JPH08111644A - Information encoding method and device, information decoding method and device, and information recording medium and information transmitting method - Google Patents

Abstract

PURPOSE: To perform an efficient encoding by making the shape of a forward conversion window function at the time of a forward spectrum conversion processing and the shape of an inverse conversion window function different. CONSTITUTION: An audio signal is inputted in an input terminal 100 and a band division is performed for the signal in a band division circuit 101 having the function of a forward spectrum conversion. The spectrum signal for which the band division is performed is decomposed into a normalizing coefficient and a signal to be normalized by the normalizing circuits 111 to 114 corresponding to each band for every certain time block. Each signal to be normalized is converted into signal to be normalized/ quantized by quantizing circuits 121 to 124 based on the quantizing accuracy information of a quantization deciding circuit 141. Each signal to be normalized/quantized, and each normalizing coefficient from the normalizing circuits 111 to 114 and each quantization accuracy information from the deciding circuit 141 successively become code strings by a multiplexor 131 and are outputted from a terminal 103. These outputted code strings are recorded in an optical disk, etc., and are transmitted by radio waves, etc.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information coding method and device for coding input digital data by so-called high efficiency coding, an information recording medium for recording the coded information, and the coded information recording medium. The present invention relates to an information transmitting method for transmitting information, and an information decoding method and apparatus for reproducing and decoding the transmitted or recorded coded information to obtain a reproduced signal.

[0002]

2. Description of the Related Art Conventionally, there are various techniques for high-efficiency coding of a signal such as audio or voice. For example, a signal on the time axis is divided into blocks in a predetermined time unit and the signal on the time axis for each block is Converting to a signal on the frequency axis (spectral conversion), dividing it into multiple frequency bands, and coding for each band, so-called transform coding, which is a block frequency band division method, audio signals on the time axis, etc. An example is so-called band division coding (sub-band coding: SBC), which is a non-blocking frequency band division method of dividing into a plurality of frequency bands and encoding without dividing into blocks. Further, a method of high efficiency coding in which the above band division coding and transform coding are combined is also considered, and in this case, for example, after performing band division by the band division coding, A signal for each band is spectrum-converted into a signal on the frequency axis, and each spectrum-converted band is encoded.

Here, as a band division filter used in the above-mentioned band division coding, there is a filter such as a so-called QMF (Quadrature Mirror filter), and this QMF filter is described in the document "Digital Coding of Speech in Subvans "(" Di
gital coding of speech in subbands "RECrochier
e, Bell Syst.Tech. J., Vol.55, No.8 1976). This QMF filter divides the band into two equal bandwidths, and is characterized in that so-called aliasing does not occur when the divided bands are combined later.

In addition, the document "Polyphase Quadrature Filters-New Band Division Coding Technology"("Po
lyphase Quadrature filters -A new subband coding t
echnique ", Joseph H. Rothweiler, ICASSP 83, BOSTO
N) describes a method of dividing the filter into equal bandwidths. This polyphase quadrature filter is characterized in that when a signal is divided into a plurality of bands of equal bandwidth, it can be divided at one time.

Further, as the above-mentioned spectrum conversion,
For example, an input audio signal is divided into blocks in a predetermined unit time (frame), and a discrete Fourier transform (DFT) or a discrete cosine transform (Discrete Cosine T) is performed for each block.
ransform: DCT), Modified Discrete Cosine T
ransform (MDCT) or the like to convert the time axis into the frequency axis. The above M
For DCT, refer to "Subband / Transform Coding Usi" in the document "Subband / Transform Coding Using Filter Bank Design Based on Time Domain Aliasing Cancellation".
ng Filter Bank Designs Based on Time Domain Aliasi
ng Cancellation, "JPPrincen ABBradley, Univ. o
f Surrey Royal Melbourne Inst. of Tech. ICASSP 198
7).

In this way, by quantizing the signal divided for each band by the filter and the spectrum conversion,
It is possible to control the band in which the quantization noise is generated, and it is possible to perform auditory and more efficient encoding by utilizing the properties such as the so-called masking effect. In addition, if the normalization is performed for each band, for example, with the maximum absolute value of the signal component in that band before performing the quantization here,
Furthermore, highly efficient encoding can be performed.

Here, as the frequency division width in the case of quantizing each frequency component divided into frequency bands, for example, a bandwidth considering human auditory characteristics is often used. That is, an audio signal may be divided into a plurality of bands (for example, 25 bands) with a bandwidth generally called a critical band in which the bandwidth increases as the frequency band increases. Also, when encoding the data for each band at this time, a predetermined bit allocation for each band, or
Coding is performed by adaptive bit allocation (bit allocation) for each band. For example, when the coefficient data obtained by the MDCT processing is encoded by the bit allocation, the MD of each block is
The MDCT coefficient data for each band obtained by the CT process is encoded with the adaptive allocation bit number. The following two methods are known as bit allocation methods.

For example, the document "Adaptive Transform Coding of Speech Signal"
s ", R. Zelinski and P. Noll, IEEE Transactions of Ac
coustics, Speech, and Signal Processing, vol.ASSP-
25, No. 4, August 1977), bit allocation is performed based on the signal size of each band. With this method,
Although the quantization noise spectrum becomes flat and the noise energy becomes the minimum, the actual noise feeling is not optimal because the masking effect is not used auditorily.

In addition, for example, the document "Critical band encoder-
Digital encoding of the auditory system's perceptual requirements "
("The critical band coder --digital encoding of
the perceptual requirements of the auditory syste
m ", MAKransner MIT, ICASSP 1980) describes a method of performing fixed bit allocation by obtaining the required signal-to-noise ratio for each band by using auditory masking. Even when the characteristic is measured with a sine wave input, the characteristic value is not so good because the bit allocation is fixed.

In order to solve these problems, all bits that can be used for bit allocation have a fixed bit allocation pattern predetermined for each small block and a bit allocation depending on the signal size of each block. A high-efficiency coding device is proposed, which is used by dividing into parts to be performed, and makes the division ratio dependent on the signal related to the input signal, and increases the division ratio into the fixed bit allocation pattern as the spectrum of the signal becomes smoother. Has been done.

According to this method, when energy is concentrated on a specific spectrum such as a sine wave input, a large number of bits are allocated to a block including the spectrum, thereby significantly improving the overall signal-to-noise characteristic. can do. In general, human hearing is extremely sensitive to a signal having a steep spectrum component. Therefore, improving the signal-to-noise characteristic by using such a method does not only improve the numerical value in measurement. Not only that, it is effective in improving the sound quality in terms of hearing.

Many other methods have been proposed for the bit allocation method. Further, if the model relating to hearing is further refined and the coding apparatus is improved, it is possible to realize more efficient coding in terms of hearing. It will be possible.

In addition, the international application PCT /
In the specification and drawings of the application filed on May 31, 1994, JP94 / 00880, a method is proposed in which a tonal component which is particularly important for auditory perception is separated from a spectral signal and is encoded separately from other spectral components. In this way, it is possible to efficiently encode an audio signal or the like at a high compression rate with almost no auditory deterioration.

When the DFT or DCT described above is used as a method for converting a waveform signal into a spectrum, M
When the conversion is performed with the time block consisting of M samples, M independent real number data are obtained. Here, in order to reduce the connection distortion between the time blocks, M1 samples are usually overlapped with the adjacent blocks on both sides, so that in the DFT or DCT, (M−M1) samples are averaged. Therefore, M pieces of real number data are quantized and encoded.

On the other hand, when the above-mentioned MDCT is used as a method of converting into a spectrum, from 2M samples which are overlapped with N times on both adjacent times,
Since M independent real number data are obtained, these are averaged, and in MDCT, M real number data are quantized and coded for M samples. In the decoding device, by adding the waveform elements obtained by performing the inverse transformation in each block from the code obtained using the MDCT in this way while interfering with each other,
The waveform signal can be reconstructed.

By the way, generally, by lengthening the time block for conversion, the frequency resolution of the spectrum is improved and the energy is concentrated on a specific spectrum component.
Therefore, DFT and DFT can be performed by using the MDCT in which the number of spectral signals obtained does not increase with respect to the number of original time samples, by performing the conversion with a long block length by overlapping the adjacent blocks by half. It is possible to perform encoding more efficiently than when using DCT. In addition, it is also possible to reduce the inter-block distortion of the waveform signal by allowing adjacent blocks to have a sufficiently long overlap.

[0017]

However, for example, when a transform such as MDCT that causes a waveform signal to interfere with adjacent waveform elements at the time of inverse transform is used, the spectrum transform / inverse transform is performed. There is a condition that should be satisfied by the window, and if this condition is not satisfied, it will not be correctly restored when the signal converted into the spectrum is inversely converted and returned to the time series signal.

Conventionally, in addition to the above constraint conditions, these window functions have been designed so that the forward conversion window function and the inverse conversion window function have the same shape.
For this reason, a window function that does not have a sufficiently smooth shape was used as the window function of the forward spectrum conversion, so that the concentration of the energy distribution becomes relatively low when converted to a spectrum, and as a result, a large number of It was necessary to encode the spectrum with high accuracy, and it was difficult to perform efficient encoding. Further, in particular, when the tone component is separated and encoded, it is possible to realize efficient encoding if the number of spectra to be separated as constituting the tone component is as small as possible. However, since sufficient frequency separation cannot be obtained, the number of spectra forming each tone component increases, and thus it is difficult to perform efficient coding.

Therefore, according to the present invention, information coding that enables efficient coding when using a transform such as MDCT that causes a waveform signal to interfere with adjacent waveform elements at the time of inverse transform is used. It is an object of the present invention to provide a method and apparatus, an information decoding method and apparatus corresponding thereto, and an information transmitting method for transmitting information similar to an information recording medium recording encoded information.

[0020]

The present invention has been made in view of such circumstances, and the information coding method and apparatus of the present invention performs forward spectrum conversion of an input signal using a forward conversion window function. Thus, at the time of the inverse transform, the forward spectrum transform process of forming a spectrum signal that causes the waveform elements to interfere with each other between adjacent blocks to form a waveform signal, and the output spectrum signal after the forward spectrum transform process is encoded. Encoding processing is performed, and the shape of the forward conversion window function used in the forward spectrum conversion processing at this time is different from the shape of the inverse conversion window function used in the inverse conversion. It is a thing. The input signal may be an acoustic signal, for example.

Further, the information decoding method and apparatus of the present invention include a decoding process for decoding an encoded spectrum signal, and a waveform element between adjacent blocks for the spectrum signal after the decoding process. Inverse spectrum conversion processing for performing inverse spectrum conversion to cause interference is performed, and the shape of the inverse conversion window function used in the inverse spectrum conversion processing at this time is used in the forward spectrum conversion processing for obtaining the spectrum signal. It is characterized in that the shape is different from the shape of the forward conversion window function. The output signal may be an acoustic signal, for example.

Further, in the information recording medium and the information transmitting method of the present invention, the input signal is subjected to forward spectrum conversion using the forward conversion window function, so that the waveform elements are interfered between adjacent blocks at the time of inverse conversion. While the output spectrum signal after the forward spectrum conversion process to form the spectrum signal that will be configured is encoded,
The shape of the forward transform window function used in the forward spectrum transform process is different from the shape of the inverse transform window function used in the inverse transform, and the encoded information is recorded or transmitted. It is a thing. The encoded signal can be, for example, an acoustic signal.

These information encoding method and apparatus of the present invention,
In the information decoding method and apparatus of the present invention, and the information recording medium and information transmitting method of the present invention, the maximum value of the inverse transform window function is a value that does not exceed twice the central value of the inverse transform window function. is there. Further, the window shape of the transient part of the forward conversion window function is obtained by dividing the one obtained by the equation (21) described later into halves and shifting it, and w ₁ (n )
Is a forward transform window function, q is a value near 1, and M is the number of output spectra. The forward spectrum transform is a modified discrete cosine transform, and the forward transform window function is symmetrical or asymmetrical. Further, the forward conversion window function includes a section in which the value is zero, or the value is non-zero in all sections. Furthermore, the process of encoding the output spectrum signal after the forward spectrum conversion process in the encoding process is a variable length encoding process, and includes a process of extracting and separating a specific frequency component and encoding.

That is, according to the present invention, when a transform such as a modified discrete cosine transform that causes a waveform signal to be formed by interfering with adjacent waveform elements at the time of inverse transform is used, the forward transform window function and the inverse transform are used. The shape of the window function is made different, and the constraint conditions to be satisfied by the window are satisfied, and a smooth shape forward spectrum conversion window is applied.

[0025]

Preferred embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block circuit diagram showing a configuration example of an information coding apparatus to which the information coding method of the present invention is applied.

In FIG. 1, the audio signal input to the encoding device through the input terminal 100 is a band division circuit 10 which also has a function as a forward spectrum conversion means.
The band is divided by 1. As the band dividing means in the band dividing circuit 101, a dividing means using a filter such as QMF described above may be used, or a means for grouping the spectrum obtained by spectrum conversion such as MDCT into each band may be used. Good. Also, once
It is also possible to use a means of performing spectrum conversion on a spectrum divided into several bands by a filter and grouping the spectrum components obtained by this into bands. Further, the width of each band resulting from these band divisions may be uniform, or may be non-uniform so as to match a so-called critical bandwidth based on human auditory characteristics. In the example of FIG. 1, the band is divided into four bands, but of course, this number may be further increased or decreased.

The spectrum signals band-divided by the band division circuit 101 are normalized circuits 111, 112, 113, 114 corresponding to respective bands for each time block.
Is subjected to normalization, and is decomposed into a normalization coefficient and a signal to be normalized, respectively. The respective normalized signals are quantized by the quantization circuits 121, 122, 1 based on the quantization accuracy information output from the quantization accuracy determination circuit 141.
It is quantized by 23 and 124, and is converted into a normalized / quantized signal, respectively. The quantization precision determination circuit 141 determines the quantization precision, that is, the bit allocation by utilizing the auditory masking effect in consideration of human auditory characteristics as described above. In FIG.
Each quantization circuit 12 from the quantization precision determination circuit 141
Of the quantization precision information to 1, 122, 123, and 124, the quantization precision information sent to the quantization circuit 122 is sent via the terminal 152, and the quantization precision information sent to the quantization circuit 123 is sent to the terminal 153. Through the quantization circuit 12
The quantizing precision information sent to each of the four channels is sent to the corresponding circuit via the terminal 154.

The quantization circuits 121, 122, 123,
Each of the normalized / quantized signals from 124, the normalization coefficients from the normalization circuits 111, 112, 113, 114, and the quantization precision information from the quantization precision determination circuit 141 is a multiplexer. The code sequence is sequentially made by 131 as described later, and this code sequence is output from the terminal 103. This code string is then recorded on a disk-shaped recording medium such as an optical disk, a magneto-optical disk, a magnetic disk, a tape-shaped recording medium such as a magnetic tape, or a recording medium such as a semiconductor memory such as an IC card, radio waves, or an optical cable. Is transmitted from a transmission system such as a signal cable including.

Here, in the example of FIG. 1, the quantization precision determination circuit 141 calculates the quantization precision information based on each signal band-divided by the band division circuit 101. It is also possible to calculate from the signal through the previous terminal 100, and each normalization circuit 1
It is also possible to calculate based on the normalization coefficient from 11,112,113,114. Further, the calculation in the quantization precision determination circuit 141 can be performed based on an auditory phenomenon such as a masking effect, but each of the quantization precision information is output via the multiplexer 131 as described above and then decoded. Since it is sent to the digitizer,
The auditory model used in this decoding device can be set arbitrarily.

On the other hand, FIG. 2 shows a block circuit diagram of a configuration example of an information decoding apparatus corresponding to the information encoding apparatus of FIG. 1 to which the information decoding method of the present invention is applied.

In FIG. 2, the code information (the code string) input to the terminal 201 of the decoding device of this configuration example is
It is sent to the demultiplexer 202, where it is separated and restored into the quantization precision information for each band, the normalization coefficient, and the signal to be normalized / quantized. The quantization accuracy information for each band, the normalization coefficient, and the normalized / quantized signal are sent to the signal component configuration circuits 211, 212, 213, 214 which also have a function as decoding means corresponding to each band. Then, a signal component is formed for each band. The signal components from the respective signal component configuration circuits 211, 212, 213 and 214 are synthesized by the band synthesizing circuit 221 which also has a function as an inverse spectrum converting means, and output from the terminal 251 as an audio signal.

Next, FIG. 3 shows the band division circuit 101 of FIG.
FIG. 4 shows a specific configuration of the forward spectrum conversion means when using the MDCT in FIG.
21 shows a specific configuration of the inverse spectrum conversion means when using IMDCT (inverse MDCT) in No. 21.

In FIG. 3, the terminal 301 is supplied with the signal from the terminal 100 of FIG. 1 or any of the signals separated by QMF. This signal is sequentially output from the waveform cutting circuit 3
02, forward conversion window circuit 303, and forward spectrum conversion calculation circuit 304. The waveform cutting circuit 30
In 2, the signal waveform supplied to the terminal 301 is cut out, the next forward conversion window circuit 303 calculates the equation (1), and the forward spectrum conversion calculation circuit 304 calculates the equation (2).

[0035]

(Equation 7)

[0036]

(Equation 8)

However, in the above equations (1) and (2), J
Is a block number, M is the number of output spectra, x (n) is an input waveform signal, X _J (k) is an output spectrum signal obtained for each block, and w ₁ (n) is a forward spectrum transform window function (forward transform window). Function).

In addition, in FIG.
The output signal of any of the signal component configuration circuits 211 to 214 is supplied. This signal is sequentially sent to the inverse spectrum transform calculation circuit 402, the inverse transform window circuit 403, and the adjacent block synthesis circuit 404. The inverse spectrum transform calculation circuit 402 calculates the formula (3), the inverse transform window circuit 403 calculates the formula (4), and the adjacent block synthesis circuit 404 calculates the formula (5).

[0039]

[Equation 9]

[0040]

[Equation 10]

[0041]

[Equation 11]

In the equations (3), (4), and (5), J is the block number, M is the number of output spectra, and X is the number of output spectra.
_J (k) is the input spectrum signal given for each block,
y (n) is an output waveform signal, and w ₂ (n) is an inverse spectrum conversion window function (inverse conversion window function).

Here, the fact that y (n) obtained from equations (1) to (5) matches x (n) means that if there is no information loss due to encoding, the original waveform signal Is correctly restored, which is an indispensable condition for the waveform signal encoding / decoding means to have sufficient performance.

Equation (6) is obtained from equations (2) and (3), but since equation (7) is established, equations (8) and (9) are established. , (10) is obtained.

[0045]

(Equation 12)

[0046]

(Equation 13)

[0047]

[Equation 14]

[0048]

(Equation 15)

[0049]

[Equation 16]

Further, from the equations (4), (5), and (10), (11)
Since the formulas are obtained, the formulas (12) and (13) are x (n) and y (n)
Are the necessary and sufficient conditions for the agreement.

[0051]

[Equation 17]

[0052]

(Equation 18)

[0053]

[Formula 19]

Further, as shown in the equations (14) and (15), if a condition that the forward transformation window function and the inverse transformation window function match and they are symmetric is added, (12)
Equation (13) is equivalent to equation (16).

[0055]

(Equation 20)

[0056]

[Equation 21]

[0057]

[Equation 22]

After all, the equations (14), (15), and (16) become x (n)
Is a sufficient condition for y and n (n) to match, and conventionally the window function of Eq. (17) that satisfies these conditions is MDCT.
It has been used as a forward transform window function and an inverse transform window function for.

[0059]

(Equation 23)

However, under the constraint condition of equation (16), the value of the forward conversion window function at n ₀ is M−1.
Since it is restricted by the value of the forward transform window function at −n _0, it was difficult to design the forward transform window function to have a sufficiently high frequency separation.

FIG. 5 shows an example of the result of the spectrum conversion when the window function of the equation (17) is used, where M = 64. FIG. 5A shows the shape of the window function, that is, the characteristic. A curve (here, the curve also includes a broken line for discrete values such as sample values), and FIG. 5B shows an input waveform signal obtained by sampling a sine wave,
FIG. 5C shows a spectrum signal obtained as a result of MDCT of this input waveform signal using the window function. In general, if the shape of the window function of the forward conversion is not sufficiently smooth, when the spectrum conversion is performed using this window function, the energy distribution of the spectrum is spread without increasing.
Therefore, for example, in the window function as shown in FIG. 5A, which approaches 0 relatively quickly toward both ends, the distribution of the spectrum when the function is used is not sufficiently small at the tail of the peak. Therefore, in order to accurately encode a signal having a spread spectrum distribution as shown in FIG. 5C, it is necessary to encode a large number of these spread spectrum signals with sufficient accuracy. It is not desirable for improving the coding efficiency.

Therefore, in the configuration example of the present invention, the forward conversion window function and the inverse conversion window function match (14).
Remove the condition of the expression, first give the forward transform window function a sufficiently high frequency separation, and set the inverse transform window function so that the relation between the forward transform window function and equations (12) and (13) is satisfied. To do.

For example, the symmetry condition of the equation (15) is defined by the equation (18),
Under the condition of symmetry in Eq. (19), w as in Eq. (20)
_{When 2} (n) is set, this satisfies the conditions of equations (12) and (13), so that the original waveform signal can be restored when forward transform and inverse transform are applied.

[0064]

[Equation 24]

[0065]

(Equation 25)

[0066]

(Equation 26)

In this case, the forward conversion window function is (16)
Since it can be set without being restricted by the constraint condition such as the formula, it is possible to make the frequency separation degree sufficiently high. It should be noted that making the window shape asymmetric is not effective in increasing the frequency separation, and the equation (18), (1
It is convenient that the window function is symmetric as shown in the equation (9) since the number of coefficient information can be halved.

FIG. 6 shows (21) as the forward conversion window function.
An example in which a function with q = 1 is used in the equation is shown as M = 64. FIG. 6A shows the shape of the forward transform window function, and FIG. 6B shows the inverse transform window function. shape,
6C shows an input waveform signal obtained by sampling a sine wave, and FIG. 6D shows a spectrum signal obtained as a result of MDCT conversion of the input waveform signal.

[0069]

[Equation 27]

The inverse transform window function is obtained by the equation (20) and is different from the forward transform window function.
In the example of FIG. 6, the skirt of the spectrum signal is narrowed and the frequency separation degree is increased, as compared with the example of FIG.

Further, FIG. 7 also shows an example in which the function of q = 2 in the equation (21) is used as the forward conversion window function, similarly M = 64. In the case of the example of FIG. 7, the base of the spectrum signal is narrower than that of the example of FIG. 6, but the maximum value of the inverse transform window function is a large value of 2 or more. When the value of the inverse transform window function is large in this way, the level of the quantization noise inversely transformed on the time axis is also amplified by this inverse transform window function, so it is preferable that the inverse transform window function has a large value. Absent. In particular, when the input signal originally has a relatively flat energy distribution on the frequency axis, the effect of improving the coding efficiency by the frequency separation of the forward transform window function is small, so the maximum value of the inverse transform window function is 2 or more. It is better not to set so that

However, as described above, since the auditory sense is more sensitive to the tone signal in which the energy is concentrated in a specific isolated frequency band, the inverse transform window function of the example of FIG. 6 is used. If the absolute value can be suppressed to about 1.2, it is better to set a forward conversion window function that improves the frequency separation degree as much as possible.

The increase in the frequency separation as described above has a great effect on improving the coding efficiency of a tone-like signal, and this effect is obtained by incorporating the method described below for coding. In some cases the effect is even greater.

In such a first method, a code having a relatively short length is assigned to a quantization value having a high frequency, and a code having a relatively long length is assigned to a quantization value having a low frequency. This is the case where a variable length code is used. That is, since many spectra are normally quantized to 0, a code having a relatively short length is assigned to the quantized value 0, so that frequency separation is increased and more spectra are quantized to 0. , The coding efficiency is high.

The second method is the above-mentioned international application PCT /
As described in the specification and drawings of JP94 / 00,880, this is a case in which a tonal component which is particularly important for auditory sense is separated from a spectral signal and is encoded separately from other spectral components. When the frequency separation increases and the number of spectra quantized to 0 increases, the number of spectrum signals regarded as constituting the tone component can be reduced, and each tone component can be expressed by a shorter code, so that the coding efficiency is improved. Get higher

In the above example, the whole conversion window function is given by the expression (21) for the sake of simplicity. However, such a restriction can be relaxed, so that this will be described below.

As shown in FIG. 8, for example, when quantization noise is added to a spectrum signal obtained by subjecting a waveform signal to forward spectrum conversion using the window function (transform window function) of FIG. 8A, When it is subjected to inverse spectrum conversion and returned to the waveform signal on the time axis again, its quantization noise spreads over the entire conversion block. Here, when the signal waveform SW rapidly increases in the middle of the conversion block as shown in FIG. 8B, in the interval where the original signal waveform SW is small, the quantization noise QN changes to the signal waveform SW.
However, simultaneous masking does not work, and it becomes a hearing loss as a pre-echo.

FIG. 9 is a diagram for explaining an example of a conventional technique devised to reduce the hearing loss caused by such pre-echo. Generally, for a quasi-stationary signal waveform, a longer transform block length concentrates energy on a specific spectral coefficient, resulting in higher coding efficiency, but the loudness of the sound changes rapidly. In the part, if the conversion block length is long, the above-mentioned pre-echo becomes a problem. Therefore, if the conversion block length is shortened in the portion where the loudness of the sound changes abruptly to shorten the pre-echo generation period sufficiently, so-called reverse masking by the original signal works, and the auditory disturbance is eliminated. In the method of FIG. 9, this is used to selectively switch the conversion block length according to the property of each part of the signal waveform.

As a modification of the method of FIG. 9, a method of inserting a block having an asymmetric window function between the short transform section and the long transform section as shown in FIG. 10 is also proposed. By doing so, a window function over the entire conversion unit can be obtained without setting both ends to 0 in the long conversion unit, so that the concentration degree of energy distribution in the case of spectrum conversion can be increased and the coding efficiency can be improved. it can.

Such a forward transform window function of MDCT can be generally expressed as having a shape as shown in FIG. 11, and in some cases, the high level part, the low level parts L1 and L2, etc. are omitted. Can be interpreted as In the present invention, it is the transient portions E1 and E2 in FIG. 11 that define the shape of the window function, and the shape of this portion divides the window of the shape given by equation (21) into halves. If so provided, then the above is included in the method of the invention.

The Japanese Patent Application No. 6-130 filed by the present applicant
As described in the specification and drawing of No. 17, the pre-echo can be prevented by gain control, and in such a case, it is not necessary to switch the block length, and therefore, it is configured only by the transient section. Different window functions can be used to increase the frequency separation of the spectral signal.

The shape and spectrum of the window function can also be obtained by multiplying the window function of forward conversion by a constant at the time of conversion in the coding means of the coding device and dividing by the constant at the time of inverse conversion in the decoding means of the decoding device. It is needless to say that the same relation as above is established for the spread of the above, and the above comparison is performed after correcting all such changes by a constant multiple with respect to the window function.

Next, FIG. 12 and FIG. 13 show the flow of encoding and decoding processes using the method of the present invention.

First, in FIG. 12, the block number J = 0 is set in step S1, the equation (1) is calculated in the next step S2, and the equation (2) is calculated in the next step S3. In step S4, the obtained spectrum signal is normalized / quantized, and in step S5, the normalization / quantization is performed.
Encode the quantized spectrum signal. In step S6, the block number J is incremented, and in step S7, it is determined whether or not to end. If No, the process returns to step S2, and if Yes, the process ends.

In FIG. 13, the block number J = 0 is set in step S11, and the normalized / quantized spectrum signal is decoded in step S12. In step S13, denormalization of the obtained spectrum signal is performed.
Inverse quantization is performed. Equation (3) is calculated in the next step S14, and equation (4) is calculated in the next step S15.
Further, in the next step S16, calculation of equation (5) is performed. Then, in step S17, the obtained waveform signal is output. In the next step S18, the block number J is incremented, and in step S19, it is determined whether or not to end, and N
If o, the process returns to step S12, and if Yes, the process ends.

Although the band division circuit 101 that directly performs MDCT conversion and the band synthesis circuit 221 that directly performs IMDCT conversion have been described above, of course, other band division and band synthesis processes will be described. It can also be applied to the case of.

For example, FIG. 14 shows a concrete example in which a band division filter 502 and forward spectrum conversion circuits 511 to 514 such as MDCT are combined as the band division means, and FIG. 15 shows IMDCT as the band synthesis means. Forward spectrum conversion circuits 611 to 614 and band synthesis filter 6
21 shows a specific example in the case of combining with 21.

First, in FIG. 14, the signal from the terminal 100 of FIG. 1 is supplied to the terminal 501. This signal is
Is divided into a plurality of bands by the band division filter 502,
The signals in each band are sent to the forward spectrum conversion circuits 511 to 514, respectively. Each forward spectrum conversion circuit 511-5
In FIG. 14, forward spectrum conversion processing using the same forward spectrum conversion window function (forward conversion window function) as described above is performed, and the spectrum-converted signals respectively correspond to the corresponding terminals 521 to 524 and the normalization circuit 11 of FIG.
1 to 114 will be sent.

Further, in FIG. 15, terminals 601 to 60
4 corresponds to the signal component configuration circuit 211 of FIG.
The signals from ˜214 are supplied and sent to the corresponding inverse spectrum conversion circuits 611 to 614, respectively. In each of the inverse spectrum conversion circuits 611 to 614, an inverse spectrum conversion window function (inverse conversion window function) similar to that described above is used.
Is used to perform the inverse spectrum conversion process, and the signals obtained by the inverse spectrum conversion are respectively subjected to the band synthesis filter 62.
1 is output to the terminal 622 after being combined here. The output signal from this terminal 622 is the terminal 2 in FIG.
It will be sent to 51.

Here, the forward spectrum conversion circuit 511 is used.
3 to 514, and the inverse spectrum conversion circuits 611 to 614 can have the configuration of FIG. 4, and the method of the present invention described above can also be applied to these cases.

The forward conversion window function determined by the equation (21) in the present invention includes all of the above.

FIG. 16 shows an example of a code (code string) obtained by applying the method of the present invention to a waveform signal. The code is recorded on a recording medium or transmitted to a transmission line. It is possible. In the case of this example, the waveform signal is subjected to MDCT conversion for each block, and the obtained spectrum signal is normalized and quantized for encoding. That is, in the example shown in FIG. 16, the code information of each block consists of quantization accuracy information, normalization coefficient information, and spectrum coefficient information.

The case where the MDCT is used as the spectrum conversion for causing the waveform elements to interfere with each other between the adjacent blocks at the time of the inverse conversion has been described above. By using the MDCT, it is convenient that the spectral transform calculated by the same formula for all blocks can be easily realized, but it is convenient. However, the spectral transform that causes waveform elements to interfere between adjacent blocks at the time of inverse transform is convenient. The present invention can be applied even when another conversion method is adopted as. Another example of such a spectral transform is, for example, the document "Analysis / Syntehsis Filter Bank Des"("Analysis / Syntehsis Filter Bank Des").
ign Based on Time Domain Aliasing Cancellation ",
J. Princen and ABBradley, IEEE Transactions on Ac
oustics, Speech and Signal Processing, VolASSP-34
No. 5, October 1986).

Furthermore, it is also possible to perform coding using the forward transform window function according to the present invention, not necessarily the spectral transform in which the waveform elements interfere with each other during the inverse transform.

Although the case where the method is applied to the acoustic waveform signal has been described above, the method of the present invention can be applied to other types of signals, for example, the image signal. It is possible. However, the effect of the method of the present invention is remarkable when it has a particularly sharp spectrum distribution, and in the case of an acoustic signal, particularly high accuracy encoding is required when it has such a spectrum distribution from the viewpoint of hearing. The method of the present invention can be used particularly effectively.
Further, it goes without saying that the method of the present invention can be applied not only when recording encoded information on a recording medium but also when transmitting information.

[0096]

As is apparent from the above description, according to the information coding method and apparatus, the information decoding method and apparatus, the information recording medium and the information transmitting method of the present invention, the modified discrete cosine transform (MDCT) is used. Even when the window function has to satisfy a predetermined constraint condition, as in (1), the energy concentration of the spectrum distribution can be increased, and efficient coding is possible.

[Brief description of drawings]

FIG. 1 is a block circuit diagram showing a schematic configuration of an information encoding device of a configuration example of the present invention.

FIG. 2 is a block circuit diagram showing a schematic configuration of an information decoding device of a configuration example of the present invention.

FIG. 3 is a block circuit diagram showing a specific configuration of forward spectrum conversion means.

FIG. 4 is a block circuit diagram showing a specific configuration of inverse spectrum conversion means.

FIG. 5 is a diagram showing an example of a result of spectrum conversion when M = 64 when the window function of expression (17) is used.

FIG. 6 is a diagram showing an example in which M = 64 when a function in which q = 1 in Expression (21) is used as a forward conversion window function.

FIG. 7 is a diagram showing an example where M = 64 when a function in which q = 1 in Expression (21) is used as a forward conversion window function.

FIG. 8 is a diagram showing a state in which quantization noise is added to a spectrum signal obtained by subjecting a waveform signal to forward spectrum conversion using a window function.

FIG. 9 is a diagram for explaining an example of a conventional window function devised to reduce a hearing loss caused by pre-echo.

FIG. 10 is a diagram for explaining an asymmetric window function between a short conversion unit and a long conversion unit.

FIG. 11 is a diagram for explaining the shape of a forward transform window function of MDCT.

FIG. 12 is a flowchart showing a processing flow of an information encoding method of the present invention.

FIG. 13 is a flowchart showing a processing flow of the information decoding method of the present invention.

FIG. 14 is a band division filter and MD as band division means.
It is a block circuit diagram which shows a specific example in the case of combining with forward spectrum conversion means, such as CT.

FIG. 15 is a block circuit diagram showing a specific example of a case in which a forward spectrum converting unit such as IMDCT and a band synthesizing filter are combined as the band synthesizing unit.

FIG. 16 is a diagram showing an example of a code obtained by applying the information coding method of the present invention to a waveform signal.

[Explanation of symbols]

 101 Band division circuit 111-114 Normalization circuit 121-124 Quantization circuit 131 Multiplexer 141 Quantization accuracy determination circuit 202 Demultiplexer 211-214 Signal component configuration circuit 221 Band synthesis circuit 302 Waveform cutout circuit 303 Forward conversion window circuit 304 Forward spectrum Transform calculation circuit 402 Inverse spectrum transform calculation circuit 403 Inverse transform window circuit 404 Adjacent block synthesis circuit 502 Band division filter 511-514 Forward spectrum conversion circuit 611-614 Inverse spectrum conversion circuit 621 Band synthesis filter

Claims (66)

[Claims] 1. An order of forming a spectrum signal that forms a waveform signal by interfering waveform elements between adjacent blocks at the time of inverse transformation by forward spectrum transforming an input signal using a forward transform window function. It has a spectrum conversion process and a coding process for coding the output spectrum signal after the forward spectrum conversion process, and the shape of the forward conversion window function used for the forward spectrum conversion process is an inverse transform used at the time of inverse transform. An information encoding method, which is different from the shape of the window function. 2. The maximum value of the inverse transform window function is
2. The information coding method according to claim 1, wherein the value does not exceed twice the value in the center of the inverse transform window function. 3. The window shape of the transient part of the forward conversion window function is obtained by dividing the window shape obtained by the following equation 1 into halves and shifting it. The information encoding method according to claim 1 or 2, wherein ₁ (n) is a forward transform window function, q is a value in the vicinity of 1, and M is the number of output spectra. [Equation 1] 4. The information coding method according to claim 1, wherein the forward spectral transform is a modified discrete cosine transform. 5. The information encoding method according to claim 1, wherein the forward transform window function is symmetric. 6. The information coding method according to claim 1, wherein the forward transform window function is asymmetric. 7. The forward conversion window function includes a section in which the value is zero.
The information encoding method according to any one of 1. 8. The forward transform window function has a non-zero value in all sections.
The information encoding method according to any one of 1. 9. The process according to claim 1, wherein the process of encoding the output spectrum signal after the forward spectrum conversion process in the encoding process is a variable length encoding process. The information encoding method according to any one of items. 10. The process of encoding the output spectrum signal after the forward spectrum conversion process in the encoding process,
The information coding method according to claim 1, further comprising a process of extracting and separating a specific frequency component and coding the extracted frequency component. 11. The information encoding method according to claim 1, wherein the input signal is an acoustic signal. 12. A decoding process for decoding an encoded spectrum signal and an inverse spectrum conversion process for performing an inverse spectrum conversion for interfering waveform elements between adjacent blocks with respect to the spectrum signal after the decoding process. And the shape of the inverse transform window function used in the inverse spectrum transform process is different from the shape of the forward transform window function used in the forward spectrum transform process for obtaining the spectral signal. Information decoding method. 13. The information decoding method according to claim 12, wherein the maximum value of the inverse transform window function is a value that does not exceed twice the value at the center of the inverse transform window function. 14. The window shape of the transient part of the forward conversion window function is obtained by dividing the window shape obtained by the following equation 2 into halves and shifting it. ₁ (n) is the forward transform window function, q
14. The information decoding method according to claim 12, wherein is a value in the vicinity of 1, and M is the number of output spectra. [Equation 2] 15. The information decoding method according to claim 12, wherein the forward spectrum transform is a modified discrete cosine transform. 16. The information decoding method according to claim 12, wherein the forward transform window function is symmetric. 17. The information decoding method according to claim 12, wherein the forward transform window function is asymmetric. 18. The information decoding method according to claim 12, wherein the forward transform window function includes a section in which a value is zero. 19. The information decoding method according to claim 12, wherein the forward transform window function has a non-zero value in all sections. 20. The encoded spectral signal is
The information decoding method according to any one of claims 12 to 19, wherein the information decoding method is variable-length coded. 21. A part of the coded spectrum signal is coded by extracting and separating a specific frequency component, and any one of claims 12 to 20. The information decoding method described in the item. 22. The information decoding method according to claim 12, wherein the output signal is an acoustic signal. 23. An order of forming a spectrum signal that forms a waveform signal by interfering waveform elements between adjacent blocks at the time of inverse conversion by performing forward spectrum conversion of an input signal using a forward conversion window function. The output spectrum signal after the spectrum conversion process is encoded, and the shape of the forward transform window function used in the forward spectrum transform process is different from the shape of the inverse transform window function used in the inverse transform, An information recording medium characterized by recording the coded information. 24. The information recording medium according to claim 23, wherein the maximum value of the inverse transform window function is a value that does not exceed twice the value at the center of the inverse transform window function. 25. The window shape of the transient part of the forward conversion window function is obtained by dividing the window obtained by the following equation 3 into halves and shifting it. ₁ (n) is the forward transform window function, q
The information recording medium according to claim 23 or 24, wherein is a value near 1 and M is the number of output spectra. (Equation 3) 26. The information recording medium according to claim 23, wherein the forward spectral transform is a modified discrete cosine transform. 27. The information recording medium according to claim 23, wherein the forward transform window function is symmetrical. 28. The information recording medium according to claim 23, wherein the forward transform window function is asymmetric. 29. The information recording medium according to claim 23, wherein the forward conversion window function includes a section in which a value is zero. 30. The information recording medium according to claim 23, wherein the forward conversion window function has a non-zero value in all sections. 31. The process according to claim 23, wherein the process of coding the output spectrum signal after the forward spectrum conversion process is a variable length coding process. Information recording medium. 32. The process according to claim 23, wherein the process of encoding the output spectrum signal after the forward spectrum conversion process includes a process of extracting and separating a specific frequency component for encoding. The information recording medium according to any one of the items. 33. The information recording medium according to claim 23, wherein the encoded signal is an acoustic signal. 34. An order of forming a spectrum signal that forms a waveform signal by causing waveform elements to interfere with each other between adjacent blocks at the time of inverse conversion by performing forward spectrum conversion of an input signal using a forward conversion window function. It has a spectrum conversion means and an encoding means for coding the output spectrum signal after the forward spectrum conversion, and the shape of the forward conversion window function used for the forward spectrum conversion is an inverse conversion window function used at the time of inverse conversion. An information encoding device having a shape different from that of. 35. The information encoding apparatus according to claim 34, wherein the maximum value of the inverse transform window function is a value that does not exceed twice the value in the center of the inverse transform window function. 36. The window shape of the transitional part of the forward transform window function is obtained by dividing the window obtained by the following equation 4 into halves and shifting, and w in the equation. ₁ (n) is the forward transform window function, q
The information coding apparatus according to claim 34 or 35, wherein is a value near 1 and M is the number of output spectra. [Equation 4] 37. The information coding apparatus according to claim 34, wherein the forward spectrum transform is a modified discrete cosine transform. 38. The information coding apparatus according to claim 34, wherein the forward transform window function is symmetric. 39. The information coding apparatus according to claim 34, wherein the forward transform window function is asymmetric. 40. The information encoding apparatus according to claim 34, wherein the forward transform window function includes a section in which a value is zero. 41. The information encoding apparatus according to claim 34, wherein the forward transform window function has a non-zero value in all sections. 42. The process of encoding the output spectrum signal after the forward spectrum conversion process in the encoding means,
The information encoding device according to any one of claims 34 to 41, which is a variable length encoding process. 43. The process of encoding the output spectrum signal after the forward spectrum conversion process in the encoding means,
The information encoding device according to any one of claims 34 to 42, further comprising a process of extracting and separating a specific frequency component and encoding the same. 44. The information encoding apparatus according to claim 34, wherein the input signal is an acoustic signal. 45. Decoding means for decoding the encoded spectrum signal, and inverse spectrum conversion means for performing an inverse spectrum conversion for interfering waveform elements between adjacent blocks with respect to the decoded spectrum signal. And the shape of the inverse transform window function used for the inverse spectrum transform is different from the shape of the forward transform window function used for the forward spectrum transform for obtaining the spectrum signal. Device. 46. The information decoding apparatus according to claim 45, wherein the maximum value of the inverse transform window function is a value that does not exceed twice the value at the center of the inverse transform window function. 47. The window shape of the transient part of the forward conversion window function is obtained by dividing the window obtained by the following equation (5) into halves and shifting, and w in the equation. ₁ (n) is the forward transform window function, q
47. The information decoding apparatus according to claim 45 or 46, wherein is a value near 1 and M is the number of output spectra. (Equation 5) 48. The information decoding apparatus according to claim 45, wherein the forward spectrum transform is a modified discrete cosine transform. 49. The information decoding apparatus according to claim 45, wherein the forward transform window function is symmetric. 50. The information decoding apparatus according to claim 45, wherein the forward transform window function is asymmetric. 51. The information decoding apparatus according to claim 45, wherein the forward transform window function includes a section in which a value is zero. 52. The information decoding apparatus according to claim 45, wherein the forward transform window function has a non-zero value in all sections. 53. The encoded spectral signal is
The information decoding device according to any one of claims 45 to 52, wherein the information decoding device is variable length coded. 54. Any one of claims 45 to 53, wherein a part of the coded spectrum signal is coded by extracting and separating a specific frequency component. The information decoding device according to the item. 55. The information decoding apparatus according to claim 45, wherein the output signal is an acoustic signal. 56. An order of forming a spectrum signal that forms a waveform signal by interfering waveform elements between adjacent blocks at the time of inverse conversion by subjecting an input signal to forward spectrum conversion using a forward conversion window function. The output spectrum signal after the spectrum conversion process is encoded, and the shape of the forward transform window function used in the forward spectrum transform process is different from the shape of the inverse transform window function used in the inverse transform, An information transmitting method comprising transmitting the coded information. 57. The information transmitting method according to claim 56, wherein the maximum value of the inverse transform window function is a value that does not exceed twice the value in the center of the inverse transform window function. 58. The window shape of the transient part of the forward conversion window function is obtained by dividing the window obtained by the equation (6) below into halves and shifting it. ₁ (n) is the forward transform window function, q
58. The information transmitting method according to claim 56 or 57, wherein is a value near 1 and M is the number of output spectra. (Equation 6) 59. The information transmitting method according to claim 56, wherein the forward spectrum transform is a modified discrete cosine transform. 60. The information transmitting method according to claim 56, wherein the forward transform window function is symmetric. 61. The information transmitting method according to claim 56, wherein the forward transform window function is asymmetric. 62. The information transmitting method according to claim 56, wherein the forward conversion window function includes a section in which a value is zero. 63. The information transmitting method according to claim 56, wherein the forward conversion window function has a non-zero value in all sections. 64. The process of encoding the output spectrum signal after the forward spectrum conversion process is a variable length encoding process, according to any one of claims 56 to 63. Information transmission method. 65. The process according to claim 56, wherein the process of encoding the output spectrum signal after the forward spectrum conversion process includes a process of extracting and separating a specific frequency component and encoding it. The method for transmitting information according to any one of the items. 66. The information transmitting method according to claim 56, wherein the encoded signal is an acoustic signal.