JP3198559B2 - Video information amount compression and / or information amount expansion device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video information compression and / or
Or, it relates to an information amount expansion device.

[0002]

2. Description of the Related Art FIG. 5 is a block diagram of a conventional still picture compression apparatus, and FIG. 6 is a view for explaining the operation of a zigzag scanner.

[0005] Referring to FIG. 5 , JPEG (Joint Photograb) is an organization that promotes international standardization of color still image coding.
phic Expert Group), a conventional still image compression device (see Journal of the Institute of Television Engineers of Japan P158 Vol44 No2 1990)
First, a DCT block obtained by extracting eight vertical and eight horizontal digital pixel data from one screen is referred to as a DCT block.
Transformer 1 (8 × 8 DCT), where two-dimensional DC
An 8 × 8 transform coefficient obtained by performing the T transform is converted to a linear quantizer 2
To supply. Here, DCT is one of orthogonal transform coding.
Abbreviation for the species Discrete Cosine Transform.

Then, linear quantization is performed using a quantization matrix obtained by multiplying a reference quantization matrix supplied from a quantization matrix generator 3 (quantization matrix) by a scaling factor S in a multiplier 4. The obtained DC conversion coefficient 2a and AC conversion coefficient 2b are supplied to the one-dimensional predictor 5 and the zigzag scanner 7, respectively.

The one-dimensional predictor 5 compresses the information amount of the DC conversion coefficient 2 a by obtaining the difference value of the continuous DC conversion coefficient 2 a and performs one-dimensional prediction, and supplies it to the first Huffman encoder 6. I do. Then, the first Huffman encoder 6 generates a Huffman code that allocates a smaller number of bits in the order of higher event occurrence probability and supplies the generated Huffman code to one input of the multiplexer 9. Note that the Huffman code is a kind of entropy code because it is formed based on the occurrence probability of an event. Also, the occurrence probability of an event may use a result measured in advance, and need not be obtained at the time of encoding.

On the other hand, the zigzag scanner 7 scans the AC conversion coefficient 2b from a low-frequency component to a high-frequency component as shown in FIG. 6 , and supplies this to a second Huffman encoder 8. Then, the value of the second Huffman encoder 8 is “0”.
Similarly, a Huffman code is generated for the combination of the run length of the transform coefficient and the value of the transform coefficient other than “0”,
The other input of the multiplexer 9 is supplied.

The multiplexer 9 outputs a multiplexed output signal 9a obtained by multiplexing the respective inputs to a transmission line (not shown).

Since the above-mentioned conventional technique uses the Huffman code, the information amount of the multiplexed output signal 9a fluctuates with respect to a fixed input information amount. For this reason, the scaling factor S is used to control, for example, to keep the information amount of the multiplexed output signal 9a per field constant.

[0009]

However, in the above-mentioned prior art, since the linear quantizer 2 generates quantization noise, which will be described later, there is a problem in that there is a DCT block whose image quality is remarkably deteriorated. Hereinafter, a one-dimensional DCT will be described as an example to facilitate understanding of the problem.

FIG. 7 is a diagram for explaining image quality degradation due to compression and decompression using DCT due to quantization noise, and FIG. 8 is a diagram for explaining a quantization matrix. Hereinafter, the problem to be solved by the present invention will be specifically described with reference to the drawings.

Here, FIG. 7 (A) having a periodic change
Consider a one-dimensional digital waveform data sequence shown in FIG. 4C composed of a one-dimensional digital waveform data sequence shown in FIG. Note that the inverse DCT transform is defined by the following equation 1. Further, FIG. 7 (A) ~ (D)
The vertical axis in the middle represents the level of the video signal.

[0012]

(Equation 1) Then, the above-mentioned data sequence is subjected to one-dimensional DCT transform defined by the following expression 2, and the transform coefficient D (7) is quantized in a coarse quantization step, and then inversely quantized and inversely DCT-transformed. The waveform data strings obtained in this manner are shown in FIGS. 3B and 3D, respectively.

[0013]

(Equation 2) Here, the portions indicated by the dotted lines in FIGS. 8B and 8D are noises based on quantization noise generated due to quantization in a coarse quantization step.
In the waveform shown in FIG. 7B, a small noise is added at a place where a large change of the video signal level occurs, so that the difference sensitivity (“digital signal processing of image”, Nikkan Kogyo Shimbun (1987) P2
3) is low, the noise is hardly detected. On the other hand, the waveform shown in FIG. 3D has a high difference sensitivity, so that the noise appearing in the flat portion is a problem.

Although the above description has been given of a one-dimensional case,
The same applies to the two-dimensional case. The value of the quantization matrix shown in FIG. 8 increases as the horizontal and vertical frequencies increase in the diagonally lower right direction. That is, the higher the order of the transform coefficient, the coarser the quantization step at the time of linear quantization. This is because the amount of information is reduced by utilizing the fact that the visibility decreases in a portion where the spatial frequency is high and the visibility decreases with respect to a waveform in an oblique direction.

However, even in the two-dimensional case,7
The waveform formed by the flat portion and the contour portion as shown in FIG.
It has frequency components, and such a waveform has a two-dimensional DC
The transform coefficient obtained by performing the T transform is calculated by the quantization matrix.
When quantized using, the higher-order (diagonally lower right) quantization transform
The quantization noise is mixed into the coefficients. And this
Is inversely quantized and subjected to inverse two-dimensional DCT,
Noise due to the noise, especially in flat areas
Appears as harm.

Accordingly, the present invention has been made in view of the above problems,
An object of the present invention is to add information representing noise based on quantization noise to a video signal whose information amount has been compressed and output the resulting signal, and to use this when expanding the information amount to remove noise based on quantization noise.

[0017]

SUMMARY OF THE INVENTION The present invention provides the following arrangement for solving the above-mentioned problems.

A main encoding means for subjecting an input block signal obtained by dividing a video signal to DCT transformation, quantization, and outputting a main encoded signal obtained by encoding, and inversely transforming quantization noise accompanying the quantization. Auxiliary encoding means for encoding noise information obtained by DCT transformation and outputting an auxiliary encoded signal to which discrimination information for distinguishing from the main encoded signal is added; Multiplexing means for outputting an information amount compressed signal obtained by multiplexing the signal and the auxiliary code.
Means for converting the noise information obtained by the inverse DCT transform.
In other words, according to a predetermined criterion representing the visual inconspicuousness of noise
On the other hand, the noise of the pixel where the noise higher than the reference
A video information amount compression or information amount compression / decompression device that encodes after reducing size information.

A video information amount expansion or information amount expansion / compression device for performing an information amount expansion on the information amount compression signal according to claim 1, wherein a decoded block signal obtained by decoding the main coded signal is: A video information amount expansion or information amount compression / expansion device, comprising: noise suppression means for performing noise suppression based on the noise information concerning the auxiliary coded signal.

[0020]

FIG. 1 is a block diagram for explaining a main part of a video information amount compression apparatus according to a first embodiment. FIG. 2 is a numerical value for explaining a main part of the video information amount compression apparatus according to the first embodiment. FIG. 3 shows an example, FIG. 3 is a block diagram for explaining a main part of the video information amount expanding apparatus of the first embodiment, and FIG. 4 is a diagram for explaining a main part of the video information amount expanding apparatus of the first embodiment. it is a diagram showing a numerical example. Hereinafter, an embodiment of a video information amount compression and / or expansion device according to the present invention will be described with reference to the drawings. The outline of the first embodiment is as follows. In the information amount compression apparatus, noise information based on quantization noise generated when the information amount is expanded is obtained in advance, and the contour (region having a high spatial frequency component) or the luminance signal level is large. in too or too small areas <br/> encodes reduce the amount of information of the upper Symbol noise information, which is to obtain the amount of information compressed signal multiplexes the main encoded output signal encoded video signal is there. Then, when performing the information amount expansion, the noise information separated from the information amount compressed signal and decoded are added to the decoded block signal, respectively. Hereinafter, an example of a main part of the video information amount compression device according to the present invention will be described, and then an example of a main portion of the video information amount expansion device according to the present invention will be described. (・ Video Information Amount Compression Device) In FIG. 1, the main parts of the video information amount compression device are a main encoding unit AA for compressing the information amount of the input video signal, and an auxiliary code for correcting the image quality deterioration due to the information amount compression. Means BB
And multiplexing means CC for multiplexing the outputs of the main coding means AA and the auxiliary coding means BB. Hereinafter, description will be made in order from the main encoding unit AA.

A DCT block 1a obtained by extracting eight vertical and eight digital pixel data from an input video signal is converted into a DCT block 1a by a DCT block divider (not shown).
It is supplied to the dimensional DCT converter 1. Here, the DCT block shown in FIG. 2A is considered as an example of the DCT block 1a, and other numerical examples will be described below based on this numerical example. It is to be noted that the number representing the digital data in FIG.
27. The input video signal may be any signal that represents image information such as a Y signal, an RY signal, and a BY signal.

The two-dimensional DCT converter 1 performs two-dimensional DC
The transform coefficient block 1 b obtained by performing the T transform is supplied to the quantizer 2. The transform coefficient block 1b represented by 12 bits is shown in FIG.

The quantizer 2 quantizes the transform coefficient block 1b by dividing each value of the quantization matrix shown in FIG. 8 as a quantization step to obtain a quantized value shown in FIG. 2C. Block 2c is converted to the first encoder 1
Supply 0.

Then, the main encoder 10a obtained by subjecting the quantized block 2c to known encoding by the first encoder 10 is input to one input of a multiplexer 26 in the multiplexer CC. To supply.

The main encoding means AA is a part corresponding to the video information amount compression apparatus according to the above-mentioned conventional technique, but differs in the configuration after the quantizer 2. That is, the first encoder 10 is limited to the conventional technology including the one-dimensional predictor 5, the first and second Huffman encoders 6, 8, the zigzag scanner 7, and the multiplexer 9. What is necessary is just to perform well-known encoding rather than a thing.

Next, the auxiliary encoding means BB will be described. The outline is that noise generated on the basis of quantization noise which is a problem when the information amount is expanded is obtained, and for pixels in which this noise is visually inconspicuous. Is an ID for reducing the noise information and distinguishing it from the main coded signal 10a.
Auxiliary coded signal 18 a obtained by adding information (determination information)
The, and supplies to the other input of the multi-duplex unit 2 6.

First, the quantization block 2c is the inverse quantizer 1
After being supplied to 1 and subjected to inverse quantization, it is supplied to one input of a subtractor 12, where it is subjected to subtraction with the transform coefficient block 1b supplied to the other input to obtain FIG. D)
Are supplied to the second quantizer 13.

FIG. 2 shows the quantization noise 12a quantized by the second quantizer 13 and rounded to the upper several bits.
The second quantizer output signal 13a shown in (E) is supplied to the inverse DCT transformer 14. Note that the second quantizer 1
Although 3 is not always necessary, the configuration of the inverse DCT converter 14 is simplified by rounding the data to be inversely DCT-converted into upper several bits. Then, supplies the zero encoder 15 obtains a noise 14a rather groups Dzu the quantization noise by inverse DCT by the inverse DCT transformer 14.

[0029] Here, the detail reason for it this seeking noise 14a rather groups Dzu to quantization noise by transforming the inverse DCT, those representing the pixel data in the DCT block 1a in the form of a matrix [Xij] , The transform coefficient block 1b is likewise [dij], and the quantization noise is likewise [Ti
j], similarly quantization noise based Dzu rather noise 14a [Ni
j] Two-dimensional DCT transform is D, and inverse two-dimensional DCT transform is D
−1, and i and j are integers of 0 to 7, respectively. Here, the inverse D performed by using a transform coefficient having quantization noise
The CT conversion is represented by the following equation 3.

[0030]

(Equation 3) Also, both the two-dimensional DCT transform and the inverse DCT transform are linear transforms, and the following equation 4 holds, so that equation 5 holds.

[0031]

(Equation 4)

[0032]

(Equation 5) The equation 5, the noise 14a rather groups Dzu the quantization noise it is understood that it is possible to obtain by inverse DCT quantization noise. The inverse DCT converter 14 performs this calculation. The quantization noise based Dzu rather noise 14a FIG 2 (H)
Illustrated in FIG.

[0033] However, the noise 14a rather than based on Dzu to quantization noise
Need not be transmitted, it is sufficient to transmit only those pixels where noise is noticeable when the information amount is expanded.

Therefore, the nullifier 15 is provided with the first discriminating means B
B1 and second by the position information of the pixel obtained noise conspicuous absence by a determination unit, the noise 14a rather groups Dzu the quantization noise
Corresponding pixel data to "0" in, if not applicable outputs the nulling output signals 15a noise 14a rather groups Dzu the quantization noise for each pixel data.

Now, the first discriminating means BB1 will be described. A DCT block 1a is supplied to a two-dimensional HPF 19, and a two-dimensional HPF obtained by applying a spatial HPF is provided here.
The output block 19a is supplied to the absolute value converter 20. Here, an example of a coefficient example of the two-dimensional HPF 19 is shown in FIG. In FIG. 8F, a coefficient “8” surrounded by a bold frame indicates a coefficient of a target pixel, and another coefficient “−1” indicates a coefficient of a pixel around the target pixel. . And
Data each constituting the two-dimensional HPF output blocks 1 9 a by adding the nine data obtained by -1 times the digital data of the surrounding pixels with multiplying 8 the digital data of the pixel of interest And If the target pixel is at the end of the DCT block and there are no eight pixels around, the coefficient of the target pixel is made equal to the number of valid pixels. The two-dimensional HPF output block 19a thus obtained is shown in FIG.

The absolute value converter 20 takes the absolute value of each data constituting the two-dimensional HPF output block 19a and supplies it to the first comparator 21 having the first threshold. The first comparator 21 detects whether or not there is a flat portion in the DCT block 1a, and obtains “1” when it is smaller than the first threshold, and obtains “0” in other cases. Is supplied to one input of an AND circuit 25 as a low-frequency block 21a. Here, a pixel having a value of "1" in the low-frequency block 21a is a pixel having no high-frequency frequency component. Note that the first threshold is “64”.

[0037] Then, supplies at the AND circuit 25 and the other NOR circuit output signal 24a to the input is supplied, to obtain a logical product of both input, zero equalizer 15 so as AND circuit output signal 2 5 a .

Next, the second discriminating means BB2 will be described. The DCT block 1a determines whether the level of the luminance signal is high and the second comparator 22 which determines if the level of the luminance signal is low. Third comparator 2 for determining a certain case
3 and supplied. Then, in the second comparator 22, D
The pixel data in the CT block 1a is compared with a second threshold value. If the pixel data exceeds the threshold value, "1" is set, and if not, "0" is set to one input of the first NOR circuit 24. Supply each time. The third comparator 23 outputs DC
The pixel data in the T block 1a is compared with the third threshold value, respectively. If the pixel data does not exceed the third threshold value, "1" is set. Otherwise, "0" is set to the other input of the NOR circuit 24 for each pixel data. Supply. Note that the second threshold is “100” and the third threshold is “−100”.

Then, a NOR circuit 24 obtains a logical sum of both inputs and inverts to obtain a NOR circuit output signal 24a. The low-frequency block 21a and the NOR circuit output signal 24
Needless to say, the timing with a is adjusted by a well-known timing adjustment circuit.

The AND circuit 25 calculates the logical product of the low-frequency block 21a and the NOR circuit output signal 24a and supplies the AND to the nullifier 15 as the AND circuit output signal 25a. Here, the pixels for which the AND circuit output signal 25a is "0" are shown by hatching in FIGS. 2 (G) and 2 (H).

The logical product of both inputs is obtained as described above.
Pixels having a value of "0" in the low-frequency block 21a are pixels having high-frequency components, and thus are pixels whose noise is visually suppressed even if noise due to quantization noise occurs.
When the NOR circuit output signal 24a is "0", the level of the luminance signal is too high or too low, so that the pixel is visually suppressed in noise even in a flat portion. In this case, the corresponding pixel data is set to “0” by the nullifier 15 to compress the data amount when encoding is performed by the second encoder 17 described later.

Then, the nullifier output signal 15a is supplied to a third quantizer 16, where the third quantizer 16 quantizes the pixel data and rounds the pixel data to higher-order bits, as shown in FIG. 2 (I). The quantizer output signal 16a is supplied to the second encoder 17.

The second encoder 17 encodes the third quantizer output signal 16a. For example, the second encoder 17 encodes the AC transform coefficient 2c described in the prior art.
A second encoder output signal 17a obtained by sequentially scanning pixel data and performing Huffman encoding by combining a run length of a pixel “0” and a coefficient other than “0” is output to the ID adder 1
8

The auxiliary coded signal 18a obtained by adding ID information for distinguishing the second coder output signal 17a from the main coded signal 10a by the ID adder 18 is multiplexed as described later. To the vessel 26.

The multiplexing means CC includes a multiplexer 26.
At supplying the main coded signal 10a and the auxiliary coded signal 18 a to the transmission path or the like (not shown) the amount of information compressed signal 26a obtained by multiplexing.

It should be noted that the inverse D in the auxiliary encoding means BB is
The CT conversion may be simplified as in the first and second simplification means described below.

[0047] The first simplification means is the second quantizer output.
Assuming that the low-order transform coefficient of the signal 13a is "0", the inverse DCT
The circuit scale of the converter 14 is simplified and the operation time is reduced.
Is to reduce. This is when the quantization matrix used in quantizing the first quantizer 2 in which the spatial frequency as shown in FIG. 8 increases the quantization step higher becomes higher, lower order transform coefficients Focuses on the fact that quantization noise does not occur as much as higher-order transform coefficients .

The second simplification means is that the inverse DCT converter 14 performs the inverse DCT on all 8 × 8 pixels, and a part of the pixels is zeroized by the nullifier 15. The AND circuit output signal 2 is directly supplied to the DCT converter 14.
5a, and performs an inverse DCT transform only when this is "1", and when it is "0", zeroes it and supplies it directly to the third quantizer 16 without passing through the zeroizer 15. It is. (Image Information Decompression Device) An example of a main part of the image information decompression device according to the present invention will be described with reference to FIG. 3A. 1 is supplied to the selector 27, and the information amount compressed signal 26a is converted into the main coded signal 10a and the auxiliary coded signal 18 based on the ID information.
a second encoder output signal 17a obtained by removing ID information from a
And supplied to the first and second decoders 28 and 32, respectively.

The first decoder 28 corresponds to the above-described first encoder 10 and has the main encoded signal 10
The above-described quantization block obtained by decoding a (FIG. 2C)
Is supplied to a first inverse quantizer 29, where the first inverse quantizer output signal 29a obtained by inverse quantization is inverted D
It is supplied to an inverse DCT converter 30 (inverse DCT) that performs a CT transform. 4 output from the inverse DCT converter 30.
The decoded DCT block 30a shown in FIG.
1 to one input.

When the first encoder 10 is the one described in the prior art, the first decoder 28 is as shown in FIG. That is, the main coded signal 10a is supplied to the second selector 35, where it is separated into the decoded DC conversion coefficient 35a related to the DC conversion coefficient 2a and the decoded AC conversion coefficient 35b related to the AC conversion coefficient 2b. ,
The former is supplied to a one-dimensional predictive decoder 37 via a first Huffman decoder 36, where a one-dimensional predictive decoder output signal 37a corresponding to the DC conversion coefficient 2a is obtained, and this is obtained as a third signal. It is supplied to one input of a selector 38. On the other hand, the decoded AC transform coefficient 35b is supplied to the memory 40 via the second Huffman decoder 39, and the memory output signal 40a obtained by performing the timing adjustment is supplied to the other input of the third selector 38. Supplied. Then, both inputs are appropriately selected by a third selector 38 to obtain a first decoder output signal 28a.

On the other hand, the second decoder 32 corresponds to the above-described second encoder 17, and decodes the second encoder output signal 17a to generate a noise based on the quantization noise. The information is obtained and supplied to the delay circuit 33 (DL) as the second decoder output signal 32a.

This delay circuit 33 is used for the decoding DCT block 3
0a and is provided to adjust the timing of the decoder output signal 32 a, for example constituted by a digital memory. Then, the delay circuit output signal 33a whose timing has been adjusted by the delay circuit 33 is supplied to the second inverse quantizer 34, where the pixel data is inversely quantized by multiplying the pixel data by 8 times. A noise information signal 34a (shown in FIG. 4A) based on the noise is supplied to the other input of the adder 31.

Then, the adder 31 adds the decoded DCT block 30a and the noise information signal 34a for each pixel data to obtain an information amount extension signal 31a shown in FIG. 4C.

Here, the DCT block 1a shown in FIG. 2A and the decoded DCT block 3 shown in FIG.
0A and the information amount decompression signal 31a shown in FIG. 3C, it can be seen that, of the noise added to each pixel data of the decoded DCT block 30a, the information amount of a pixel whose pixel data is originally “0” is Noise is suppressed by the extension signal 31a. On the other hand, noise is not suppressed for a pixel whose original value is “120” or for a pixel at the boundary between “0” and “120”. However, there is no problem because the noise added to the contour portion or the pixel having an extremely large luminance level is hardly noticeable visually.

In the above-described embodiment, the video information amount compression device and the video information amount expansion device have been described separately.
Needless to say, an information amount compression / decompression device in which both are integrated may be used.

[0056] Incidentally, in the embodiment described above, first in the auxiliary coding device BB, is provided with the second discriminating means may be provided with either one of the these it is of course is there. Furthermore, of course also it may be one which does not include both, such a case, the noise 1 rather groups Dzu the quantization noise
4a without passing through the zeroizer 15 to the third quantizer 16
Directly to At this time, the encoding performed by the second encoder 17 is not limited to the entropy encoding.

As described in the above embodiment,
When compressing the amount of information, the quantization noise caused by quantization is directly
Since noise information was obtained by performing T-transform, the transform coefficient block 1b and the quantized block 2c before quantization were respectively subjected to inverse DCT.
There is an effect that the configuration can be simplified as compared with a case where noise information is obtained by converting and obtaining a difference value between them.

As described in the above embodiment ,
When encoding the noise information generated by the quantization noise caused by quantization, the information amount of the noise information in the region contour (region having a high spatial frequency components) or the luminance signal level is too large or too small Reduce and encode,
There is an effect that the information amount of the information amount compressed signal can be reduced.

[0059]

Effects of the Invention According to the present invention as described above, to encode the noise information generated by the quantization noise due to quantization
The noise information obtained by the inverse DCT transform
For a given criterion that represents the visual inconsistency of noise
The noise of the pixel where the noise above the standard is not conspicuous
To encode after reducing the information, the amount of compressed information signal
This has the effect of reducing the amount of information.

[Brief description of the drawings]

FIG. 1 is a block diagram for explaining a main part of a video information amount compression device according to a first embodiment.

FIG. 2 is a diagram showing a numerical example for explaining a main part of the video information amount compression device of the first embodiment.

FIG. 3 is a block diagram for explaining a main part of the video information amount expanding device according to the first embodiment;

FIG. 4 is a diagram showing a numerical example for explaining a main part of the video information amount expanding device of the first embodiment.

FIG. 5 is a block diagram of a conventional still image compression device.

FIG. 6 is a diagram for explaining the operation of the zigzag scanner.
You.

FIG. 7 shows the results of compression / expansion using DCT due to quantization noise .
FIG. 7 is a diagram for explaining image quality deterioration.

FIG. 8 is a diagram for explaining a quantization matrix;
You.

[Explanation of symbols]

1a DCT block (input block signal) 10a main coded signal 12a quantizing noise 18a auxiliary coded signal 2 6 a information amount compressed signal 30a decoded DCT block (decoded block signal) AA main encoding unit BB auxiliary coding means CC multiplex Means

────────────────────────────────────────────────── ─── Continued on the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) H04N ^7/ 24-7/68 H04N 1/41-1/419 H03M 7/30 JICST file (JOIS)

Claims (2)

(57) [Claims] 1. A main encoding means for subjecting an input block signal obtained by dividing a video signal to DCT transformation, quantization, and encoding to output a main encoded signal, and a quantization noise accompanying the quantization. Auxiliary encoding means for encoding the noise information obtained by performing the inverse DCT transform on the signal and outputting an auxiliary coded signal to which discrimination information for distinguishing from the main coded signal is added; Multiplexing means for outputting a compressed information amount signal obtained by multiplexing the coded signal with the coded signal , wherein the auxiliary coding means comprises:
Of the noise information that indicates the inconspicuousness of the noise
Noise that is higher than the above standard is not noticeable
Encoding after reducing the noise information
A video information amount compression or information amount compression / decompression device. 2. A video information amount expansion or information amount compression / expansion apparatus for performing an information amount expansion on said information amount compression signal according to claim 1, wherein said decoded block signal is obtained by decoding said main coded signal. A video information amount expansion or information amount compression / expansion apparatus, further comprising noise suppression means for performing noise suppression based on the noise information concerning the auxiliary coded signal.