JPH08167247A - High-efficiency encoding method and device as well as transmission medium - Google Patents

Abstract

PURPOSE: To make it possible to execute high-efficiency encoding with good efficiency in effectively utilizing a sound hearing characteristic, to prevent the deterioration in sound quality at a low bit rate and to improve the sound quality at the same bit rate. CONSTITUTION: This high-efficiency encoding device breaks down input signals to plural frequency band components to provide signal components within plural two-dimensional blocks relating to time and frequency, determines the quantization coefft. expressing the characteristics of the signal components by every two-dimensional block, determines a bit distribution quantity in accordance therewith, compresses the information by quantizing the signal components therein by every two-dimensional block and outputs the compressed information together with the information compression parameters by every two-dimensional block. The device has a harmonic component detecting circuit 813 for detecting the harmonic relation possessed by the input signals and a bit fraction adjusting circuit 814 from a band-by-band energy calculating circuit 803 as constitution to determine the bit distribution quantity. The degree of concn. of the bit distribution is enhanced for the two-dimensional blocks in which the respective harmonic components are included at the time of determining the bit distribution quantity.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a high-efficiency coding method and device for bit-compressing a digital signal such as a digital audio signal, and a transmission medium (including a recording medium) for transmitting the compressed data, and more particularly to a processing block. The present invention relates to a high-efficiency coding method and device, and a transmission medium, which changes the bit allocation amount of a corresponding processing block according to the shape of energy for each.

[0002]

2. Description of the Related Art There are various techniques and devices for high-efficiency coding of audio or voice signals. For example, a time domain audio signal is divided into blocks for each unit time, and a signal on the time axis of each block is used. Is converted into a signal on the frequency axis (orthogonal conversion), divided into multiple frequency bands, and coded for each band, so-called transform coding method, which is a so-called transform coding method, time domain audio signal, etc. , Which is a non-blocking frequency band division method that encodes by dividing into a plurality of frequency bands without blocking each unit time (sub-band coding: S
BC) method etc. can be mentioned. Further, a method and apparatus for high efficiency coding in which the above band division coding and transform coding are combined are also considered, and in this case, for example, after performing band division by the above band division coding method. The signals in the respective bands are orthogonally transformed into signals in the frequency domain by the transform coding method, and the respective orthogonally transformed bands are encoded.

Here, as a band division filter used in the above-mentioned band division encoding method, for example, QMF (Q
There is a filter such as uadrature Mirror filter), which is, for example, the document "Digital coding of speech in subvans".
n subbands "RE Crochiere, Bell Syst.Tech. J., Vo
l.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 Quadratic
Filter-New Band Division Coding Technique "(" Polyphase
Quadrature filters -A new subband coding techniqu
e ", Joseph H. Rothweiler ICASSP 83, BOSTON)
Equal bandwidth filter partitioning techniques are described. 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 once.

As the above-mentioned orthogonal transform, for example, the input audio signal is divided into blocks in a predetermined unit time (frame), and the fast Fourier transform (F
FT), Discrete Cosine Transform (DCT), Modified DCT Transform (MDCT), etc. are used to transform the time axis into the frequency axis. This MDCT
For the paper "Subband / Transform Coding Using Fil".
ter Bank Designs Based on Time Domain Aliasing Can
cellation, "JPPrincen ABBradley, Univ. of Surr
ey Royal Melbourne Inst. of Tech. ICASSP 1987).

Further, as a frequency division width in the case of quantizing each frequency component divided into frequency bands, for example, there is a band division considering human auditory characteristics. That is, an audio signal may be divided into a plurality of bands (for example, 25 bands) with a bandwidth that is generally called a critical band and has a wider bandwidth in a higher band. Further, at the time of encoding the data for each band at this time, encoding is performed by predetermined bit allocation for each band or adaptive bit allocation for each band. For example, when the MDCT coefficient data obtained by the MDCT process is encoded by the bit allocation, the MDCT for each band obtained by the MDCT process for each block is performed.
Coding is performed on the coefficient data with an adaptive distribution bit number.

The following two methods and apparatuses are known as the above-mentioned bit allocation method and an apparatus therefor. For example,
Reference "Adaptive Transf.
ormCoding of Speech Signals ", IEEE Transactions of
Accoustics, Speech, and Signal Processing, vol.ASS
P-25, No.4, August 1977) allocates bits based on the signal size of each band. In addition, for example, the document "Critical Band Coding-Digital Coding for Perceptual Requirements of Auditory System"("The critical band cod
er --digital encoding of the perceptual requireme
nts of the auditory system ", MAKransner MIT, ICA
SSP 1980) describes a method and a device for performing fixed bit allocation by obtaining a necessary signal-to-noise ratio for each band by using auditory masking.

[0007]

By the way, in the above-mentioned conventional high-efficiency coding method and device, when a digital audio signal having a harmonic relationship, such as a reproduced sound of a musical instrument, is coded, it is important for hearing. It is difficult to improve the degree of concentration of bit allocation with respect to a quantized block including a large overtone component. Therefore, more adaptive bit allocation cannot be performed for the input signal, and particularly when the bit rate is low and the number of usable bits is small, it is not possible to obtain good sound quality in terms of hearing.

The present invention has been made in view of the above circumstances, and provides an efficient high-efficiency encoding method and apparatus and a transmission medium that make better use of auditory characteristics.
The purpose is to prevent deterioration of sound quality at a low bit rate and improve sound quality at the same bit rate.

[0009]

The high-efficiency coding method and apparatus of the present invention has been proposed in order to achieve the above-mentioned object, and decomposes an input signal into a plurality of frequency band components, And a signal component in a plurality of two-dimensional blocks regarding frequency are obtained, a quantized coefficient representing a feature of the signal component in the two-dimensional block is obtained for each of the two-dimensional blocks, and a bit allocation amount is determined based on the quantized coefficient. For each two-dimensional block, the signal component in the two-dimensional block is quantized and information is compressed, and the compression information for each two-dimensional block is output together with the information compression parameter for each two-dimensional block. When performing the adaptive bit allocation, the harmonic relationship of the input signal is detected, and each 2
The degree of concentration of the bit allocation amount is increased with respect to the dimensional block.

Here, when performing bit allocation adapted to the input signal, based on the spectrum information on the frequency axis of the input signal and the quantization coefficient representing the signal component in the two-dimensional block, the input signal is Detects the overtone relationship of.
Further, when detecting the above-mentioned overtone relationship, the spectrum information of all or part of the frequency band of the input signal and / or 2
Based on the quantized information representing the signal component in the dimension block, only the low-order harmonic component and / or the fundamental and the first harmonic overtone component are detected, and the frequency band of the other harmonic components based on these harmonic components. Can be calculated to detect the overtone relationship of all frequency bands. Further, when performing bit allocation adapted to an input signal, a two-dimensional block containing all or some harmonic components, all or some two-dimensional blocks containing even harmonic components, and odd-numbered The degree of concentration of bit allocation is increased for all or some of the two-dimensional blocks containing overtone components. Further, when the bit allocation amount adapted to the input signal is determined, when the bit allocation amount is small with respect to the predetermined bit rate, the bit is preferentially applied to all or some of the two-dimensional blocks including the overtone component. Is added to adjust the fraction of bits, or bits are preferentially added to all or some of the two-dimensional blocks including even harmonic components to adjust the fraction of bits. On the contrary, when determining the bit allocation amount adapted to the input signal, if the bit allocation amount is large with respect to the predetermined bit rate, priority is given to all or some of the two-dimensional blocks that do not include the overtone component. The bit is adjusted by deleting the bit, or the bit is adjusted by deleting the bit preferentially for all or some of the two-dimensional blocks that do not include the odd harmonic component.

Further, the transmission medium of the present invention includes not only the transmission medium but also the recording medium, and transmits or records the compressed information and the information compression parameter compression-encoded by the high efficiency encoding method or apparatus of the present invention. To do.

[0012]

According to the present invention, when bit allocation is performed, the harmonic relationship of the input signal is detected, and the degree of concentration of bit allocation to the two-dimensional block for quantization including each harmonic component is improved. , It is possible to perform efficient bit allocation by making better use of the auditory characteristics, and it is possible to perform efficient compression without lowering the block floating efficiency.

From the above, it becomes possible to obtain better sound quality at the same bit rate. Also,
It can be performed at a lower bit rate to obtain the same sound quality.

[0014]

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

In this embodiment, an input digital signal such as an audio PCM signal is subjected to band division coding (SBC), adaptive transform coding (ATC), and adaptive bit allocation (APC).
-High efficiency coding using each technique of AB). This technique will be described with reference to FIG.

In the concrete high-efficiency coding apparatus of the present embodiment shown in FIG. 1, the input digital signal is obtained by dividing the input digital signal into a plurality of frequency bands by a filter and performing orthogonal transformation for each frequency band. The frequency axis spectrum data is coded by adaptively allocating bits for each so-called critical band (critical band) in consideration of human auditory characteristics described later. At this time, a band obtained by further dividing the critical bandwidth is used in the high band. Of course, the non-blocking frequency division width by the filter or the like may be equal division width.

Further, in the embodiment of the present invention, the block size (block length of the orthogonal transform) is adaptively changed according to the input signal before the orthogonal transform, and the critical bandwidth (in the critical band unit or in the high band) is changed. The floating process is performed in a block composed of signals in a band in which the critical band) is further subdivided. The critical band is a frequency band divided in consideration of human auditory characteristics, and when the pure tone is masked by a narrow band noise of the same strength in the vicinity of the frequency of a pure tone, the noise of that pure tone is masked. It is the bandwidth that you have. The critical band has a wider bandwidth in a higher frequency range, and for example, the entire frequency band of 0 to 20 kHz is 25, for example.
Is divided into critical bands.

That is, in FIG. 1, an audio PCM signal of 0 to 22 kHz, for example, is supplied to the input terminal 10. This input signal is divided into a band of 0 to 11 kHz and a band of 11 to 22 kHz by a band dividing filter 11 such as a so-called QMF, and 0 to 11 kHz.
Similarly, the band signals are 0 to 5.5 kHz band and 5.5k to 11k by the band division filter 12 such as so-called QMF.
And the Hz band.

11k from the band division filter 11
The signal in the 22 kHz band is M, which is an example of an orthogonal transformation circuit.
DCT (Modified Discrete Cosine Transform) circuit 1
5.5k from the band division filter 12 described above.
The signal in the ~ 11 kHz band is sent to the MDCT circuit 14,
The signals in the 0 to 5.5 kHz band from the band division filter 12 are sent to the MDCT circuit 15 to be subjected to MDCT processing. Each MDCT circuit 13-
In block 15, block decision circuits 19 and 2 provided for each band
MDCT processing is performed based on the block size (orthogonal transform block size) determined by 0 and 21.

Here, the block decision circuits 19 to 21.
Specific examples of the block sizes in the MDCT circuits 13 to 15 determined by are shown in (A) and (B) of FIG. In addition, in FIG. 2A, when the orthogonal transform block size is long (orthogonal transform block size in the long mode).
2B shows a case where the orthogonal transform block size is short (orthogonal transform block size in short mode).
Is shown.

In the specific example of FIG. 2, the three filter outputs each have two orthogonal transform block sizes. That is, for a signal in the low frequency band of 0 to 5.5 kHz and a signal in the middle frequency band of 5.5 to 11 kHz, if the block length is long ((A) of FIG. 2), the samples in one block are sampled. The number of samples is 128, and when a short block is selected ((B) of FIG. 2), the number of samples in one block is a block for every 32 samples. On the other hand, for a signal in the 11 kHz to 22 kHz band on the high frequency side, in the case of a long block length ((A) of FIG. 2), the number of samples in one block is set to 256 samples, and a short block is selected. In this case ((B) of FIG. 2), the number of samples in one block is a block for every 32 samples. When a short block is selected in this way, the number of samples of orthogonal transform blocks in each band is set to be the same, the time resolution is increased in the higher frequency range, and the types of windows used for blocking are reduced.

Referring again to FIG. 1, the information indicating the block size determined by the block determination circuits 19 to 21 is:
It is sent to adaptive bit allocation coding circuits 16, 17, and 18 which will be described later, and is also output from output terminals 23, 25, and 27.

The spectrum data in the frequency domain or the MDCT coefficient data obtained by the MDCT processing in each of the MDCT circuits 13 to 15 is summarized by a so-called critical band or a band obtained by further dividing the critical band in the high band. And transmitted to the adaptive bit allocation encoding circuits 16-18.

In the adaptive bit allocation encoding circuits 16 to 18, according to the information of the block size and the number of bits allocated for each band obtained by dividing the critical band (critical band) or the critical band in the high band,
Each spectrum data (or MDCT coefficient data) is requantized (normalized and quantized).

Each of these adaptive bit allocation encoding circuits 16 to
The data encoded by 18 is output to the output terminals 22, 2
It is taken out via 4, 26. Further, the adaptive bit allocation encoding circuits 16 to 18 also have a scale factor indicating what kind of signal magnitude is normalized and bit length information indicating what bit length is quantized. I am looking for these, and these are also output terminals 22,
It is output from 24 and 26.

Further, the MDCT circuits 13 to 13 in FIG.
From the output of 15, by calculating the energy of each band obtained by further dividing the critical band in the above critical band (critical band) or high band, for example, calculating the square root of the root mean square of each amplitude value in the band. Desired. Of course, the scale factor itself may be used for subsequent bit allocation. In this case, the calculation of new energy is not required, and the hardware scale is saved. Further, instead of the energy for each band, it is also possible to use the peak value, the average value, etc. of the amplitude values.

Next, the adaptive bit allocation encoding circuits 16 to 1
8, a concrete method of the bit allocation will be described.

The operation of the adaptive bit allocation encoding circuit in this case will be described with reference to FIG. 3. The magnitude of the MDCT coefficient is obtained for each block, and the MDCT coefficient is input terminal 801.
Is supplied to. MDC supplied to the input terminal 801
The T coefficient is supplied to the energy calculation circuit 803 and the spectrum smoothness calculation circuit 808 for each band. The energy calculation circuit 803 for each band calculates the signal energy for each band obtained by further dividing the critical band or the critical band in the high band. The energy related to each band calculated by the energy calculation circuit 803 for each band is supplied to the energy-dependent bit allocation circuit 804, the spectrum smoothness calculation circuit 808, and the overtone component detection circuit 813 which is the overtone relationship detection means in the present invention. To be done.

The overtone component detection circuit 813 detects the frequency band related to the overtone of the input signal using the energy for each band from the energy calculation circuit 803 for each band.

Overtones will be described with reference to FIG. An overtone is a signal having periodicity, such as a musical sound, which is indicated by F in the figure.
It is a frequency component having an integral multiple relationship such as 0, 2F0, 3F0, 4F0, .... F0 is called the fundamental tone or pitch, 2F0 is called the second overtone, 3F0 is called the third overtone, and 4F0 is called the fourth overtone. That is, a component having a frequency n times that of the fundamental tone is called an nth harmonic. Further, these overtones are separated into even-numbered (2n) or odd-numbered (2n + 1) components (n = 1, 2, 3, ...) Of the frequency of the fundamental tone F0.
・ ・), And they are called even harmonics and odd harmonics, respectively.
For example, the performance sound of a clarinet of a musical instrument is characterized in that the intensity of even-order overtones is significantly smaller than that of odd-order overtones.

The overtone component detection circuit 813 uses the difference between the block floating coefficients of adjacent blocks as an index as a method of detecting the overtone component of the input signal. If the frequency band of a block in which the difference between the block floating coefficients is large is in an integral multiple relationship, it is highly likely that the block contains a harmonic component of the input signal.
In this method, the difference between the block floating coefficients is obtained instead of the MDCT coefficient, so the amount of calculation is reduced and the hardware scale can be reduced.

It is also possible to detect the overtone component by directly using the MDCT coefficient instead of the block floating coefficient. If this method is used, more precise detection becomes possible. In particular, this is an effective means when one block contains a plurality of overtone components.

As another method for detecting the overtone component of the input signal, a method of obtaining only the fundamental tone and the first overtone and predicting the frequencies of all the overtone components based on the frequency relationship between them can be considered. This utilizes the above-mentioned property of overtones that the frequencies of overtones have an integer multiple relationship. By using this method, it is only necessary to check the band containing the fundamental and the first overtone without checking the difference between adjacent block floating coefficients or MDCT coefficients in all bands, so the amount of calculation is reduced, It is possible to reduce the hardware scale.

Information on the band containing the overtone component detected by the overtone component detection circuit 813 is supplied to the energy-dependent bit distribution circuit 804 and the bit fraction adjustment circuit 814.

In the energy-dependent bit allocation circuit 804,
Bits depending on the energy of the input signal by using the band including the energy from the band-by-band energy calculation circuit 803, the total usable bit from the total usable bit generation circuit 802, and the overtone component from the overtone component detection circuit 813. In this embodiment, the distribution is performed using a certain ratio of 128 kbps. The higher the tonality of the input signal, that is, the larger the unevenness of the spectrum of the input signal, the more the ratio of this bit amount to the 128 kbps increases. In addition,
To detect the unevenness of the input signal, the sum of the absolute values of the differences between the block floating coefficients of adjacent blocks is used as an index. Then, for the obtained usable bit amount, bit distribution proportional to the logarithmic value of the energy of each band is performed. At that time, more bits are allocated to the band containing the overtone component than to the other bands. Explaining with reference to FIG. 5, SP in the figure is the spectrum of the input signal, B1 to B
Reference numeral 8 indicates a quantization block. As for the overtone component of the input signal in this figure, the fundamental tone is block B2, the second overtone is block B4,
The third overtone is included in the block B6, and the fourth overtone is included in the block B8. In the figure, L1 represents the level of the quantization noise in each block when bit allocation is performed without considering the overtone relationship, and L2 in the figure is each block when bit allocation is performed in consideration of the overtone relationship. Represents the level of quantization noise at. As can be seen by comparing L1 and L2 in the figure, when bit distribution is performed in consideration of the overtone component, the quantization noise level in the block including the overtone component is lowered, and the sound quality is good for hearing. It is possible to obtain

In addition, when bit distribution is performed in consideration of harmonic components, it is possible to divide into even-order harmonic overtones and odd-order harmonic overtones. As described above, the overtones are divided into even-order overtones and odd-order overtones. Among various input signals, for example, there is a property that the odd harmonics have a greater strength than the even harmonics, such as the playing sound of a clarinet of a musical instrument. When bits are distributed to such an input signal, by setting so that more bits are distributed to a block including odd-order overtones than even-order harmonics, it is possible to obtain a good sound quality in terms of hearing. May be possible. On the other hand, on the contrary, there is an input signal in which the even-numbered overtone has a greater strength than the odd-numbered overtone. In such an input signal, it may be possible to obtain a good sound quality by allocating more bits to a block including an even-order harmonic than a block including an odd-order harmonic.

Further, it is desirable to improve the degree of concentration of bit distribution with respect to all blocks containing each overtone detected by the overtone component detecting circuit 813. For example, a fundamental component and a lower harmonic component are included. It is also possible to obtain good sound quality by applying it to only some blocks such as only blocks and / or only even harmonics or odd harmonics.

The bit allocation calculation circuit 805, which depends on the permissible noise level for hearing, first determines the permissible noise amount for each critical band in consideration of so-called masking effect based on the spectrum data divided for each of the critical bands. Bits obtained by subtracting energy-dependent bits from the total available bits are distributed so as to give a perceptible noise spectrum to the.

The energy-dependent bit number and the hearing-acceptable noise level-dependent bit number thus obtained are added in the adder 806, and the bit fraction adjustment circuit 81 is added.
4 is supplied.

In the bit fraction adjusting circuit 814, the adder 8
0 in this embodiment, the total available bit from the available total bit generation circuit 802, and the band containing the harmonic component from the harmonic component detection circuit 813.
Fractional adjustment of bits is performed to match the bit rate of 28 kbps.

For example, the output of the adder 806 is 128 kb
In the case where the number of bits exceeds ps, the band that has the least effect on hearing, for example, a method of sequentially decreasing one bit from the block on the high frequency side toward the low frequency side, or the high frequency side There is a method of setting the number of allocated bits to 0 from the block, that is, performing band limitation. In the present embodiment, a method of preferentially decreasing each block from the high frequency side to the low frequency side by 1 bit is used for the block not including the overtone component, and the fraction adjustment is performed so that the bit rate becomes equal to 128 kbps. To do. Explaining with reference to FIG. 5, it is assumed that the number of bits exceeds 128 kbps by one bit. First, focusing on the block B9 on the highest frequency side, since this block B9 contains harmonic components, no bit reduction is performed. Next, focusing attention on the adjacent low-frequency block B8, this block B8
Since 8 does not include the overtone component, reduction by 1 bit is performed. At this point, the bit number becomes equal to the bit rate of 128 kbps, and the fraction adjustment is completed. Note that, for example, when the number of bits is sequentially reduced from the high frequency side to the low frequency side, and when the block B1 including the lowest frequency component is reached, it is necessary to further reduce the number of bits. Returning to the block B9 including the component, the same operation is repeated. It should be noted that there is a possibility that the number of bits may be reduced for a block including an overtone component, such as when the number of bits is significantly insufficient. In addition, a method of classifying overtones into odd-order overtones and even-order overtones, and preferentially decreasing one or more bits from the high frequency side for all or some of the blocks that do not include odd-order overtones or even-order overtones Can also be considered.

On the contrary, for example, the output of the adder 806 is 128
When the number of bits is less than kbps, there is a method of sequentially increasing the bit by 1 bit from a block on the low frequency side toward a high frequency side, for example, a band that most affects the hearing. In the present embodiment, a method of preferentially increasing by 1 bit from the low frequency side for a block containing a harmonic component is used.
Adjust to the same number of bits as 28 kbps. Explaining with reference to FIG. 5, it is assumed that the number of bits exceeds 128 kbps by one bit. First, focusing on the block B1 on the lowest frequency side, this block B1
Since 1 does not include a harmonic component, the number of bits is not increased. Next, focusing on the adjacent block B2 on the high frequency side, since this block B2 also does not include a harmonic component, the number of bits is not increased. Next, focusing on the adjacent high-frequency side block B3, since this block B3 contains a harmonic component (fundamental tone), it is increased by one bit. At this point, the bit number becomes equal to the bit rate of 128 kbps, and the fraction adjustment is completed. Note that, for example, when the number of bits is sequentially increased from the low frequency side to the high frequency side, and when it reaches the block B9 including the highest frequency band, if it is necessary to further increase the number of bits, the lowest frequency is again set. Returning to the block B1 including the band, the same operation is repeated. It should be noted that there is a possibility that the number of bits may be increased even for a block that does not include an overtone component, such as when the number of bits is excessive. In addition, a method of classifying overtones into odd-order harmonic overtones and even-order harmonic overtones, and preferentially increasing by 1 bit from the low-frequency side for all or some of the blocks containing odd-order harmonic overtones or even-order harmonic overtones Can also be considered.

As described above, by using the fractional bit adjustment method considering the overtone relation, it is possible to reduce the quantization noise in the band including the overtone component, and it is possible to obtain a good sound quality in hearing. In particular, when the output of the adder 806 has a bit rate exceeding 128 kbps, it is less likely that high-frequency overtone components will be deleted, and the possibility of destroying the overtone structure of the input signal is reduced. Good sound quality can be obtained. In particular, it works effectively in a high-efficiency coding apparatus using a bit rate lower than the bit rate of 128 kbps in this embodiment.

The bits adjusted to the bit rate of 128 kbps in this way are divided into a plurality of bands by dividing the critical band in each critical band or in the high band by the adaptive bit allocation coding circuits 16 to 18 in FIG. Each spectrum data (or MDCT coefficient data) is re-quantized according to the number of bits assigned to. The data encoded in this way is taken out via the output terminals 22, 24 and 26 of FIG.

The auditory permissible noise spectrum calculating circuit in the bit allocation circuit 805 depending on the permissible auditory noise spectrum will be described in more detail. The MDCT circuits 13 to 1 will be described below.
The MDCT coefficient obtained in 5 is the bit allocation circuit 805.
It is given to the allowable noise spectrum calculation circuit therein.

FIG. 6 is a block circuit diagram showing a schematic configuration of a specific example in which the allowable noise spectrum calculating circuits are collectively described. In FIG. 6, the input terminal 521 has:
The spectrum data in the frequency domain is supplied from the MDCT circuits 13, 14, and 15.

The input data in the frequency domain is sent to the energy calculation circuit 522 for each band, and the energy of each critical band (critical band) is calculated, for example, as the sum of the squared amplitude values in the band. It is required by doing. Instead of the energy for each band, a peak value, an average value, etc. of the amplitude value may be used. As an output from the energy calculation circuit 522, for example, the spectrum of the total sum value of each band is generally called a Bark spectrum. FIG. 7 shows the Bark spectrum SB for each such critical band. However, in FIG. 7, in order to simplify the illustration, the number of bands of the critical band is represented by 12 bands (B1 to B12).

Here, in order to consider the influence of so-called masking of the Bark spectrum SB, a convolution process is performed such that the Bark spectrum SB is multiplied by a predetermined weighting function and added. Therefore, the output of the energy calculation circuit 522 for each band, that is, each value of the Bark spectrum SB is sent to the convolution filter circuit 523. The convolution filter circuit 523
Is, for example, a plurality of delay elements that sequentially delay input data, and a plurality of multipliers (for example, 25 multipliers corresponding to each band) that multiply outputs from these delay elements by a filter coefficient (weighting function). , A sum total adder that sums the outputs of the respective multipliers.

The above-mentioned masking means a phenomenon in which one signal is masked by another signal and becomes inaudible due to human auditory characteristics. This masking effect is caused by a time domain audio signal. There are an axial masking effect and a simultaneous time masking effect by a signal in the frequency domain. Due to these masking effects, even if there is noise in the masked portion, this noise cannot be heard. Therefore, in the actual audio signal, the noise within the masked range is regarded as an acceptable noise.

A specific example of the multiplication coefficient (filter coefficient) of each multiplier of the convolution filter circuit 523 will be described. When the coefficient of the multiplier M corresponding to an arbitrary band is 1, the multiplier M −1 gives a coefficient of 0.15, multiplier M-2 gives a coefficient of 0.0019, and multiplier M-3 gives a coefficient of 0.0000.
086, multiplier M + 1 gives a coefficient of 0.4, multiplier M + 2
By multiplying the output of each delay element by a coefficient of 0.06 and a coefficient of 0.007 by a multiplier M + 3, the convolution processing of the Bark spectrum SB is performed. However, M is 1 to
It is an arbitrary integer of 25.

Next, the output of the convolution filter circuit 523 is sent to the subtractor 524. The subtractor 524 calculates a level α corresponding to an allowable noise level described later in the convoluted area. The level α corresponding to the permissible noise level (permissible noise level) is a level at which the critical noise band becomes the permissible noise level for each band by performing inverse convolution processing, as described later. is there.

Here, the subtractor 524 is supplied with an allowance function (function expressing a masking level) for obtaining the level α. The level α is controlled by increasing or decreasing this allowance function. The permissible function is the (n-ai) function generating circuit 52 as described below.
It is supplied from No. 5.

That is, the level α corresponding to the allowable noise level can be obtained by the following equation, where i is the number given in order from the low band of the critical band.

Α = S- (n-ai) In this equation, n and a are constants and a> 0, and S is the intensity of the convolved Bark spectrum.
i) is the tolerance function. As an example, n = 38, a = -0.
5 can be used.

In this way, the level α is obtained, and this data is transmitted to the divider 526. The divider 526 is for deconvolution of the level α in the convolved area. Therefore, by performing the inverse convolution processing, the masking threshold can be obtained from the level α. That is, this masking threshold becomes the allowable noise spectrum. Although the above-mentioned inverse convolution processing requires complicated calculation, in this embodiment, the inverse convolution is performed using the simplified divider 526.

Next, the masking threshold is transmitted to the subtractor 528 via the synthesizing circuit 527. Here, the output from the energy detection circuit 522 for each band, that is, the above-described Bark spectrum SB is supplied to the subtractor 528 via the delay circuit 529. Therefore, the subtractor 528 performs a subtraction operation on the masking threshold and the Bark spectrum SB, so that the Bark spectrum SB shows the masking threshold MS as shown in FIG.
The level below the level indicated by will be masked. The delay circuit 529 is the same as the synthesis circuit 527.
Energy detection circuit 5 considering the delay amount in each circuit before
It is provided to delay the Bark spectrum SB from 22.

The output from the subtractor 528 is taken out via the allowable noise correction circuit 530 and the output terminal 531. For example, the ROM in which the distribution bit number information is stored in advance.
Etc. (not shown). The ROM or the like is distributed for each band according to the output (the level of the difference between the energy of each band and the output of the noise level setting means) obtained from the subtraction circuit 528 through the allowable noise correction circuit 530. Outputs bit number information.

In this way, the energy-dependent bits and the bits depending on the permissible noise level of the hearing are added, and the information on the number of allocated bits thereof is the adaptive bit allocation coding circuits 16-1.
By sending the data to the M.C.R.8, the spectrum data in the frequency domain from the MDCT circuits 13 to 15 are quantized by the number of bits assigned to each band.

In summary, in the adaptive bit allocation encoding circuits 16 to 18, in each band band (each critical band) of the critical band or in the high band, the critical band is further divided into a plurality of bands. The spectrum data for each band is quantized by the number of bits distributed according to the level of the difference between the energy or peak value and the output of the noise level setting means.

By the way, at the time of synthesizing by the above-mentioned synthesizing circuit 527, data showing a so-called minimum audible curve RC which is the human auditory characteristic as shown in FIG. The masking threshold MS can be combined. In this minimum audible curve, if the absolute noise level is below this minimum audible curve, the noise will not be heard. This minimum audible curve is different even if the coding is the same, for example, due to the difference in playback volume during playback,
In a realistic digital system, there is not much difference in how music is put into a 16-bit dynamic range, so if the quantization noise in the most audible frequency band around 4 kHz is inaudible, for example, other It is considered that quantization noise below the level of this minimum audible curve is inaudible in the frequency band. Therefore, for example, the dynamic range of the system is 4 kHz.
Assuming that the noise is not heard in the vicinity, and the allowable noise level is obtained by synthesizing the minimum audible curve RC and the masking threshold MS together, the allowable noise level in this case is shown in FIG. It will be possible to go up to the part shown by the diagonal line of. In this embodiment, the level of 4 kHz of the minimum audible curve is set to the minimum level equivalent to 20 bits, for example. Further, FIG. 9 also shows the signal spectrum SS at the same time.

Further, in the allowable noise correction circuit 530,
The allowable noise level in the output from the subtractor 528 is corrected based on the information of the equal loudness curve sent from the correction information output circuit 533, for example. Here, the equal loudness curve is a characteristic curve relating to human auditory characteristics, for example, a curve obtained by obtaining the sound pressure of sound at each frequency that sounds the same as a pure tone of 1 kHz, and connecting the curves. Also called sensitivity curve. Further, this equal loudness curve is the minimum audible curve RC shown in FIG.
It draws almost the same curve as. In this equal loudness curve, for example, at 4 kHz, even if the sound pressure drops by 8 to 10 dB from 1 kHz, it sounds as loud as 1 kHz, and conversely, at around 50 Hz, it must be about 15 dB higher than the sound pressure at 1 kHz. It doesn't sound the same.
Therefore, it is understood that the noise exceeding the level of the minimum audible curve (allowable noise level) should have the frequency characteristic given by the curve corresponding to the equal loudness curve. From this, it can be seen that correcting the permissible noise level in consideration of the equal loudness curve is suitable for human hearing characteristics.

The spectrum shape depending on the permissible hearing noise level described above is created by bit allocation using a certain ratio of the total usable bits of 128 Kbps. This ratio decreases as the tonality of the input signal increases.

Next, a bit amount dividing method between the two bit allocation methods will be described.

Returning to FIG. 3, the signal from the input terminal 801 to which the output of the MDCT circuit is supplied is also given to the spectrum smoothness calculating circuit 808, and the smoothness of the spectrum is calculated here. In this embodiment, a value obtained by dividing the sum of the absolute values of the differences between the adjacent values of the signal spectrum by the sum of the absolute values of the signal spectrum is calculated as the smoothness of the spectrum.

The above-mentioned spectrum smoothness calculation circuit 808
Is supplied to the bit division rate determination circuit 809, and the bit division rate between the energy-dependent bit allocation and the bit allocation according to the perceptual noise spectrum is determined. The bit division rate is the smoothness calculation circuit 80 of the spectrum.
It is considered that the larger the output value of 8, the smoother the spectrum is, and the bit allocation is performed with more emphasis on the bit allocation by the perceptual noise spectrum than the energy-dependent bit allocation. The bit division ratio determination circuit 809 sends control outputs to multipliers 811 and 812 which control the amount of bit distribution depending on the energy and the permissible noise spectrum of the auditory sense, respectively. Here, if the spectrum is smooth and the output of the bit division rate determination circuit 809 to the multiplier 811 is 0.8 so as to emphasize the energy-dependent bit allocation.
, The output of the bit division rate determination circuit 809 to the multiplier 812 is 1-0.8 = 0.2. The outputs of these two multipliers are the adders 806.
Are added together to form final bit allocation information, which is output from the output terminal 807.

The state of bit allocation at this time is shown in FIGS. 12 and 13 show the state of quantization noise corresponding to this. FIG. 10 shows the case where the spectrum of the signal is flat, and FIG. 11 shows the case where the signal spectrum shows high tonality. Also,
In FIGS. 10 and 11, QS indicates the bit amount corresponding to the signal level, and QN in the diagrams indicates the bit amount corresponding to the bit allocation depending on the permissible hearing noise level. 12 and 13, L represents the signal level, NS in the figures represents the noise reduction due to the signal level dependency, and NN in the figures represents the noise reduction due to the permissible noise level dependent bit allocation. .

First, in FIG. 12, which shows a case where the spectrum of the signal is flat, the bit allocation depending on the permissible noise level of the hearing aids in obtaining a large signal-to-noise ratio over the entire band. However, relatively low bit allocations are used in the low and high frequencies. This is because auditory sensitivity to noise in this band is low. The bit allocation depending on the signal energy level is small in amount, but in this case, it is concentrated in the high frequency region of the signal level in the middle and low frequencies so as to generate a white noise spectrum.

On the other hand, as shown in FIG. 13, when the signal spectrum shows a high tonality, the amount of bit allocation depending on the signal energy level increases, and the quantization noise is reduced in a very narrow band. Used to reduce The concentration of bit allocation depending on the permissible noise level of the hearing is less tight than this.

As shown in FIG. 13, the improvement of the characteristics in the isolated spectrum input signal is achieved by the sum of the bit allocations of the both.

FIG. 14 shows a basic high efficiency decoding apparatus of the present invention for decoding the signal coded by the high efficiency coding apparatus shown in FIG. 1 again.

In FIG. 14, the quantized MDCT coefficients of each band are the decoding device input terminals 122 and 124,
The block size information given to 126 and used is given to the input terminals 123, 125, 127. The decoding circuits 116, 117 and 118 cancel the bit allocation using the adaptive bit allocation information.

Next, the IMDCT (Inverse Change Discrete Cosine Transform) circuits 113, 114 and 115 convert the frequency domain signals into time domain signals. The time domain signals of these sub-bands are IQMF (inverse QMF) circuits 112 and 11
1, it is decoded into a full-range signal and sent to the output terminal 110.

Next, the transmission medium of the embodiment of the present invention transmits the signal coded by the high-efficiency encoder of the embodiment of the present invention as described above. Therefore, the recording medium (recording medium) of the present invention includes, for example, a disc-shaped recording medium such as an optical disc, a magneto-optical disc, and a magnetic disc in which the above-mentioned encoded signal is recorded, and a magnetic tape. Examples thereof include a tape-shaped recording medium in which the above-mentioned coded signal is recorded, a semiconductor memory in which the coded signal is stored, an IC card, and the like. When a transmission system is used as the transmission medium of the present invention, examples of the transmission medium (transmission medium) include electric wires, optical cables, and radio waves.

The present invention is not limited to this embodiment. For example, the recording / reproducing medium described above and the other recording / reproducing medium do not need to be integrated, and data is transferred between them. It is also possible to connect by a line or the like. Furthermore, for example, the present invention can be applied not only to audio PCM signals but also to signal processing devices for digital audio (speech) signals, digital video signals, and the like. Further, the above-described minimum audible curve synthesizing process may not be performed. In this case, the minimum audible curve generating circuit 532 and the synthesizing circuit 527 in FIG. 6 are unnecessary, and the output from the subtractor 526 is immediately transmitted to the subtractor 528.

There are various bit allocation methods,
In the simplest case, it is possible to use fixed bit allocation, simple bit allocation by each band energy of the signal, or bit allocation combining fixed and variable components.

A method is also conceivable in which it is determined whether or not the input signal has a harmonic relationship and the bit allocation is changed. For example, when a clear overtone relation cannot be detected by the overtone component detection circuit 813 in FIG. 3, it is possible to adopt a configuration in which output to the energy-dependent bit distribution circuit 804 and the bit fraction adjustment circuit 814 is not performed. .

Various modifications as described above can be considered in the embodiment of the present invention.

[0078]

As is clear from the above description, the following effects can be obtained in the present invention. In other words, by allocating bits in consideration of the harmonic relationship of the input signal, it becomes possible to allocate bits adaptively to the input signal, and at the same bit rate, good sound quality can be obtained in terms of hearing. It becomes possible. Further, at a low bit rate, it is possible to prevent sound quality deterioration.

[Brief description of drawings]

FIG. 1 is a block circuit diagram showing a configuration example of a high-efficiency coding apparatus according to an embodiment of the present invention.

FIG. 2 is a diagram showing a specific example of orthogonal transform block sizes in the frequency and time domains in the device of this embodiment.

FIG. 3 is a block circuit diagram showing a configuration example of a bit allocation function of the embodiment of the present invention.

FIG. 4 is a diagram showing a harmonic relationship of input signals.

FIG. 5 is a diagram showing changes in the amount of quantization noise in each block, with and without consideration of a harmonic relationship of an input signal.

FIG. 6 is a block circuit diagram showing a configuration example of a hearing masking threshold calculation function of an embodiment of the present invention.

FIG. 7 is a diagram showing masking by each critical band signal.

FIG. 8 is a diagram showing a masking threshold by each critical band signal.

FIG. 9: Information spectrum, masking threshold,
It is a figure which shows the minimum audible limit.

10 is a diagram showing signal level-dependent and auditory permissible noise level-dependent bit allocation for an information signal having a flat signal spectrum. FIG.

FIG. 11 is a diagram showing signal level-dependent and auditory permissible noise level-dependent bit allocation for an information signal having a high tonality of a signal spectrum.

FIG. 12 is a diagram showing a quantization noise level for an information signal having a flat signal spectrum.

FIG. 13 is a diagram showing a quantization noise level for an information signal having high tonality.

FIG. 14 is a block circuit diagram showing a configuration example of a high efficiency decoding device according to an embodiment of the present invention.

[Explanation of symbols]

 801 MDCT circuit output input terminal 802 Total usable bit generation circuit 803 Energy calculation circuit for each band 804 Energy-dependent bit allocation circuit 805 Hearing permissible noise level-dependent bit allocation circuit 806 Adder 807 Bit allocation amount output terminal for each band 808 Spectrum smoothness calculation circuit 809 Bit division rate determination circuit 811, 812 Multiplier 813 Overtone component detection circuit 814 Bit fraction adjustment circuit

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Office reference number FI Technical indication location H03M 7/02 9382-5K 7/30 A 9382-5K H04B 14/00 E

Claims (30)

[Claims] 1. An input signal is decomposed into a plurality of frequency band components to obtain signal components in a plurality of two-dimensional blocks with respect to time and frequency, and features of the signal components in the two-dimensional block for each of the two-dimensional blocks. , The bit allocation amount is determined based on the quantized coefficient, and the signal component in the two-dimensional block is quantized for information compression for each two-dimensional block, and the compression information for each two-dimensional block is obtained. In the high-efficiency coding method for outputting together with the information compression parameter for each two-dimensional block, when performing bit allocation adapted to the input signal, the overtone relation of the input signal is detected, and each two-dimensional block containing the overtone component is detected. A high-efficiency coding method characterized by increasing the degree of concentration of bit allocation amounts for blocks. 2. The harmonic relation of the input signal is detected on the basis of spectrum information on the frequency axis of the input signal when performing bit allocation adapted to the input signal. High efficiency coding method. 3. When performing bit allocation adapted to an input signal, a harmonic relationship of the input signal is detected based on a quantized coefficient representing a signal component in a two-dimensional block. The high-efficiency coding method according to Item 1. 4. When performing bit allocation adapted to an input signal, all or part of the frequency band is examined to detect the overtone relationship of the input signal.
High efficiency coding method described. 5. When performing bit allocation adapted to an input signal, only low-order harmonic components are detected, the frequency bands of other harmonic components are calculated based on these harmonic components, and the entire frequency band is calculated. The high-efficiency coding method according to claim 1, wherein the overtone relation possessed by is detected. 6. When performing bit allocation adapted to an input signal, only the overtone components of the fundamental tone and the first overtone are detected, and the frequency bands of other overtone components are calculated based on these overtone components, The high-efficiency coding method according to claim 4, wherein the harmonic relations of all frequency bands are detected. 7. When performing bit allocation adapted to an input signal, based on spectrum information of all or a part of the frequency band of the input signal and / or quantization information representing a signal component in a two-dimensional block, Detect only low-order harmonic components and / or fundamental and first harmonic overtone components, calculate the frequency band of other harmonic components based on these harmonic components, and detect the harmonic relationship of all frequency bands. The high-efficiency coding method according to claim 1. 8. The method according to claim 1, wherein, when performing bit allocation adapted to an input signal, the degree of bit allocation is increased with respect to a two-dimensional block containing all or a part of overtone components. High efficiency coding method. 9. When performing bit allocation adapted to an input signal, the degree of concentration of bit allocation is increased for all or some of the two-dimensional blocks containing even harmonic components. The high-efficiency coding method described in 1. 10. When performing bit allocation adapted to an input signal, all or a part of 2 including an odd-order harmonic component is included.
The high-efficiency coding method according to claim 1, wherein the degree of concentration of bit allocation is increased with respect to the dimensional block. 11. When determining a bit allocation amount adapted to an input signal, when the bit allocation amount is small with respect to a predetermined bit rate, priority is given to all or some of the two-dimensional blocks containing overtone components. 2. The high-efficiency coding method according to claim 1, further comprising adding bits to the bits to adjust the fraction of the bits. 12. When determining the bit allocation amount adapted to an input signal, when the bit allocation amount is large with respect to a predetermined bit rate, priority is given to all or some of the two-dimensional blocks that do not include overtone components. 2. The high-efficiency coding method according to claim 1, wherein the bits are deleted to adjust the fraction of the bits. 13. When determining a bit allocation amount adapted to an input signal, when the bit allocation amount is small with respect to a predetermined bit rate, priority is given to all or some two-dimensional blocks containing even harmonic components. 2. The high-efficiency encoding method according to claim 1, wherein the fractional adjustment of bits is performed by adding bits. 14. When determining a bit allocation amount adapted to an input signal, when the bit allocation amount is large with respect to a predetermined bit rate, for all or some of the two-dimensional blocks that do not include an odd harmonic component. 2. The high efficiency encoding method according to claim 1, wherein the bits are preferentially deleted to adjust the fraction of the bits. 15. A block signal component forming means for decomposing an input signal into a plurality of frequency band components to form signal components in a plurality of two-dimensional blocks regarding time and frequency, and the two-dimensional block for each two-dimensional block. An information compression means for obtaining a quantized coefficient representing a characteristic of a signal component in a block, determining a bit allocation amount based on the quantized coefficient, and quantizing the signal component in the two-dimensional block for each two-dimensional block to compress information. In the high-efficiency coding apparatus for outputting the compression information for each of the two-dimensional blocks together with the information compression parameter for each of the two-dimensional blocks, the information compression means detects the overtone relationship of the input signal. Detecting means and bit allocation amount determining means for determining the bit allocation amount, and when performing bit allocation adapted to the input signal, harmonic components are included. High-efficiency encoding apparatus characterized by increasing the concentration of the bit allocation with respect to the original block. 16. The overtone relationship detecting means detects an overtone relationship of the input signal based on spectrum information on the frequency axis of the input signal.
The high-efficiency coding device described. 17. The high-efficiency code according to claim 15, wherein the overtone relationship detecting means detects an overtone relationship of the input signal based on a quantization coefficient representing a signal component in a two-dimensional block. Device. 18. The high-efficiency coding apparatus according to claim 1, wherein the overtone relation detecting means detects the overtone relation of the input signal by examining all or a part of frequency bands. 19. The overtone relationship detecting means detects only low-order overtone components, calculates frequency bands of other overtone components based on these overtone components, and detects overtone relationships of all frequency bands. The high-efficiency coding device according to claim 15, wherein 20. The overtone relation detecting means detects only the overtone components of the fundamental tone and the first overtone, calculates the frequency bands of the other overtone components based on these overtone components, and the overtones of the entire frequency band. The high efficiency encoding device according to claim 18, wherein the relationship is detected. 21. The harmonic overtone detection unit detects low order harmonic overtone components based on spectrum information of frequency bands of all or a part of an input signal and / or quantized information representing a signal component in a two-dimensional block. And / or detecting only the overtone components of the fundamental tone and the first overtone, calculating the frequency bands of other overtone components based on these overtone components, and detecting the overtone relationship of all frequency bands. 15. The high-efficiency coding device according to item 15. 22. The bit allocation amount determining means enhances the degree of concentration of bit allocation with respect to a two-dimensional block including all or some of the overtone components.
The high-efficiency coding device described. 23. The high efficiency according to claim 15, wherein the bit allocation amount determining means increases the degree of concentration of bit allocation for all or some of the two-dimensional blocks including even harmonic components. Encoding device. 24. The high efficiency according to claim 15, wherein the bit allocation amount determining means increases the degree of bit allocation concentration for all or some of the two-dimensional blocks containing odd harmonic components. Encoding device. 25. When the bit allocation amount adapted to an input signal is determined, the bit allocation amount determining means includes all or a part of the 2 or more parts which include harmonic components when the bit allocation amount is smaller than a predetermined bit rate. 16. The high-efficiency encoding apparatus according to claim 15, wherein bits are preferentially added to the dimensional block to adjust the fraction of the bits. 26. When determining the bit allocation amount adapted to an input signal, the bit allocation amount determining means does not include all or a part of overtone components when the bit allocation amount is large with respect to a predetermined bit rate. 16. The high-efficiency coding device according to claim 15, wherein bits are preferentially deleted from the two-dimensional block to adjust the fraction of the bits. 27. When determining the bit allocation amount adapted to an input signal, the bit allocation amount determining means includes all or a part of even harmonic components if the bit allocation amount is smaller than a predetermined bit rate. 16. The high-efficiency coding apparatus according to claim 15, wherein bits are preferentially added to the two-dimensional block to adjust the fraction of the bits. 28. The bit allocation amount determining means, when determining a bit allocation amount adapted to an input signal, does not include an odd harmonic component when the bit allocation amount is large with respect to a predetermined bit rate, or all or part of the odd harmonic components are not included. 16. The high efficiency encoding device according to claim 15, wherein bits are preferentially deleted in the two-dimensional block to adjust the fraction of the bits. 29. A transmission medium, which transmits compression information and information compression parameters compression-encoded by the high-efficiency encoding method according to any one of claims 1 to 14. . 30. A transmission medium, which transmits compression information and information compression parameters compression-encoded by the high-efficiency encoding device according to any one of claims 15 to 28. .