the variable length encoder is configured so that In such a case, as long as the input image is an image with little movement, D PCM transmission can be performed with the minimum amount of information generated.
When the input image changes to a moving and complex picture and correspondingly switches to the intra-frame prediction method, the variable length encoder assigns a short digital signal to a small i-order prediction error. As the encoder operates, in this case the variable length encoder operates to allocate long digital signals to large prediction errors that occur frequently.

Therefore, in this case, the amount of digital information output by the variable length encoder is not minimized.

As described above, the conventional variable-length quantization-applied predictive DPCM transmission apparatus does not take into account differences in prediction error occurrence frequency due to differences in prediction methods. Therefore, variable length encoders do not always apply variable length encoding to patents that are characterized by the amount of information generated.
There were times when the amount of information generated was unnecessarily large. When the amount of generated information increases in this way, if a plurality of quantizers are provided, the quantizers are switched, the number of quantization steps decreases, and the image quality deteriorates.

SUMMARY OF THE INVENTION Therefore, the present invention aims to solve the above-mentioned problems and provide a variable length quantization predictive DPCM transmission device that always minimizes the amount of generated information even if the nature or characteristics of the input signal changes. It is something to do.

Means for solving the problem, that is, according to the present invention, includes a first adder that receives an input signal at a first input and outputs a difference between the input signal and the second input; a variable length encoder that receives the output of the quantizer and outputs a DPCM signal; and an adder that receives the output of the quantizer at a first input and outputs the sum of the output and the second input. and a plurality of predictors connected to the output of the adder, and outputs of the predictors are selected and output to a second input of the first adder and a second input of the second adder. A variable-length quantization adaptive predictive DPCM transmission device that selects and adaptively controls a predictor so that a prediction error is minimized, comprising a selection means that A plurality of variable length encoders are provided which are optimized to reduce be done.

In the variable-length quantization predictive DPCM transmission device as described above, a variable-length encoder is provided for each prediction method to minimize the amount of generated information depending on the frequency of occurrence of prediction encapsulation errors. Therefore, if the prediction scheme is adaptively selected to minimize the prediction error, the variable-length coder is also selected accordingly, so that even if the nature or characteristics of the input signal changes, DPC that minimizes the amount of information
M code is output.

Embodiments Hereinafter, embodiments of a variable length quantization applied predictive DPCM transmission apparatus according to the present invention will be described with reference to the accompanying drawings. Note that the same parts as those of the conventional predictive DPCM transmission apparatus using variable length quantization shown in FIG. 2 are given the same reference numerals.

The illustrated predictive DPCM transmission device applying variable length quantization has an adder 10 that receives a digital image signal The difference, or prediction error, is output to two quantizers 12A and 12B.

The quantized prediction error qt quantized by the quantizers 12A and 12B is input to two variable length encoders 14A and 14B via a selection switch SWI.

This selection switch SWI is a buffer memory 2 which will be described later.
In order to prevent overflow and underflow of 2, the quantizer selection signal bt from the buffer memory 22 is used.

Furthermore, when intra-frame prediction is performed, the variable length encoder 14A converts the quantized prediction error q into a variable length so that the amount of generated information is minimized depending on the frequency of occurrence of each part of the quantized prediction error Qt. The variable length encoder 14B encodes the quantized prediction error Q when interframe prediction is performed.
The quantization prediction error q is variable-length coded so that the amount of generated information is minimized depending on the frequency of occurrence of each part of t.

The output end of the selection switch SWI is further connected to the plus input of an adder 16, and the output of the adder 16 is input to two predictors 18A and 18B. Those predictors 1
Predictor 18A of 8A and 18B operates according to the intra-frame prediction method, and predictor 18B operates according to the inter-frame prediction method. The outputs of predictors 18A and 18B are selected by selection switch SW2 and supplied to a minus input of adder 10 and another plus input of adder 16.

The selection switch SW2 is controlled by a prediction method selection signal a from the optimal prediction selection controller 22 that receives the input digital image signal X. The optimal prediction selection controller 22 determines the state of the input image and selects and adaptively controls one of the predictors 18A and 18B so that the prediction error is minimized.

Furthermore, the outputs of the variable length encoders 14A and 14B are selected by the selection switch SW3, applied to the buffer memory 22, and output as variable length DPCM codes. This selection switch SW3 is also controlled by the prediction method selection signal a from the optimal prediction selection controller 22.

Thus, if SW2 selects the predictor 18A based on the prediction method selection signal a output from the optimal prediction selection controller 20, then the variable length encoder 14A is selected by SW3, and the quantization prediction error Qt is variable-length encoded by the variable-length encoder 14Aat. Therefore, DPCM encoding is performed using intraframe prediction, variable length encoding suitable for intraframe prediction is performed, and images with rapid motion are transmitted as DPCM with a minimum amount of information.

On the other hand, if SW2 selects the predictor 18B based on the prediction method selection signal at output from the optimal prediction selection controller 20, then the variable length encoder 14B is selected by SW3. At this time, the predictor 18B performs inter-frame prediction, and the variable-length encoder 14B performs variable-length encoding suitable for inter-frame prediction, so that complex images with little motion are transmitted by DPCM with a minimum amount of information. be done.

In the above embodiment, two quantizers 12A and 12
The reason for providing B is that the prediction error may become large depending on the image, so we prepare multiple quantizers with different quantization steps and adjust the quantizer adaptively depending on the amount of data occupied in the buffer memory. control to select,
This is to prevent overflow and underflow of the buffer memory. Therefore, although two or more predictors are required, one quantizer may be used.

Furthermore, in the above embodiment, an intra-frame predictor and an inter-frame predictor are used as two predictors, but three predictors including an inter-field predictor and three variable-length coders may also be used. The same is true.

Effects of the Invention As is clear from the above, according to the present invention, the optimum variable length encoder is selected depending on the prediction method.
The amount of information generated by the variable length encoder is smaller than in the conventional variable length encoding method, which uses common variable length encoding regardless of the prediction method.

[Brief explanation of the drawing]

FIG. 1 shows a predictive DPCM using variable length quantization according to the present invention.
FIG. 1 is a block diagram of one embodiment of a transmission device. FIG. 2 is a block diagram of an example of a conventional variable length quantization applied predictive DPCM transmission device. [Main reference numbers] 10.16...Adder 12A, 12B...Quantizer 14.14A, 14B...Variable length encoder 18A, 18B
... Predictor 20 ... Optimal prediction selection controller 22 ... Buffer memory

Claims (2)

[Claims] (1) A first adder that receives an input signal at its first input and outputs the difference from the second input; a quantizer that receives the output of the first adder; and a quantizer that receives the output of the quantizer. a variable length encoder that receives the output of the quantizer and outputs a DPCM signal; an adder that receives the output of the quantizer at a first input and outputs the sum of the DPCM signal and a second input; and a plurality of encoders connected to the output of the adder. predictors and the outputs of those predictors are selected to be connected to the second input of the first adder and the second input of the first adder.
A variable length quantization adaptive predictive DPCM transmission device that selects and adaptively controls a predictor so that a prediction error is minimized, comprising a second input of an adder and a selection means for outputting to a second input of an adder. A plurality of variable length encoders are provided which are optimized to reduce the amount of generated information corresponding to each of the plurality of predictors. Variable length quantization applied predictive DPCM characterized in that the output of a long encoder is selected and output
Transmission device. (2) The input signal is a digital image signal, and the plurality of predictors are at least two predictors selected from an intraframe predictor, an interfield predictor, and an interframe predictor. Claim No. (1)
The predictive DPCM transmission device applying variable length quantization as described in .