JPH07101849B2 - Run length 1 / n compression floating code method and decoding method thereof - Google Patents

Description

Detailed Description of the Invention

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a code compression method used for high-quality recording and transmission of image data, and particularly to a run-length 1 / n compression floating system having a good instantaneous S / N ratio characteristic over a wide input range. The present invention relates to a coding method and a decoding method thereof.

2. Description of the Related Art In conventional satellite PCM broadcasting and digital audio tape recorders (hereinafter abbreviated as DAT), a comparatively light compression ratio is taken into consideration in consideration of a trade-off with the amount of hardware while suppressing quality deterioration due to compression. Has been adopted. An example of these is the line code, DAT
In, a 12-bit 13-line code is used. For detailed technical standards, refer to "EI
AJ CP-2305 DAT Cassette Sy
system Part1: Dimension and
Characteristics ". FIG. 6B shows a 13-line code (1
It is a figure explaining the conversion method of 2 bits), and shows a positive range. In the frame in the figure, the numbers in parentheses are decimal numbers, and the other numbers are binary numbers. The most significant bit of the binary number (hereinafter referred to as MSB) is the polarity bit P, which is "0" in the positive range and "1" in the negative range. "A in the figure
BCDEFGH ”is from“ 00000000 ”to“ 111
The value is up to 11111 ", and" * "is an arbitrary binary number. Further, L0 is the run length of continuous data Q0 in which “0” continues, excluding the polarity bit of the original data. The procedure of the above coding method will be described below. FIG. 6B is a diagram specifically showing conversion between encoding and decoding. First, a method for encoding 13 polygonal line codes (12 bits) will be described. The 16-bit area is divided into seven in the positive range, and the shift amount is manipulated according to each area. As is clear from the figure, the area division is M
It can be performed at a position where "1" is first seen from the SB side.
In other words, it can be determined by the run length L0 of continuous data of "0" excluding the polarity bit. There are eight run lengths L0 from "0" to "7 or more", and the complement of L0 is expressed exponentially as range data. For example, the original data is "1216 (decimal)", that is, "000001001".
In the case of "1000000 (binary)", the run length L0 of the continuous data Q0 is "4". The range data Q2 is “011” which is the inverse of “100”. Also, "AB
"CDEFGH" is "00110000". Therefore, the output data is "001100110000". FIG. 6A is a diagram schematically showing the above concept. The output data is 1 by combining the polarity bit P, the range data Q2 representing the area, and the shifted mantissa data D2.
Set to 2 bits. Range data Q2 has a fixed word length,
It is 3 bits. When the original data is 2S (toes) complementary, the negative range can be encoded in exactly the same way as the positive range by inverting all the bits of P or less. The code in which the bit is inverted in the negative range is called a folded binary code. In the following description, a folded binary code will be used unless otherwise specified, but conversion with the 2S complementary is mutually possible. Next, a method of decoding 13 polygonal line codes (12 bits) will be described. In decoding, since the position of the range data Q2 of the polygonal line code is known, the exponent part, that is, the range is reproduced by looking at the 3 bits after the polarity bit P. Since the position of the mantissa data D2 is also known, the range data Q
Run length L0 of linear code (16 bits) by 2
Of the inversion bit T0 and the mantissa data D2
Attach. When the code length W0 of the original data is less than, a fixed value is used. Decoding processing is performed according to such a procedure. When the encoding and decoding described above are performed, the precision of expression is changed by rounding the information. Rounding of information occurs due to shift operations. Since the run length L0 is the same for "6" and "7 or more", there are seven types of shifts. Therefore, there are 14 types of positive and negative, but the same shift operation is performed in the central region with positive and negative, so the input / output characteristics of the entire region can be represented by 13 types of broken lines. Figure 6
In each region of (b), the expression precision, that is, the resolution, changes depending on the run length L0. When the run length L0 is "0" to "6", the resolution is 10 to 16 bits in terms of the number of bits converted into word length. The contents described above are summarized in (Table 1).

[Table 1] FIG. 7 is a characteristic diagram of a decompression code obtained by decoding and decompressing a compression code using a polygonal line code. The horizontal axis of the figure shows the input level, and the vertical axis shows the instantaneous S / N ratio. However, as for the instantaneous S / N ratio, about 2 dB for improving the S / N ratio of the rectangular wave is omitted for simplicity. In the figure, Z1 indicates the characteristic of 13 polygonal line code (12 bits). Z1 will be described.
The input level of 0 dB is based on the maximum amplitude of a sine wave that can be represented by a code. The straight line code "P1111111111
"11111" corresponds to this. In the input level range of 0 dB to -6 dB, the straight line code is "P1 *** ...".
Since the run length L0 is “0” from (Table 1), the resolution is 10 bits. The quantization noise of 10-bit data is -60 dB, with -6 dB per bit. Therefore, the instantaneous S / N ratio of the input level 0 dB to -6 dB is 60 dB to 54 dB. Next, in the input level range of -6 dB to -12 dB, the linear code is "P01 *** ...", and since the run length L0 is "1" from (Table 1), the resolution is 11 bits. The quantization noise of 11-bit data is -66 dB. Therefore, the instantaneous S / N ratio is 60 dB to 54 dB. Similarly, the instantaneous S / N in the area up to -96 dB
Z1 is the characteristic curve obtained by obtaining the ratio. Here, at an input level of −36 dB or less, the characteristic curve is the same as that of a linear code (16 bits).

However, in the above-mentioned conventional method, the instantaneous S / N ratio is only 60 dB at the best, and the instantaneous S / N ratio is the same as the linear code (16 bits) at a finer input level. descend. In a commercial digital recording / reproducing apparatus that requires a margin (headroom) for excessive input, the average input level is set low and recording is performed, so that the dynamic range that is practically used is narrowed, and Since the instantaneous S / N ratio becomes insufficient in a considerable range, there is a problem in that deterioration is increased in editing in which amplification and attenuation are repeated.
As described above, this problem also applies to linear codes (16 bits). Preferably, the hearing detection limit is about 48 dB
The above-mentioned instantaneous S / N ratio characteristic is input level 0 to -96d
A wide range of B is required to be satisfied. As an improved method therefor, it is possible to easily consider a method of extending the conventional 13-line code (12 bits) to 16 bits.
(Table 2) shows straight line code (20 bits) to broken line code (1
6 bit) is shown.
Since the word length of the mantissa data D2 in Table 2 is 12, the resolution is increased by 4 bits as a whole.

[Table 2] The characteristic diagram of the decompression code that decodes and expands the compression code using the polygonal line code (16 bits) shown in (Table 2) is Z in FIG.
It becomes like 2. With this improved method, the instantaneous S / N ratio is improved by 24 dB over the entire range, but the input level is -72.
The instantaneous S / N ratio is 48d in the range of dB to -96dB.
It is B or less, which is insufficient. Therefore, if this signal is amplified by an amplifier, there is a problem that the quantization noise is auditorily detected. The present invention solves the above-mentioned problems of the prior art, and is about 48d which is the hearing detection limit of the instantaneous S / N ratio in a wide range of input level 0 to -96dB.
The purpose is to secure B or higher.

In order to achieve this object, a run length 1 / n compression floating code method of the present invention uses a linear code as the original data, that is, continuous data Q0 in which bits of a predetermined logic are continuous in the higher order. And a linear code composed of an inversion bit T0 that breaks the continuity of the continuous data Q0 and a lower data D0 after the inversion bit T0 is encoded to obtain compressed data, that is, a run length of the continuous data Q0. The compressed continuous data Q1 obtained by compression, the inversion bit T1 that breaks the continuity of the compressed continuous data Q1, the compressed residual data C1 representing the residual F1 generated when compressing the run length, and the lower data D0. Mantissa data D1 rounded and obtained
It is converted into compressed data composed of and. However, when the run length of the continuous data Q0 is L0, the run length of the compressed continuous data Q1 is L1, and n is an integer of 2 or more, L1 = int (L0 / n) F1 = L0 mod n.

According to the present invention, the run length L0 of the original data is compressed to about 1 / n by the above method. The exponent part, that is, the range is specified by the compressed continuous data having the compressed run length, the inversion bit, and the compression remainder data.
The exponent part or range has a function of self-expanding according to the run length of the original data. That is, when the run length is small, the exponent part or range is specified with a small number of bits, and when the run length is large, the exponent part or range is subdivided, and the exponent part or range is specified with a large number of bits. When the word length of the entire code is finite and has a predetermined value, the number of bits assigned to the mantissa is changed according to the run length. By these actions, the space represented by the compression code is expanded, and the expression precision can be maximized.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1A is a conceptual diagram showing the configuration of a run length 1/2 compression floating code in the first embodiment of the present invention, and FIG. 1B is a diagram showing a concrete example for explaining the conversion processing. Is. In the figure, the same parts as those in the conventional example are given the same symbols. The procedure of compression (encoding) will be described below with reference to the drawings. First
In the embodiment of the present invention, a straight code (30 bits) is used as the run length 1 /
This is a coding method that compresses and expands into a 2-compression floating code (16 bits). The compressed continuous data Q1 is the continuous data Q
It is continuous data having the length of the run length L1 obtained by dividing the run length L0 of 0 by 2 to obtain an integer. That is, L1 = int (L0 / 2). If the remainder term of the integer division is the compression remainder F1, then F1 = L0 mod 2. The inversion bit T1 is an inversion bit that breaks the run of the compressed continuous data Q1. The compression residue data C1 is a complement representation of the compression residue F1. In addition, the mantissa data D1 is higher-order partial data of the data D0. The run length 1/2 compression floating code is arranged in the order of the polarity bit P, the compressed continuous data Q1, the inversion bit T1, the compression remainder data C1, and the mantissa data D1. Below, Figure 1
Based on (b), the run length L0 is "0" or "2".
The case of "8" will be described. When L0 = 0, L1 and F1 are: L1 = int (0/2) = 0 F1 = 0 mod 2 = 0. Since the run length L1 is "0", there is no compressed continuous data Q1. The compression residue data C1 is “1”. The data D0 is 30 bits, and the upper 13 bits "ABCDEFGHIJKLM" of the data D0 is the mantissa data D1. In the run length 1/2 compression floating code, the polarity bit P, the inversion bit T1, the compression remainder data C1 and the mantissa data D1 are arranged in this order, and “P11ABC
DEFGHHIJKLM ". Similarly, the following is obtained. When L0 = 1, L1 = int (1/2) = 0 F1 = 1 mod 2 = 1 Therefore, there is no compressed continuous data Q1 and compressed remainder data C1
Is "0". Upper 13 bits of data D0 "ABC
"DEFGHIJKLM" is the mantissa data D1. Run length 1/2 compression floating code is "P10AB
CDEFGHIJKLM ". When L0 = 2, L1 = int (2/2) = 1 F1 = 2 mod 2 = 0 Therefore, the continuous data Q1 is “0” and the compression remainder data C1 is “1”. Upper 12 bits of data D0 "A
BCDEFGHIJKL ”is the mantissa data D1. Run length 1/2 compression floating code is "P011
ABCDEFGHIJKL ". When L0 = 3, L1 = int (3/2) = 1 F1 = 3 mod 2 = 1 Therefore, the continuous data Q1 is “0” and the compression remainder data C1 is “0”. Upper 12 bits of data D0 "A
BCDEFGHIJKL ”is the mantissa data D1. Run length 1/2 compression floating code is "P010
ABCDEFGHIJKL ". Further, when L0 = 20, L1 = int (20/2) = 10 F1 = 20 mod 2 = 0 Therefore, the continuous data Q1 is “0000000000” and the compression residue data C1 is “1”. The upper 3 bits “ABC” of the data D0 are the mantissa data D1. Therefore, the run length 1/2 compression floating code is "P
000000000011 ABC ". Further, when L0 = 28, L1 = int (28/2) = 14 F1 = 28 mod 2 = 0 Therefore, the continuous data Q1 is “0000000000000000.
The compression remainder data C1 is "1". There is no data D0 and no mantissa data D1. Therefore, the run length 1/2 compression floating code is "P00000.
0000000001 ". In this way, the linear code (30 bits) is compressed into the run length 1/2 compression floating code (16 bits). Next, a decompression (decoding) procedure will be described. In the decoding procedure, the polarity bit P is removed from the compressed data, and the run length L1 is obtained from the compressed continuous data Q1 and the inversion bit T1. Further, the compression residue data C1 (1 bit) and the mantissa data D1 immediately after the inversion bit T1 are obtained. A compression residue F1 is obtained by inverting the compression residue data C1. From these, the run length L0 of the continuous data Q0 of the original data is obtained from L0 = 2 * L1 + F1. The continuous data Q0 is restored by connecting "0" having the length of the run length L0. An inversion bit T0 is added after the continuous data Q0, and then mantissa data D1 is added. The polarity bit P is added to the head to form the original data. When this code length is less than W0, a fixed value is assigned to the lower order of the mantissa data D1 to set the code length to W0. Decoding processing is performed according to this procedure. Hereinafter, some cases of the run length L1 will be specifically described with reference to FIG. When L1 = 0 and F1 = 0, since L0 = 2 * 0 + 0 = 0, there is no continuous data Q0. The mantissa data D1 is 13 bits and is "ABCDEFGHIJKLM".
At this time, the original data is the polarity bit P and the inversion bit T0.
And mantissa data D1 are arranged, and “P1ABCDEF
GHIJKLM ******************** ”. "*** ..." is a fixed value "011 ..."
Shall be satisfied. Similarly, L1 = 10 and F1
When L = 0, L0 = 2 * 10 + 0 = 20, so the continuous data Q0 is “0000000000.
0000000000 "and mantissa data D1 is" AB
C ". At this time, the original data is obtained by arranging the polarity bit P, the continuous data Q0, the inversion bit T0, and the mantissa data D1 into “P000000000000000000”.
001ABC ******* ”. In this way, the linear code (30 bits) is decoded and expanded from the run length 1/2 compressed floating code (16 bits). In the same manner as above, the results summarized in all cases are shown in (Table 3).

[Table 3] In Table 3, the linear code (30 bits) is a folded binary code, and the floating code is a folded run length 1/2 compression floating code. Run length L0, run length L1 and resolution column is 1
It is a decimal number. The expression accuracy, that is, the resolution of the decompressed code obtained by decoding and decompressing the compressed code is determined by the rounding accuracy of the linear code in the linear code, and changes depending on the run length L0. From (Table 3), according to the first embodiment, a precision of up to 30 bits to 15 bits can be obtained. Here, the instantaneous S / N ratio in the case where a linear code is input, compression coding is performed using a run length 1/2 compression floating code, and this code is decoded and decompressed to output a decompressed code will be described. However, for simplification, the instantaneous S / N ratio omits the S / N ratio improvement (about 2 dB) of the rectangular wave. The input level is based on the maximum sine wave amplitude that can be represented by a sign (0 dB)
To The straight line code “P1111 ...” Corresponds to this. The range of the input level from 0 dB to -6 dB is "P1ABC ..." With a linear code, and the resolution is 15 bits from (Table 3). The quantization noise of 15-bit data is -90 dB, where -6 dB per bit. Therefore, the instantaneous S in the area of the input level 0 dB to -6 dB
The / N ratio is 90 dB to 84 dB. Input level-
In the range of 6 dB to -18 dB, the linear code "P01
“ABC ...” Or “P001ABC ...”, and the resolution is 16 bits from (Table 3). Since the quantization noise of 16-bit data is -96 dB, the instantaneous S / N ratio is 90 dB to 78 dB. In the input level range of -18 dB to -30 dB, the linear code "P
0001ABC ... "or" P00001ABC.
··· ”, and the resolution is 17 bits from (Table 3). Since the quantization noise of 17-bit data is -102 dB, the instantaneous S / N ratio is 84 dB to 72 dB. Similarly, the instantaneous S / N ratio is obtained in the input level region up to -180 dB. FIG. 3 is an instantaneous S / N ratio characteristic diagram of a restored code obtained by decoding and expanding a compressed code using a run length 1/2 compression floating code. The horizontal axis of the figure shows the input level, and the vertical axis shows the instantaneous S / N ratio. The characteristic curve X1 in the figure represents the instantaneous S / N ratio of the run length 1/2 compressed floating code shown in (Table 3). From the figure, it can be said that the characteristic curve X1 is substantially above the hearing detection limit of 48 dB of the instantaneous S / N ratio in the input level range of 0 dB to -96 dB. Compared to the conventional linear code (16 bits), an instantaneous S / N ratio improvement of about 42 dB is achieved at an input level of -96 dB. The improvement of the instantaneous S / N ratio works over a wide range from the input level of -18 dB to -180 dB. The instantaneous S / N ratio deteriorates in the input level range of 0 dB to -6 dB, but the deterioration is 6 dB at most, and the instantaneous S / N ratio is still 90 dB to 84 dB. This area is reserved as a margin and is a range that is rarely used normally, so there is no problem at all. Further, also in comparison with the polygonal line code (16 bits) shown in the characteristic curve Z2 of FIG. 7, 6d in the input level range of 0 dB to -12 dB.
There is an improvement of B, and an improvement of 18 dB is obtained at -96 dB. As described above, according to the first embodiment, the quantization noise can be made smaller than the auditory sense limit over a wide range up to the input level -96 dB even if the signal of the minute level is amplified. Therefore, the headroom can be sufficiently secured and the recording level can be set, and it can be amplified arbitrarily in the later editing. Further, in order to improve the problem of the quantization noise of the unique spectrum that occurs in a signal of a minute level, it is possible to eliminate the processing such as adding white noise (dither) to disperse the noise spectrum. With current technology, it is difficult to amplify a wide range of signals from input level 0 dB to -180 dB, or to digitize it with an AD converter, but due to future development of functional elements and circuit technology utilizing high temperature superconductivity, The effect of the present invention can be maximized. In addition, in Table 3, the increase characteristic of the run length 1/2 compression floating code is considered. In the first embodiment, since the compression residue data C1 is the complement of the compression residue F1, the numerical value of the upper range is always larger than the lower range at the boundary of the range. Therefore, the run length 1/2 compressed floating code of the first embodiment has a monotonically increasing characteristic. Even if the compressed data is erroneously reproduced as a linear code (16 bits), the level inversion or jump does not occur with the monotonically increasing characteristic, which is advantageous. This action converts the compression residue data C1 into the compression residue F.
Although it is based on the one's complement, it is not always necessary to use the complement and other expressions may be used. As described above, it is difficult with the current technology to amplify a wide range of signals from an input level of 0 dB to −180 dB or to digitize it with an AD converter. Therefore, when it is not possible to prepare 30 bits of the linear code, the linear code may be shortened. Table 4 shows a code decoding method for compressing and expanding a linear code (28 bits) into a run length 1/2 compression floating code (16 bits).

[Table 4] In this case, the run length L0 is cut off at a maximum of 26. When the run length L0 = 26, the inversion bit T0 is omitted,
The remaining data is D0. The encoding is the same, but the compression remainder data C1 is omitted and mantissa data D1, that is, D0 is arranged instead. Due to the termination of the run length L0, the instantaneous S / N ratio characteristic becomes the characteristic curve X2 in FIG. Input level-162dB to 0dB
The range is the same as X1 in the figure. Further, in the same manner, the linear code may be shortened to 24 bits. Table 5 shows a code decoding method for compressing and expanding a linear code (24 bits) into a run length 1/2 compression floating code (16 bits).

[Table 5] In this case, the run length L0 is cut off at the maximum "18". When the run length L0 = 18, the inversion bit T0 is omitted and the remaining data is D0. Although the encoding is the same, it is known that the maximum value of the run length L1 is "9", so that the inversion bit T1 may be any value. Further, the compression remainder data C1 is omitted and mantissa data D1, that is, D0 is arranged in its place. Due to the termination of the run length L0, the instantaneous S / N ratio characteristic becomes the characteristic curve X3 in FIG. The input level is the same as X1 in the figure in the range of -114 dB to 0 dB. With any of the characteristic curves X1, X2, and X3 described above, the hearing detection limit of the instantaneous S / N ratio can be almost cleared in the range of the input level 0 dB to -96 dB. Next, the second aspect of the present invention
Embodiments will be described with reference to the drawings. Figure 2
(A) is a run length 1 in the second embodiment of the present invention.
It is a conceptual diagram which shows the structure of a / 4 compression floating code | symbol, and FIG.2 (b) is a figure showing the specific example for demonstrating conversion processing. In the figure, the same parts as those in the conventional example are given the same symbols. The procedure of compression (encoding) will be described below with reference to the drawings. In the second embodiment, the linear code (32
(Bit) to a run length 1/4 compression floating code (16 bits). The compressed continuous data Q1 is continuous data having a length of the run length L1 which is an integer obtained by dividing the run length L0 of the continuous data Q0 by "4". L1 = int (L0 / 4) When the remainder term of integer division is the compression remainder F1, F1 = L0 mod 4. The inversion bit T1 is an inversion bit that breaks the run length of the compressed continuous data Q1. The compression residue data C1 is the complement of the compression residue F1. The mantissa data D1 is partial data on the upper side of the data D0. The run length 1/4 compression floating code has a polarity bit P, compressed continuous data Q1, and inversion bit T.
1, the compression remainder data C1, and the mantissa data D1 are arranged in this order. Hereinafter, the case where the run length L0 is "0" to "28" will be described with reference to FIG. When L0 = 0, L1 = int (0/4) = 0 F1 = 0 mod 4 = 0 Since the run length L1 is “0”, there is no compressed continuous data Q1. The compression residue data C1 is “11”. The data D0 is 30 bits, and the upper 12 bits "ABCDEFGHIJKL" of the data D0 is the mantissa data D1. In the run length 1/4 compression floating code, the polarity bit P, the inversion bit T1, the compression remainder data C1 and the mantissa data D1 are arranged in this order, and “P111ABC
DEFGHHIJKL ". Similarly, when L0 = 1, L1 = int (1/4) = 0 F1 = 1 mod 4 = 1 Therefore, there is no compressed continuous data Q1 and compressed remainder data C1
Is "10". Upper 12 bits of data D0 "AB
"CDEFGHIJKL" is the mantissa data D1. Run length 1/4 compression floating code is "P110A
BCDEFGHIJKL ”. When L0 = 3, L1 = int (3/4) = 0 F1 = 3 mod 4 = 3 Therefore, the continuous data Q1 is “0” and the compression remainder data C1 is “00”. The upper 12 bits "ABCDEFGHIJKL" of the data D0 is the mantissa data D1. Run length 1/4 compression floating code is "P
100ABCDEFGHIJKL ". Further, when L0 = 20, L1 = int (20/4) = 5 F1 = 20 mod 4 = 0 Therefore, the continuous data Q1 is “00000” and the compression remainder data C1 is “11”. The upper 7 bits “ABCDEFG” of the data D0 is the mantissa data D1. Therefore, the run length 1/4 compression floating code is "P
00000111ABCDEFG ". Furthermore, when L0 = 23, L1 = int (23/4) = 5 F1 = 23 mod 4 = 3 Therefore, the continuous data Q1 is “00000” and the compression remainder data C1 is “00”. The upper 7 bits “ABCDEFG” of the data D0 is the mantissa data D1. Therefore, the run length 1/4 compression floating code is "P
00000100ABCDEFG ". In this way, the linear code (32 bits) is compressed into the run length 1/4 compression floating code (16 bits). Next, a decompression (decoding) procedure will be described. In the decoding procedure, the polarity bit P is removed from the compressed data, and the run length L1 is obtained from the compressed continuous data Q1 and the inversion bit T1. Further, the compression remainder data C1 (2 bits) and the mantissa data D1 immediately after the inversion bit T1 are obtained. A compression residue F1 is obtained by inverting the compression residue data C1. From these,
The run length L0 of the continuous data Q0 of the original data is obtained from L0 = 4 * L1 + F1. The continuous data Q0 is restored by connecting "0" having the length of the run length L0. An inversion bit T0 is added after the continuous data Q0, and then mantissa data D1 is added. The polarity bit P is added to the head to form the original data. When this code length is less than W0, a fixed value is assigned to the lower order of the mantissa data D1 to set the code length to W0. Decoding processing is performed according to this procedure. Hereinafter, some cases of the run length L1 will be specifically described with reference to FIG. When L1 = 0 and F1 = 0, L0 = 4 * 0 + 0 = 0, so there is no continuous data Q0. The mantissa data D1 is 12 bits and is "ABCDEFGHIJKL". At this time, the original data is obtained by arranging the polarity bit P, the inversion bit T0, and the mantissa data D1 into “P1ABCDEFG”
HIJKL ************** ”.
Note that "*** ..." is a fixed value of "1000 ...". Similarly, when L1 = 5 and F1 = 0, since L0 = 4 * 5 + 0 = 20, the continuous data Q0 is “0000000000”.
0000000000 "and mantissa data D1 is" AB
CDEFG ". At this time, the original data is obtained by arranging the polarity bit P, the continuous data Q0, the inversion bit T0, and the mantissa data D1 into “P0000000000000000”.
0000001ABCDEFG *** ”. In this way, the linear code (32 bits) is decoded and expanded from the run length 1/4 compression floating code (16 bits). In the same manner as above, the results summarized in all cases are shown in (Table 6).

[Table 6] From Table 6, according to the present embodiment, the accuracy of up to 32 bits to 14 bits can be obtained. Here, the instantaneous S / N ratio in the case where a linear code is input, compression coding is performed using a run length 1/4 compression floating code, and this code is decoded and expanded to output a decompressed code will be described.
The range of the input level from 0 dB to -6 dB is "P1ABC ..." With a straight line code, and the resolution is 1 from (Table 6).
It is 4 bits. Quantization noise of 14-bit data is 1
It becomes -84 dB as -6 dB per bit. Therefore,
Instantaneous S / N in the input level range of 0 dB to -6 dB
The ratio is 84 dB to 78 dB. Input level-6d
In the range of B to -12 dB, the straight line code indicates "P01AB
C ... ”, and the resolution is 15 bits from (Table 6). Since the quantization noise of 15-bit data is -90 dB, the instantaneous S / N ratio is 84 dB to 78 dB. In the input level range of -12 dB to -18 dB, the linear code is "P001ABC ..." (Table 6).
More resolution is 16 bits. Since the quantization noise of 16-bit data is -96 dB, the instantaneous S / N ratio is 8
It becomes 4 dB to 78 dB. In the input level range of -18 dB to -30 dB, the linear code "P0001AB
C ... "or" P00001ABC ... ", and the resolution is 17 bits from (Table 6). Since the quantization noise of 17-bit data is -102 dB, the instantaneous S / N ratio is 84 dB to 72 dB. Similarly, the instantaneous S / N ratio is obtained in the input level region up to -180 dB. FIG. 4 is an instantaneous S / N ratio characteristic diagram of a restoration code obtained by decoding and expanding a compression code using a run length 1/4 compression floating code. Input level on the horizontal axis of the figure,
The vertical axis shows the instantaneous S / N ratio. The characteristic curve Y1 in the figure represents the instantaneous S / N ratio of the run length 1/4 compression floating code shown in (Table 6). The characteristic curve Y1 clearly exceeds the hearing detection limit of 48 dB of the instantaneous S / N ratio in the input level range of 0 dB to -96 dB. Input level compared to conventional linear code (16 bits)
At 96 dB, an instantaneous S / N ratio improvement of about 60 dB is achieved. The improvement of the instantaneous S / N ratio is the input level -18d.
It acts over a wide range from B to -196 dB. Although the instantaneous S / N ratio deteriorates in the input level range of 0 dB to -12 dB, it deteriorates at most 12 dB, and still has the instantaneous S / N ratio of 84 dB to 78 dB. Since this area is a range that is not often used as a margin, there is no problem at all. Also, in comparison with the polygonal line code (16 bits) shown in the characteristic curve Z2 of FIG. 7, there is an improvement in the input level range of −54 dB to −196 dB, and −9 dB.
At 6 dB, an improvement of 36 dB is obtained. Second like this
According to this embodiment, even if a signal having an input level of -144 dB, which is much smaller than that in the first embodiment, is amplified, the auditory detection of quantization noise is completely eliminated. Therefore, the headroom can be sufficiently secured and the recording level can be set, and the recording level can be arbitrarily amplified in the later editing. (Table 7) shows run length 1/29 with linear code (29 bits).
This figure shows a code decoding method for compressing and expanding a 4-compressed floating code (16 bits).

[Table 7] In this case, the run length L0 is cut off at a maximum of "20".
When the run length L0 = 20, the inversion bit T0 is omitted and the remaining data is D0. The encoding is the same, but the maximum value of the run length L1 is known to be "5", so the inversion bit T1 is omitted. Further, the compression residue data C1 may be any data. Run length L
By terminating 0, the instantaneous S / N ratio characteristic becomes the characteristic curve Y2 in FIG. Input level -126 dB to 0
In the range of dB, it is the same as Y1 in the figure. Further, in the same manner, the linear code may be shortened to 26 bits.
(Table 8) shows run length 1 /
This figure shows a code decoding method for compressing and expanding a 4-compressed floating code (16 bits).

[Table 8] In this case, the run length L0 is cut off at the maximum "16". When the run length L0 = 16, the inversion bit T0 is omitted and the remaining data is D0. The encoding is the same, but the run length L1 is known to have a maximum value of 4, so the inversion bit T1 is omitted. Furthermore, the compression residue data C1 may be any data. Run length L
By terminating 0, the instantaneous S / N ratio characteristic becomes the characteristic curve Y3 in FIG. Input level -102 dB to 0
In the range of dB, it is the same as Y1 in the figure. With any of the characteristic curves Y1, Y2, and Y3 described above, the hearing detection limit of the instantaneous S / N ratio can be completely cleared within the range of the input level 0 dB to -96 dB. Next, a third embodiment of the present invention will be described. In the third embodiment, the linear code (2
This is a coding method for compressing and expanding 4 bits) into a run length 1/4 compression floating code (8 bits). That is, the number of bits of the code in the second embodiment is changed. Therefore, the procedure of compression (encoding) and decompression (decoding) is the same as that of the second embodiment, and therefore its explanation is omitted. The results summarized for all exponents or ranges are shown in (Table 9).

[Table 9] From Table 9, according to the present embodiment, a precision of up to 24 bits to 6 bits can be obtained. FIG. 5 is an instantaneous S / N ratio characteristic diagram of a decompressed code obtained by decoding and expanding the compressed code using the run length 1/4 compressed floating code of the third embodiment. The horizontal axis of the figure shows the input level, and the vertical axis shows the instantaneous S / N ratio. The characteristic curve V1 in the figure represents the instantaneous S / N ratio of the run length 1/4 compression floating code shown in (Table 9). The characteristic curve Y1 clearly shows a characteristic of a substantially flat instantaneous S / N ratio in a much wider input level range as compared with the linear code (8 bits). Although it is less than the hearing detection limit of 48 dB of the instantaneous S / N ratio as it is, it is possible to make use of a vast input level range and high resolution by combining with a differential code. It can be used for redundant compression that uses the correlation between samples of image data. In the first to third embodiments described above, the linear code is a folded binary code, but other linear codes such as a 2'S complementary code and an offset binary code are also converted to each other. Alternatively, it can be applied in exactly the same manner only by changing a predetermined logical value. Also, n has been described only in the case of “2” and “4”, but n may be any integer of “2 or more”. In this case, since the number of cases of the compression residue varies depending on the value of n, it goes without saying that the word length of the compression residue data may be changed.

As described above, according to the present invention, the exponent is obtained from the compressed continuous data obtained by compressing the upper run length of the original data to about 1 / n, the inversion bit, and the residual data derived when compressing the run length. It constitutes a section or range. The exponent part, that is, the word length of the range changes according to the run length of the original data. With this operation, when the run length is small, the exponent part, that is, the range is configured with a small word length, and the word length of the mantissa part is increased, so that the expression precision can be maximized. Also, when the run length is large, the exponent part, or range, is configured with a large word length. Therefore, the exponent part, or range, can be subdivided and expressed to widen the input level range, and the space represented by the compression code can be expanded. To be done. As a result, both the accuracy and the input level range are maintained at or above the specified characteristics.When applied to digital recording / playback equipment for business editing, in particular, quantization noise will be heard even if mixing and dubbing involving amplification and attenuation are repeated. No detection, input level 0 to -96 dB
Over a wide range, it is possible to secure the hearing detection limit of the instantaneous S / N ratio of about 48 dB or more.

[Brief description of drawings]

FIG. 1A is a conceptual diagram showing a configuration of a run length 1/2 compression floating code according to a first embodiment of the present invention, and FIG. 1B is an explanatory diagram for explaining the same conversion process.

FIG. 2A is a conceptual diagram showing a configuration of a run-length 1/4 compression floating code in the second embodiment of the present invention, and FIG. 2B is an explanatory diagram showing a specific example for explaining the conversion process.

FIG. 3 is a run length 1 according to the first embodiment of the present invention.
S / N ratio characteristic diagram of decompressed code that is decompressed by decompressing compressed code using 1/2 compressed floating code

FIG. 4 is a run length 1 according to the second embodiment of the present invention.
Instantaneous S / N ratio characteristic diagram of decompressed code obtained by decoding and expanding compressed code using / 4 compressed floating code

FIG. 5 is a run length 1 according to the third embodiment of the present invention.
Instantaneous S / N ratio characteristic diagram of decompressed code obtained by decoding and expanding compressed code using / 4 compressed floating code

FIG. 6A is a conceptual diagram showing a configuration of a conventional 13-segment line code (12 bits), and FIG. 6B is an explanatory diagram showing a specific example for explaining the conversion method.

FIG. 7 is a characteristic diagram of a restoration code obtained by decoding and expanding a compression code using a conventional polygonal line code.

[Explanation of symbols]

 P Polarity bit Q0 Continuous data Q1 Compressed continuous data Q2 Range data L0 Continuous data Q0 run length L1 Compressed continuous data Q1 run length L2 Range data Q2 run length T0, T1 Inverted bit D0 Remaining part data D1, D2 Mantissa data W0 Code length of original data W1, W2 Code length of compressed data

Claims (12)

[Claims] 1. A straight line code as original data, that is, continuous data Q0 in which bits of a predetermined logic are continuous in the upper order, and an inversion bit T0 that breaks the continuity of the continuous data Q0.
And a low-order data D0 after the inversion bit T0 are encoded to encode compressed data, that is, compressed continuous data Q1 obtained by compressing the run length of the continuous data Q0, and the compressed continuous data. Inversion bit T1 that breaks the continuity of Q1 and compressed residue data C1 that represents the residue F1 that occurs when the run length is compressed.
And a run length 1 / n compression floating coding method for converting the lower data D0 into compressed data composed of mantissa data D1 obtained by rounding. However, when the run length of the continuous data Q0 is L0, the run length of the compressed continuous data Q1 is L1, and n is an integer of 2 or more, L1 = int (L0 / n) F1 = L0 mod n. 2. The run length 1 / according to claim 1, wherein the compressed data is arranged in the order of compressed continuous data Q1, inverted bit T1, compression remainder data C1, and mantissa data D1.
An n-compression floating code method. 3. A linear code as the original data, that is, a polarity bit P, continuous data Q0 in which bits of a predetermined logic are continuous except the polarity bit P, and inversion that breaks the continuity of the continuous data Q0. A straight line code composed of a bit T0 and lower data D0 after the inversion bit T0 is encoded to obtain compressed data, that is, a compression bit, that is, a polarity bit P and a continuous length obtained by compressing the run length of the continuous data Q0. Data Q1 and the compressed continuous data Q1
To a compressed data composed of an inversion bit T1 that breaks the continuity of C, compressed residual data C1 representing a residual F1 generated when the run length is compressed, and mantissa data D1 obtained by rounding the lower data D0. A run length 1 / n compressed floating code method for converting. However, when the run length of the continuous data Q0 is L0, the run length of the compressed continuous data Q1 is L1, and n is an integer of 2 or more, L1 = int (L0 / n) F1 = L0 mod n. 4. The run length 1 / n compression floating code method according to claim 1, wherein the compression residue data C1 is a complement of the compression residue F1. 5. The run length 1 / n compression floating coding method according to claim 1, wherein the lower bits of the data D1 are deleted so that the compressed data has a predetermined code length W1. 6. The method according to claim 1, wherein L1 and F1 are L1 = int (Lmin / n) F1 = Lmin mod n, where Lmin is a smaller one of the predetermined limit length LS and the run length L0. Run length 1 / n compression floating code method. 7. The compressed data are, from the upper order, a polarity bit P,
4. The run length 1 / n compression floating coding method according to claim 3, wherein the compressed continuous data Q1, the inverted bit T1, the compression remainder data C1, and the mantissa data D1 are arranged in this order. 8. When the smaller of the predetermined limit length LS and the run length L0 is Lmin, L1 and F1 are L1 = int (Lmin / n) F1 = Lmin mod n, and the run length L0 is 7. The run length 1 / n compression floating code method according to claim 6, wherein when the length is larger than the limit length LS, the inversion bit T0 and the inversion bit T1 are omitted and left justified. 9. When the smaller of the predetermined limit length LS and the run length L0 is Lmin, L1 and F1 are L1 = int (Lmin / n) F1 = Lmin mod n, and the run length L0 is When it is larger than the limit length LS, the inversion bit T0 and the inversion bit T1 are omitted and left-justified, and the compression remainder data C1 representing the compression remainder F1
Is set to a predetermined fixed value.
An n-compression floating code method. 10. Compressed data, that is, compressed continuous data Q1 in which upper bits of a predetermined logic are continuous, and an inversion bit T1 that breaks the continuity of the compressed continuous data Q1.
And the compressed residual data C1 representing the residual F1 generated when the run length is compressed, and the compressed data composed of the mantissa data D1 are decoded, and the linear code serving as the original data, that is, the bits of a predetermined logical value at the higher order are continuous. Continuous data Q0
An inversion bit T0 that breaks the continuity of the continuous data Q0, and lower data D0 after the inversion bit T0
Run length 1 / n converted to a linear code composed of
Decoding method of compressed floating code. However, when the run length of the continuous data Q0 is L0, the run length of the compressed continuous data Q1 is L1, the remainder obtained from the compression remainder data C1 is F1, and n is an integer of 2 or more, L0 = L1 * n + F1 D0 = Let it be D1. 11. Compressed data, ie polarity bit P
A compressed continuous data Q1 in which bits of a predetermined logic are continuous at a higher level excluding the polarity bit P, and the compressed continuous data Q.
An inversion bit T1 that breaks the continuity of 1s, compressed remainder data C1 that represents a remainder F1 generated when compressing run lengths, and mantissa data D1.
A straight line code that is decoded and becomes original data, that is, a polarity bit P, continuous data Q0 in which bits of a predetermined logic are continuous in a higher order except the polarity bit P, and the continuous data Q
A decoding method of a run length 1 / n compression floating code for converting into a linear code composed of an inversion bit T0 that breaks the continuity of 0s and lower data D0 after the inversion bit T0. However, when the run length of the continuous data Q0 is L0, the run length of the compressed continuous data Q1 is L1, the remainder obtained from the compression remainder data C1 is F1, and n is an integer of 2 or more, L0 = L1 * n + F1 D0 = Let it be D1. 12. The decoded linear code has a predetermined code length W0.
Fixed value "100 ...
. "Or" 011 ... "or" 000 ... "or" 111 ... ", or deletes the lower bits. The run length 1 / n compression floating code according to claim 10 or 11. Decryption method.