JP2821124B2 - Decoding method and decoding device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a decoding method and a decoding device for decoding a coded code formed by predictive difference coding of an information signal such as a video signal. [Prior Art] As a system for performing this kind of differential encoding / decoding,
In this specification, a digital video tape recorder (DVTR) that performs band compression processing will be described as an example. In recent years, in addition to DVTRs for broadcast stations that can record and reproduce extremely high-quality images without performing band compression processing, in addition to DVTRs for broadcasting stations, the amount of data is reduced by band compression processing for home use, and compact and long-term recording is possible. What is possible is under development. A typical example of the band compression processing is a combination of offset subsampling and a differential PCM code (DPCM). Offset subsampling is a method of thinning out pixels so as to maintain the resolution in the horizontal and vertical directions of an image and reduce the resolution in an oblique direction that is difficult for human eyes to perceive. On the other hand, DPCM is a method of reducing the number of bits per pixel by utilizing the correlation of images. For example, the difference value between adjacent pixels is non-linearly quantized to reduce the number of bits. After performing offset subsampling as described above, DP
By performing the CM processing, the data amount can be significantly reduced without greatly deteriorating the image quality. [Problems to be Solved by the Invention] In the above-mentioned DPCM, when the difference value between pixels is small, the difference value is finely quantized, and when the difference value is large, non-linear quantization such as coarse quantization is performed. Therefore, a quantization error is likely to occur in a portion having a large difference value between adjacent pixels, such as an outline portion of an image. In addition, since the quantization error changes with time, a phenomenon in which the outline of the image appears to flicker, a so-called edge business, occurs. The occurrence of the quantization error as described above is a problem peculiar to the DPCM, and in order to reduce the quantization error as much as possible, an encoder having various configurations that makes the predicted value that takes the difference from the true value as close to the true value as possible. Has been proposed. However, even if the encoder is improved, the possibility that the difference value becomes large remains, and it is necessary to construct a complicated predictor. The present invention has been made in view of the above-described problems, and provides a decoding method and decoding that can reduce a quantization error in DPCM by a simple method without changing an encoder itself and reduce signal degradation. The purpose is. [Means for Solving the Problems] Under such a purpose, in the decoding method and the decoding device of the present invention, the decoding method and the decoding device are limited to a predetermined cutoff band,
In addition, an information code in which a predetermined band located in a high frequency band within the cutoff band is attenuated, is subjected to predictive difference coding using a correlation between samples obtained by sampling at a predetermined sampling interval. In a decoding method for decoding, the received encoded code is decoded, and the predetermined band of the decoded information signal is emphasized and output. Further, the decoding apparatus of the present invention obtains an information signal obtained by sampling at a predetermined sampling interval an information signal which is limited to a predetermined cutoff band and in which a predetermined band located in a high band within the cutoff band is attenuated. Means for receiving an encoded code obtained by predictive difference encoding using the correlation between the samples, means for decoding the received encoded code, and emphasizing the predetermined band of the decoded information signal. Output means. [Operation] According to the method and the apparatus as described above, the steep rising portion and the falling portion of the information signal are transmitted in a state where the quantization error at the time of encoding is small because the predetermined band is attenuated in advance. Since the frequency characteristics are restored to the same as those of the original signal by decoding and emphasizing the predetermined band, it is possible to obtain a decoded signal in which the characteristic is not deteriorated much due to the occurrence of the quantization error. [Example] Hereinafter, an example in which the present invention is applied to a DVTR will be described. First Embodiment FIG. 1 is a diagram showing a schematic configuration of a DVTR as one embodiment of the present invention, in which 101 is an input terminal of a video signal, and 102 is an analog-digital (A / D) conversion. 103, a spatial filter that limits the spatial frequency of the input video signal, 104, a subsampling circuit that performs offset subsampling by thinning out pixels, 105, a DPCM encoder that performs DPCM, and 106, an error correction code is added. Encoder 107, a recording amplifier; 108, a recording head; 109, a recording medium such as a magnetic tape; 108 ', a reproducing head; 110, a reproducing amplifier; 111, an error correction code for correcting data errors. A correction circuit, 112, a DPCM decoder that expands the DPCM-converted signal and restores the original data, 113, a 0 inserter that inserts 0 as pixel data thinned out by subsampling, and 114, a 0 inserter Pixel row Interpolation filter of a signal serving as a location filter after performing interpolation by spatially band limitation, 115 is a digital - analog (D / A)
The converter 116 is a terminal for outputting the restored video signal. Next, the operation of each unit of the present embodiment will be described in detail.
The video signal input to the terminal 101 is converted into a digital signal by the A / D converter 102 and then supplied to the spatial filter 103. The spatial filter 103 is provided for removing interference caused by aliasing components generated at the time of sampling by the offset sub-sampling circuit 104 at the subsequent stage. Assume that the sub-sampling circuit 104 performs pixel thinning in a pattern as shown in FIG. However, in FIG. 2 (A), .largecircle. Indicates a pixel transmitted to the subsequent stage, and X indicates a pixel not transmitted to the subsequent stage due to thinning. In this case, a spatial filter having a pass band of a spatial frequency region indicated by a hatched portion in FIG. 2B is usually used. Fig. 2 (B)
The middle h-axis indicates the frequency in the horizontal direction, and the υ-axis indicates the frequency in the vertical direction. Assuming that the sampling intervals of adjacent pixels (including the pixels to be decimated) are set to the horizontal direction th and the vertical direction t, The cut-off frequency fch is 1 / (2th), and the cut-off frequency fc in the vertical direction is 1 / (2t), and the cut-off frequencies fch, fc in the horizontal and vertical directions are the same as in the case where all pixels are transmitted. In the DVTR of this embodiment, the characteristics of the spatial filter 103 are represented by the horizontal cut-off frequency fch as shown by the solid line in FIG.
A high-frequency attenuation characteristic in which the vicinity is attenuated. The frequency limit in the vertical direction is not changed from the above-mentioned normal spatial filter. In FIG. 3, the upper right diagram shows the amplitude (a) limiting characteristic of υ on the axis of h = 0, and the lower left diagram shows the amplitude limiting characteristic of h on the ０ = 0 axis. The solid line is the characteristic of the spatial filter 103 in this embodiment, and the dotted line denoted by G (υ, h) is the characteristic of the ordinary spatial filter (hereinafter N (υ,
h)). The signal which is band-limited by the spatial filter 103 and further subjected to pixel thinning by the sub-sampling circuit 104 is a DPCM signal.
Input to encoder 105. The DPCM encoder 105 performs so-called previous value prediction, which calculates a prediction value using only all pixels, that is, only transmission pixels adjacent in the horizontal direction. The edge business described above occurs at a portion where a change between adjacent pixels in the horizontal direction is large, such as a contour portion of a pixel, for example, a contour portion extending in a vertical direction of an image when performing the DPCM. However, in the case of the present embodiment, the signal waveform of the above-mentioned contour portion becomes blunt due to the action of the spatial filter 103 described above. FIGS. 4 (A) and 4 (B) show this in an extreme manner, and the signal waveform as shown in FIG. 4 (A) is fch
Due to the attenuation in the vicinity, the signal becomes a dull edge portion as shown in FIG. 4 (B). For this reason, the difference value in the DPCM encoder 105 becomes small, and relatively fine quantization is performed at the time of nonlinear quantization, so that the quantization error becomes small. The signal whose data amount has been compressed by the DPCM encoder 105 is recorded on a recording medium 109 by a recording head 108 via an error correction code encoder 106 and a recording amplifier 107. Next, the operation at the time of reproduction will be described. Playhead 10
The signal reproduced from the recording medium 109 at 8 'is supplied to the reproduction amplifier 11
The signal is supplied to the error correction circuit 111 via 0, and a code error is corrected. The output signal of the error correction circuit 111 is the DPCM decoder 1
The data is supplied to the image data 12 and the data is expanded to become the original image signal. In the image signal combined by the DPCM decoder 112, 0 is inserted into the pixels decimated at the time of sub-sampling by the 0 inserter 113 and supplied to the post-interpolation filter 114. 5th characteristic
The characteristics are as shown in the figure. The meaning of each part in FIG. 5 is the same as that in FIG. 3, and has a characteristic that emphasizes the vicinity of fch, that is, the high range as shown by the solid line in the lower left diagram. Here, assuming that the characteristic of the interpolation filter 114 is F (υ, h), N (υ, h) = G (υ, h) F (υ,
h), that is, N (υ, h) is set to be a convolution of G (υ, h) and F (υ, h), and the spatial filter 103 and the interpolation filter 114 have complementary characteristics. It has. Therefore, the signal output through the interpolation filter 114 has the same frequency spectrum as the signal obtained through the normal pre-spatial filter, the interpolation filter, and the DP signal.
The quantization error accompanying CM is small. So D /
From the video signal output via the A converter 115 and the terminal 116, a good image can be obtained with little edge business. In the spatial filter 103 of the present embodiment, the gain at the horizontal cut-off frequency fch is 0.5 in the characteristic N (h, h) of the ordinary spatial filter, whereas the gain is fch / f as shown in FIG. It is attenuated from 2 so that it becomes 0.2 in fch. On the other hand, the high-frequency emphasis characteristic of the interpolation filter 114 is set so as to emphasize from fch / 2 as shown in FIG. 5, that is, the gain becomes 1 or more and the gain becomes 2 in fch. Needless to say, the characteristic G (Ｇ, h) of the spatial filter 103 and the characteristic F (υ, h) of the interpolation filter 114 can be easily adjusted by changing the tap coefficient of the filter. In the DVTR as described above, the quantization error due to the DPCM can be suppressed small, and a pre-filter for sub-sampling as a means for attenuating the high frequency band of the input signal, a signal decoded by the reproduction system Since a post-filter is used as means for emphasizing the high frequency band, the suppression of the quantization error can be realized without adding any circuit configuration as compared with the related art. Second Embodiment Next, a DVTR according to another embodiment of the present invention will be described.
The configuration of the DVTR according to the present embodiment is the same as the configuration shown in FIG. 1 except that the DPCM encoder 105 performs so-called two-dimensional prediction, and the spatial filter 103, the DPCM decoder 112, and the interpolation The configuration of the filter 114 is different from that of the first embodiment. Hereinafter, only these components will be described. First, a DPCM encoder (this embodiment is hereinafter referred to as 105 ')
The method of calculating the predicted value will be described with reference to FIG. In FIG. 6, ◎ indicates a target pixel for which a predicted value is to be calculated, ○ indicates a subsampled pixel, and × indicates a pixel thinned out by subsampling. Assuming that the values of the pixels P1, P2, and P3 are p1, p2, and p3, the predicted value q of the pixel indicated by ◎ is q = p3 + p1
Set to -p2. Of course, DPCM decryption (hereinafter 112 ') is D
It goes without saying that processing corresponding to the PCM encoder 105 'is performed. The edge business when performing such DPCM encoding is unlikely to occur for edges extending in the vertical and horizontal directions, and is likely to occur only for edges extending in oblique directions. Accordingly, the characteristic of the front value space filter (hereinafter, referred to as 103 ') is attenuated near the cutoff frequency in the oblique direction as shown by the solid line in FIG. In FIG. 7, x and y are oblique frequency axes. On the other hand, for the post-interpolation filter, the eighth
As shown by the solid line in the figure, the characteristic is such that the vicinity of the cut-off frequency in the oblique direction is emphasized. The characteristic shown in FIG.
(Υ, h), when the characteristic shown in FIG. 8 is F ′ (υ, h), G ′ (υ, h) F ′ (υ, h) = N (υ, h). In the second embodiment, the improvement of the DPCM encoder 105 'makes it possible to improve the quantization error that could not be improved by the spatial filter 103' and the interpolation filter 105 '. Very good images can be reproduced. Needless to say, the characteristics shown in FIGS. 7 and 8 can be realized by changing the coefficient of each tap. In the first and second embodiments, the present invention is applied to a DVTR. However, the present invention is applicable not only to video signals but also to other information signals such as audio signals. The present invention can be applied to a case where the transmission path is not limited to the magnetic recording / reproducing system but is transmitted using various communication paths. [Effects of the Invention] As described above, according to the decoding method and the decoding apparatus of the present invention, the encoding method itself is not particularly complicated, and the deterioration of the characteristics due to the occurrence of the quantization error is small. It is now possible to obtain a decoded signal.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram showing a schematic configuration of a digital video tape recorder (DVTR) as a first embodiment of the present invention, and FIGS. 2 (A) and (B) are shown in FIG. For explaining the operation of the sub-sampling circuit and the pre-value space filter in FIG. 3, FIG. 3 is a diagram showing the characteristics of the pre-value space filter in FIG. 1, and FIGS. 4 (A) and 4 (B) are FIG. 1 is a diagram for explaining the operation of a front value space filter in FIG. 1, FIG. 5 is a diagram showing characteristics of a post-interpolation filter in FIG. 1, and FIG. 6 is a second embodiment of the present invention. DP in DVTR as an example
FIG. 7 is a diagram for explaining the configuration of a CM encoder. FIG. 7 is a diagram showing characteristics of a pre-spatial filter in the second embodiment. FIG. 8 is a post-interpolation filter in the second embodiment. FIG. 6 is a diagram showing characteristics of the present invention. In the figure, 103 (103 ') is a previous value space filter, 104 is a subsampling circuit, 105 (105') is a DPCM encoder, and 112 (11
2 ') is a DPCM decoder, 113 is a 0 inserter, and 114 (114') is a post interpolation filter.

Claims (1)

(57) [Claims] Prediction is performed using a correlation between samples obtained by sampling an information signal, which is limited to a predetermined cutoff band and attenuated by a predetermined band located in a high band within the cutoff band, at a predetermined sampling interval. What is claimed is: 1. A decoding method for decoding an encoded code obtained by differential encoding, comprising decoding the received encoded code, and enhancing and outputting the predetermined band of the decoded information signal. Decryption method. 2. Prediction is performed using a correlation between samples obtained by sampling an information signal, which is limited to a predetermined cutoff band and attenuated by a predetermined band located in a high band within the cutoff band, at a predetermined sampling interval. Means for receiving an encoded code obtained by differential encoding, means for decoding the received encoded code, and means for enhancing and outputting the predetermined band of the decoded information signal. A decoding device characterized by the following. 3. The decoding device according to claim 2, wherein the predetermined band of the information signal is emphasized by an interpolation filter that increases the number of samples of the decoded information signal.