JP2797411B2 - Encoding device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an encoding device used in a digital transmission system for images and sounds.

[Prior Art] When digitally transmitting information such as images and sounds, various types of encoding are performed to reduce the amount of transmitted information.
As one of encoding methods, there is a predictive encoding method (hereinafter, referred to as DPCM) for compressing the amount of information by utilizing the correlation between adjacent sample values. In DPCM, a predicted value for a sample value to be encoded is obtained from a decoded value that has already been transmitted, and a difference value (prediction error) between the predicted value and the sample value is quantized and transmitted. There are various prediction encoding methods depending on the method of generating the prediction value. FIG. 3 is a block diagram showing a configuration of an encoding apparatus of the simplest preceding value predictive coding method (a predictive coding method using one previous decoded value as a predictive value). In the following, for a sample value x _i , the difference value is e _i , the DPCM code is y _i , DP
The representative value based on the CM code y _i is Q (e _i ), the decoded value is _i ,
The predicted value is represented as _xip with the addition of a subscript p.

In FIG. 3, a subtractor 12 is a sample value input to an input terminal 10.
The prediction value x _ip (the decoded value of the previous value in this embodiment) is subtracted from x _i to output a difference value e _i . Quantizer 14, the difference value e _i are quantized outputs a quantization code y _i, the quantization code y _i
Is transmitted from the output terminal 16 to the transmission path. The DPCM code y _i is also applied to the quantization representative value generation circuit 18. The quantization representative value generation circuit 18 converts the DPCM code y _i into a quantization representative value Q.
(E _i ). An adder is added to the quantized representative value Q (e _i ).
By adding the predicted values at 20, the input sample values can be restored. Since the reconstructed input samples contain quantization errors, they may exceed the range of the original input samples.
Therefore, the range is amplitude-limited by the limiter 22. Limiter 22
Is the local decoded value _i, which is applied to a D flip-flop (predictor) 24. Since the decoded value is a predicted value, the predictor becomes a D flip-flop. The D flip-flop 24 outputs the local decoded value in the next clock cycle.
_i is applied to the subtractor 12 and the adder 20 as a predicted value.

In general, the difference value between the predicted value and the input sample value is such that the probability distribution is biased toward a small value portion, and the quantization step at a small difference portion is finely divided, and a large difference portion is coarsened to obtain an information amount. Can be compressed.

FIG. 4 is a block diagram showing the configuration of a decoding apparatus corresponding to the encoding apparatus shown in FIG. The DPCM code y _i input to the input terminal 26 is represented by the quantized representative value Q
(E _i ), and the adder 20 adds the predicted value. The limiter 32 limits the amplitude of the output of the adder 30. limiter
The output _{i of} 32 is output from the output terminal 34 as a decoded value. This decoded value is delayed by one cycle clock by the D flip-flop 36 and applied to the adder 30 as a predicted value _xip .

[Problems to be Solved by the Invention] In the conventional example shown in FIG. 4, when the dynamic range (hereinafter referred to as D range) of an input sample value is set to n levels of 0 to n−1, the difference value D The range is -n + 1 to n-1
(2n-1) level, which is about twice the input sample value. As a result, when the compression ratio is increased, when the quantization characteristic is set to the above-described non-linear characteristic (a characteristic in which the quantization step is made fine near the difference value of 0 and the quantization step is made coarser as the difference value becomes larger than 0), The maximum value of the absolute value of the quantization error (difference between before and after the quantization) of the portion distant from the predicted value becomes a very large value. This is the image quality degradation at the edge of the image where the difference value is large (edge business)
Was a major factor.

On the other hand, if the quantization step is made closer to linear to reduce this deterioration, the quantization error near the predicted value increases,
There is a problem that the decoded image is rough (increase in granular noise), and degradation such that contour lines (false contours) such as contour lines of a map are seen in a portion where the level change is gentle like a face.

Further, in the conventional example, when a transmission error occurs, the decoded value of the decoding device becomes a value different from the local decoded value of the encoding device. Since the decoded value becomes the predicted value of the next pixel, the transmission error propagates to subsequent pixels, and the original image cannot be restored.

Therefore, an object of the present invention is to provide an encoding device that solves such a disadvantage.

[Means for Solving the Problems] An encoding apparatus according to the present invention comprises: an input unit for inputting sample data; a first encoding mode for encoding the sample data by a first quantization characteristic; Encoding means having a second encoding mode for encoding difference data between sample data and prediction data using a second quantization characteristic different from the first quantization characteristic; and the first encoding mode Decoding means having a first decoding mode for decoding the coded data coded in the first mode, and a second decoding mode for decoding the coded data coded in the second coding mode; Generating means for generating the prediction data based on the decoded data decoded by the decoding means;
Selecting means for selecting one of the first encoding mode and the first decoding mode, and one of the second encoding mode and the second decoding mode based on the difference data; It is characterized by having.

Embodiment An embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.
40 is an input terminal of a sample value x _i to be encoded, 42 is a subtractor, 44
Is a PCM quantizer, 46 is a DPCM quantizer, 48 and 50 are inverse quantizers corresponding to the quantizers 44 and 46, 52 is an adder, 54 is a limiter, and 56 is a D flip-flop as a predictor. , 58
Is a selection switch for the output of the inverse quantizer 48 or the output of the limiter 54, 59 is a determination circuit for determining an edge portion, and 60 is a quantizer.
44, 46 output selection switches, 62 is buffer, 64 is DPCM
Output terminal of code y _i .

The subtractor 42 calculates a predicted value x _ip from the image data x _i of the input terminal 40.
, And outputs a difference value e _i . The quantizer 46 calculates the difference value e _i
Is quantized, and the DPCM code is applied to the contact b of the switch 60 and the inverse quantizer 50. The inverse quantizer 50 converts the DPCM code into a quantized representative value, and the adder 52 adds the predicted value of the output of the D flip-flop 56 to restore the original input sample value. The limiter 54 limits the amplitude of the restored sample value within the dynamic range of the input sample value.

The input sample values x _i of the input terminal 40 is also applied to the quantizer 44, the quantizer 44 quantizes it, and outputs the PCM code. The output of the quantizer 44 is applied to the a-contact of the switch 60 and the inverse quantizer 48. The inverse quantizer 48 decodes the output (PCM code) of the quantizer 44 into the original input sample value and applies it to the a-contact of the switch 58 as a local decoded value.

The edge part determination circuit 59 compares the absolute value of the output (difference value e _i ) of the subtractor 42 with a predetermined threshold value.
A control signal for connecting the switches 58 and 60 to the a contact side, and connecting the switches 58 and 60 to the b contact side when the threshold value or less is obtained,
Applied to switches 58 and 60, respectively. However, switch 6
When 0 is connected to the contact a side, it is necessary to transmit the control code indicating the switching of the transmission code first. Therefore,
The quantizer 46 has a function of generating the control code,
The switch 60 is switched to the contact b once and then to the contact a. For example, the DP of the DPCM quantizer 46
The surplus code in the CM code is used as the control code. The amount of information increased by the control code is adjusted by the buffer.

In the difference value between the predicted value and the input sample value, the occurrence probability is concentrated near 0, and therefore, normally, the switches 58 and 60 are connected to the contact b side. That is, the dynamic representative value of the quantization
Range limited DPCM coding is performed. By the way, if the number of quantization levels is fixed, the smaller the dynamic range of the representative quantization value, the finer the quantization step, so that the above-mentioned granular noise, false contour and edge business can be improved. On the other hand, as the dynamic range of the quantization representative value is reduced, the followability of the edge of the decoded value is deteriorated, and a gradient overload in which the edge is blunted occurs. Therefore, in the present embodiment, the PCM quantizer 44 is used for the pixel at the edge where the gradient overload occurs,
Deterioration due to gradient overload is suppressed.

FIG. 2 is a block diagram showing a configuration of a decoding apparatus corresponding to FIG. 66 is the input terminal for the encoded code _yi , 68 is PCM
Inverse quantizer, 70 is a DPCM inverse quantizer, 72 is an adder, 74 is a limiter, 76 is a D flip-flop as a predictor, 78 is
A control code detector for switching from the DPCM code to the PCM code, 80 is a changeover switch, and 82 is an output terminal for the decoded value _i .

The encoded code y _i input to the input terminal 66 is the inverse quantizer 6
8, 70 and the code detector 78 rejects it. Dequantizer 68,7
0 has the same function as the inverse quantizers 48 and 50 in FIG. 1, respectively. The inverse quantizer 68 outputs the decoded value of the PCM code, and the inverse quantizer 70 outputs the representative quantization value. The adder 72 adds the predicted value to the output of the inverse quantizer 70, and the limiter 74 limits the amplitude.

The code detector 78 generates a control signal for connecting the switch 80 to the a-contact when the input code of the input terminal 66 is the above-described control code indicating switching from DPCM to PCM. That is, the switch 80 is normally connected to the contact b, but when the control code is input to the input terminal 66,
The output is switched to the contact a side, and the output of the inverse quantizer 68 is selected.

The decoded value selected by the switch 80 is output from the output terminal 82. The signal is sent to the D flip-flop 76 and applied to the adder 72 as a predicted value.

Since the code input to the input terminal 66 includes two types of codes, a PCM code and a DPCM code, the inverse quantizers 68 and 70 must be selected in accordance with each code. The code detector 78 detects switching from the DPCM code to the PCM code, and switches the switch 80 to the contact a side. DPCM to PCM
Unless the control code for switching to
80 is connected to the contact b side.

[Effects of the Invention] As described above, according to the present invention, the first encoding mode for encoding the sample data using the first quantization characteristic, and the difference data between the sample data and the prediction data are stored in the first encoding mode. 1
Is selected based on the difference data, so that signal degradation as a whole can be reduced and efficient. Encoding can be performed.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the configuration of an encoding apparatus according to an embodiment of the present invention, FIG. 2 is a block diagram showing the configuration of a decoding apparatus corresponding to FIG. 1, and FIG. FIG. 4 is a block diagram showing the configuration of a conventional decoding device. 40: input terminal, 44: PCM quantizer, 46: DPCM quantizer, 48,6
8: PCM inverse quantizer, 50, 70: DPCM inverse quantizer, 59: decision circuit, 62: buffer, 64: output terminal, 66: input terminal, 78: code detector, 82: output terminal

Claims (1)

(57) [Claims] An input unit for inputting sample data; a first encoding mode for encoding the sample data with a first quantization characteristic; and a first encoding mode for inputting difference data between the sample data and the prediction data to the first encoding mode. Encoding means having a second encoding mode for encoding according to a second quantization characteristic different from the quantization characteristic; and a first encoding means for decoding encoded data encoded in the first encoding mode. And a decoding unit having a second decoding mode for decoding the encoded data encoded in the second encoding mode, and decoding data decoded by the decoding unit. Generating means for generating the prediction data based on the difference data; the first encoding mode and the first decoding mode; and the second encoding mode and the second decoding based on the difference data. Mode of And a selecting means for selecting either one.