JP2002049399A - Digital signal processing method, learning method, and their apparatus, and program storage media therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a digital signal processing method capable of further improving the waveform reproducibility of a digital signal, a learning method, and their apparatus, and a program storage media therefor. SOLUTION: In this digital signal processing method, power spectrum data are calculated from a digital audio signal D10, the calculated power spectrum is normalized at the maximum value width and normalization data are calculated, the class of the digital audio signal D10 is categorized on the basis of the calculated normalization data, and the digital audio signal D10 is converted by a prediction system corresponding to the categorized class. Thus the method makes it possible to perform conversion further adaptive to the characteristics of the digital audio signal D10.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a digital signal processing method, a learning method, a device thereof, and a program storage medium, and more particularly to a rate converter or a PCM (Pulse Code).
Modulation) The present invention is suitably applied to a digital signal processing method and a learning method for performing data interpolation processing on a digital signal in a decoding device or the like, a device thereof, and a program storage medium.

[0002]

2. Description of the Related Art Conventionally, before a digital audio signal is input to a digital / analog converter, an oversampling process for converting a sampling frequency to several times the original value is performed. As a result, the digital audio signal output from the digital / analog converter maintains the phase characteristic of the analog anti-aliasing filter constant at high audio frequencies, and eliminates the influence of digital image noise caused by sampling. It has been made to be.

In such oversampling processing, a digital filter of a linear primary (linear) interpolation system is usually used. Such a digital filter generates linear interpolation data by calculating the average value of a plurality of existing data when the sampling rate changes or data is lost.

[0004]

However, although the digital audio signal after the oversampling process has a data amount several times more dense in the time axis direction by linear linear interpolation, the digital audio signal after the oversampling process has been used. The frequency band of the audio signal is not much different from that before conversion, and the sound quality itself has not been improved. Furthermore, since the interpolated data is not necessarily generated based on the waveform of the analog audio signal before A / D conversion, the waveform reproducibility is hardly improved.

Further, when dubbing digital audio signals having different sampling frequencies, the frequency is converted using a sampling rate converter. Even in such a case, only linear data interpolation is performed by a linear primary digital filter. Can not
It was difficult to improve sound quality and waveform reproducibility. The same applies to a case where a data sample of a digital audio signal is missing.

The present invention has been made in view of the above points, and proposes a digital signal processing method, a learning method, a device thereof, and a program storage medium capable of further improving the waveform reproducibility of a digital audio signal. Things.

[0007]

According to the present invention, power spectrum data is calculated from a digital audio signal, and the calculated power spectrum data is normalized by a maximum value width to calculate normalized data. By classifying the class based on the calculated normalized data and converting the digital audio signal by a prediction method corresponding to the classified class,
It is possible to perform conversion that is more adapted to the characteristics of the digital audio signal.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below in detail with reference to the drawings.

In FIG. 1, an audio signal processing device 10
When increasing the sampling rate of a digital audio signal (hereinafter referred to as audio data) or interpolating audio data, audio data that is close to a true value is generated by class classification application processing.

[0010] Incidentally, the audio data in the present embodiment is tone data representing human voices and the sounds of musical instruments, and data representing various other sounds.

That is, in the audio signal processing apparatus 10, the spectrum processing section 11 converts the input audio data D10 supplied from the input terminal T _IN into a region for every predetermined time (in this embodiment, for example, every six samples).
After constructing a class tap which is the time axis waveform data cut out in the above, log data is calculated for the constructed class tap according to control data D18 supplied from the input unit 18 by a log data calculation method described later.

The spectrum processing unit 11 calculates log data D11, which is a result of the log data calculation method and is to be classified, with respect to the class taps constructed at this time of the input audio data D10. To supply.

The classifying unit 13 includes the spectrum processing unit 1
An ADRC (Adaptive Dynamic Range Coding) circuit unit for compressing the log data D11 and generating a compressed data pattern for the log data D11 supplied from 1;
A class code generation circuit for generating a class code to which the logarithmic data D11 belongs.

The ADRC circuit section performs an operation to compress the logarithmic data D11 from, for example, 8 bits to 2 bits to form pattern compression data. The ADRC circuit section performs adaptive quantization. Here, since a local pattern of a signal level can be efficiently represented by a short word length, the ADRC circuit section is used for generating a code for classifying a signal pattern. Used for

[0015] More specifically, when attempting to classification six 8-bit data (log data), must be classified into enormous number of classes 2 ^48, the greater the burden on the circuit. Therefore, the class classification unit 14 of this embodiment classifies the data based on the compressed pattern data generated by the ADRC circuit unit provided therein.
For example, if 1-bit quantization is performed on six logarithmic data, the six logarithmic data can be represented by six bits and can be classified into 2 ⁶ = 64 classes.

Here, the ADRC circuit section calculates the dynamic range in the cut-out area as DR, the bit allocation as m, the data level of each logarithmic data as L, and the quantization code as Q, as follows:

[0017]

(Equation 1)

In accordance with the above, quantization between the maximum value MAX and the minimum value MIN in the area is equally divided by the designated bit length. In equation (1), {} means truncation processing below the decimal point. Assuming that each of the six logarithmic data calculated in the spectrum processing unit 11 is composed of, for example, 8 bits (m = 8), these are each compressed to 2 bits in the ADRC circuit unit.

Assuming that the log data thus compressed is q _n (n = 1 to 6), the class classification unit 14
Is based on the compressed log data q _n , the following equation:

[0020]

(Equation 2)

By executing the operation shown in ( ₁ ), a class code class indicating the class to which the block (q _{1 to} q ₆ ) belongs is calculated, and the calculated class code class
Is stored in the prediction coefficient memory 15
To supply. This class code class indicates a read address when a prediction coefficient is read from the prediction coefficient memory 15. In the expression (2), n represents the number of compressed logarithmic data q _n , n = 6 in this embodiment, P represents bit allocation, and P = 2 in this embodiment. It is.

As described above, the classifying unit 14 calculates the logarithmic data D1 calculated from the input audio data D10.
1 is generated and supplied to the prediction coefficient memory 15.

The prediction coefficient memory 15 stores a set of prediction coefficients corresponding to each class code at an address corresponding to the class code.
Based on the class code data D14 supplied from
The set of prediction coefficients W _{1 to} W _n stored at the address corresponding to the class code is read and supplied to the prediction calculation unit 16.

The prediction calculation unit 16 includes a prediction calculation unit extraction unit 13
, Audio waveform data (prediction taps) D13 (X _{1 to} X _n ) cut out from the input audio data D10 in the time domain to be calculated, and a prediction coefficient W
_{For 1 to} W _n ,

[0025]

(Equation 3)

The prediction result y 'is obtained by performing the product-sum operation shown in FIG. The prediction value y 'is output from the prediction calculation unit 16 as audio data D16 with improved sound quality.

Although the functional blocks described above with reference to FIG. 1 are shown as the configuration of the audio signal processing device 10,
As a specific configuration of the functional blocks, in this embodiment, an apparatus having a computer configuration shown in FIG. 2 is used. That is, in FIG. 2, the audio signal processing device 10 includes a CPU 21 and a ROM via a bus BUS.
(Read Only Memory) 22, a RAM (Random Access Memory) 15 constituting the prediction coefficient memory 15, and each circuit unit are connected to each other.
By executing various programs stored in the storage unit 2, the function blocks (the spectrum processing unit 11, the prediction calculation unit extraction unit 13, and the class classification unit 1) described above with reference to FIG.
4 and a prediction calculation unit 16).

The audio signal processor 10 has a communication interface 2 for communicating with a network.
4. A hard disk drive 25 having a removable drive 28 for reading information from an external storage medium such as a floppy disk or a magneto-optical disk, and performing the class classification application processing described above with reference to FIG. , And the classification adaptive processing can be performed according to the read program.

The user inputs various commands through input means 18 such as a keyboard and a mouse,
The CPU 21 is caused to execute the class classification processing described above with reference to FIG. In this case, the audio signal processing device 1
0 is audio data (input audio data) D10 whose sound quality is to be improved via the data input / output unit 27.
After the input audio data D10 is subjected to the class classification application processing, the audio data D16 with improved sound quality can be output to the outside via the data input / output unit 27.

FIG. 3 shows the audio signal processing device 1
The processing procedure of the class classification adaptation processing at 0 is shown. When the audio signal processing apparatus 10 enters the processing procedure from step SP101, the logarithmic data D11 of the input audio data D10 is calculated by the spectrum processing unit 11 in the following step SP102.

The calculated log data D11 represents the characteristics of the input audio data D10, and the audio signal processing apparatus 10 proceeds to step SP103 and classifies the class based on the log data D11 by the class classification unit 14. . Then, the audio signal processing device 10 reads a prediction coefficient from the prediction coefficient memory 15 using the class code obtained as a result of the classification. The prediction coefficients are stored in advance for each class by learning, and the audio signal processing device 10 reads out the prediction coefficients corresponding to the class codes, thereby obtaining the logarithmic data D at this time.
A prediction coefficient matching the eleven characteristics can be used.

The prediction coefficient read from the prediction coefficient memory 15 is used in the prediction operation of the prediction operation unit 16 in step SP104. As a result, the input audio data D10 is converted into desired audio data D16 by a prediction operation adapted to the characteristics of the log data D11. Thus, the input audio data D10 is converted into audio data D16 having improved sound quality,
The audio signal processing device 10 proceeds to step SP105 and ends the processing procedure.

Next, the input audio data D10 in the spectrum processing section 11 of the audio signal processing device 10
The method of calculating the log data D11 will be described.

That is, FIG. 4 shows a logarithmic data calculation processing procedure of the logarithmic data calculation method in the spectrum processing section 11. When the processing procedure starts from step SP1, the spectrum processing section 11 converts the input audio data D10 in step SP2. A class tap, which is time-axis waveform data cut out for each predetermined time, is constructed, and the process proceeds to step SP3.

In step SP3, the spectrum processing unit 11 sets the window function to "W (k)" for the class tap.
Then the following equation:

[0036]

(Equation 4)

The multiplication data is calculated according to the Hamming window shown in FIG. Incidentally, in the multiplication processing of the window function, the first value and the last value of each of the class taps constructed at this time are made equal in order to improve the accuracy of the frequency analysis performed in the subsequent step SP4. ing. Further, in equation (1), “N” represents the number of samples of the Hamming window,
“K” indicates the order of the sample data.

In step SP4, the spectrum processing unit 11 performs a fast Fourier transform (FF) on the multiplied data.
T: Fast Fourier Transform)
Then, the power spectrum data as shown in FIG.

At step SP5, the spectrum processing section 11 extracts only significant power spectrum data from the power spectrum data.

In this extraction processing, among the power spectrum data calculated from the N multiplied data, the power spectrum data group AR2 (FIG. 5) on the right side from N / 2 is
The components are almost the same as the power spectrum data group AR1 (FIG. 5) on the left side from the zero value to N / 2 (that is, symmetric). This indicates that the components of the power spectrum data at two frequency points equidistant from both ends within the frequency band of the N multiplied data are conjugate to each other. Therefore, the spectrum processing unit 11
Only the left power spectrum data group AR1 (FIG. 5) from the zero value to N / 2 is to be extracted.

Then, the spectrum processing unit 11 selects m power spectrums from the power spectrum data group AR1 to be extracted which are not selected and set in advance by the user via the input means 18 (FIGS. 1 and 2). Extract without data.

More specifically, when the user makes a selection setting through the input means 18 so that, for example, a human voice has a higher sound quality, the control data D18 corresponding to the selection operation is set.
Is output from the input means 18 to the spectrum processing unit 11 (FIGS. 1 and 2), whereby the spectrum processing unit 11
Is the power spectrum data group AR1 extracted at this time.
From (Fig. 5), 500 Hz that is significant in human voice
, And extracts only the power spectrum data in the vicinity of 4 kHz (that is, the power spectrum data other than the vicinity of 500 Hz to 4 kHz is m power spectrum data to be removed).

When the user makes a selection through the input means 18 (FIGS. 1 and 2) so that, for example, the music has a higher sound quality, the control data D18 corresponding to the selection operation is input to the input means 18. 18 to the spectrum processing unit 11, whereby the spectrum processing unit 11 outputs the power spectrum data group AR 1 (FIG. 5)
Only power spectrum data around 20 Hz to 20 kHz which is significant in music is extracted (ie, 20 Hz).
The power spectrum data other than the frequency from about 20 Hz to 20 kHz is m power spectrum data to be removed.)

Thus, the input means 18 (FIGS. 1 and 2)
The control data D18 output from the CPU determines the frequency components to be extracted as significant power spectrum data, and reflects the intention of the user to manually perform the selection operation via the input means 18 (FIGS. 1 and 2). are doing.

Therefore, the spectrum processing unit 11 for extracting the power spectrum data according to the control data D18
The frequency component of the specific audio component that the user desires to output with high sound quality is extracted as significant power spectrum data.

Incidentally, since the spectrum processing unit 11 represents the pitch of the original waveform in the power spectrum data group AR1 to be extracted, the spectrum processing unit 11 excludes the DC spectrum power spectrum data having no significant feature. It has been made to extract.

As described above, in step SP5, the spectrum processing unit 11 removes the m pieces of power spectrum data from the power spectrum data group AR1 (FIG. 5) according to the control data D18, and also removes the DC component power spectrum data. Only the necessary minimum power spectrum data to be removed, that is, only significant power spectrum data is extracted, and the process proceeds to subsequent step SP6.

In step SP6, the spectrum processing section 11 performs processing on the extracted power spectrum data.
The following formula,

[0049]

(Equation 5)

The maximum value (ps_max) of the power spectrum data (ps [k]) extracted at this time is calculated according to the following equation:

[0051]

(Equation 6)

According to the above, the power spectrum data (ps [k]) extracted at this time is normalized (divided) by the maximum value (ps_max), and the reference value (psn) obtained at this time is obtained.
[k]),

[0053]

(Equation 7)

In accordance with the above, logarithmic (decibel) conversion is performed.

In the expression (7), log is a common logarithm. In the logarithmic conversion, a small waveform can be represented as a decibel value (sound pressure level) by an arbitrary reference value. Therefore, for example, when the spectrum processing unit 11 does not logarithmically convert audio data having a significant small waveform near a large waveform, the audio data is generally quantized with a large bit number such as 16 bits. As a result, a significant small waveform portion is masked into a large waveform.

For this reason, the spectrum processing section 11 cannot find a characteristic portion (significant small waveform portion). Therefore, the spectrum processing unit 11 also performs a logarithmic transformation to find a characteristic portion (a significant small waveform portion).

Further, since a human sensation to a stimulus such as a sound pitch is almost proportional to the logarithm of the intensity, a quantity expressed by logarithmic conversion (ie, a decibel value) indicates a degree of sensation. Therefore, the spectrum processing unit 11 performs logarithmic conversion, so that the person who is to hear the sound can comfortably hear the sound.

As described above, in step SP6, the spectrum processing unit 11 performs the normalization at the maximum amplitude and the logarithmic conversion of the amplitude to find a characteristic portion (significant small waveform portion). The log data D11 is calculated so that the person who is to listen to can listen comfortably, and the process proceeds to step SP7 where the log data calculation procedure ends.

In this way, the spectrum processing unit 11 can calculate log data D11 in which the characteristics of the signal waveform represented by the input audio data D10 are found further by the log data calculation procedure of the log data calculation method. .

Next, a learning circuit for obtaining a set of prediction coefficients for each class stored in the prediction coefficient memory 15 described above with reference to FIG. 1 by learning in advance will be described.

In FIG. 6, the learning circuit 30 converts the high-quality teacher audio data D30 into the student signal generation filter 3.
Receive at 7. The student signal generation filter 37 thins out the teacher audio data D30 by a predetermined number of samples at a predetermined time interval at the thinning rate set by the thinning rate setting signal D39.

In this case, the generated prediction coefficient differs depending on the thinning rate in the student signal generation filter 37.
The audio data reproduced by the above-described audio signal processing device 10 differs accordingly. For example, when the audio signal processing device 10 attempts to improve the sound quality of audio data by increasing the sampling frequency, the student signal generation filter 37 performs a thinning process to reduce the sampling frequency. On the other hand, when the audio signal processing device 10 described above aims to improve the sound quality by compensating for the missing data sample of the input audio data D10, the student signal generation filter 37 responds accordingly. A thinning-out process is performed to remove the data.

Thus, the student signal generation filter 37 generates the student audio data D37 from the teacher audio data 30 by a predetermined thinning process, and supplies this to the spectrum processing unit 31 and the prediction calculation unit extraction unit 33.

The spectrum processing section 31 divides the student audio data D37 supplied from the student signal generation filter 37 into regions at predetermined time intervals (in this embodiment, for example, every six samples), and then performs the division. For each of the time domain waveforms, log data D31, which is a result of the log data calculation method described above with reference to FIG.
To supply.

The class classifying unit 34 includes the spectrum processing unit 3
For the log data D31 supplied from No. 1, an ADRC circuit section that compresses the log data D31 to generate a compressed data pattern, and a class code generation circuit section that generates a class code to which the log data D31 belongs.

The ADRC circuit unit performs an operation for compressing the logarithmic data D31 from, for example, 8 bits to 2 bits to form pattern compression data. The ADRC circuit section performs adaptive quantization. Here, since a local pattern of a signal level can be efficiently represented by a short word length, the ADRC circuit section is used for generating a code for classifying a signal pattern. Used for

[0067] More specifically, when attempting to classification six 8-bit data (log data), must be classified into enormous number of classes 2 ^48, the greater the burden on the circuit. Therefore, the class classification unit 34 of this embodiment classifies the data based on the compressed pattern data generated by the ADRC circuit unit provided therein.
For example, if 1-bit quantization is performed on six logarithmic data, the six logarithmic data can be represented by six bits and can be classified into 2 ⁶ = 64 classes.

Here, the ADRC circuit section calculates the dynamic range in the cut-out area as DR, the bit allocation as m, the data level of each logarithmic data as L, the quantization code as Q, and the above equation (1). By the same operation, quantization between the maximum value MAX and the minimum value MIN in the area is equally divided by the designated bit length. Thus, the six logarithmic data calculated by the spectrum processing unit 31 are:
Assuming that each is composed of, for example, 8 bits (m = 8), these are each 2 bits in the ADRC circuit section.
Compressed to bits.

Assuming that the log data thus compressed is q _n (n = 1 to 6), the class classification unit 34
Class class code generating circuit section provided, on the basis of the compressed log data q _n, by performing a calculation similar to the above equation (2), the block (q ₁ to q ₆₎ belongs to the Is calculated, and class code data D34 representing the calculated class code class is supplied to the prediction coefficient calculation unit 36. In the expression (2), n represents the number of compressed logarithmic data q _n , n = 6 in this embodiment, P represents bit allocation, and P = 2 in this embodiment. It is.

As described above, the class classification unit 34 generates the class code data D34 of the logarithmic data D31 supplied from the spectrum processing unit 31, and supplies this to the prediction coefficient calculation unit 36. In addition, the prediction coefficient calculation unit 36 cuts out and supplies the audio waveform data D33 (x ₁ , x ₂ ,..., X _n ) in the time axis region corresponding to the class code data D 34 in the prediction calculation unit extraction unit 33. Is done.

The prediction coefficient calculation unit 36 includes a class classification unit 34
A normal equation is established by using the class code class supplied from, the audio waveform data D33 cut out for each class code class, and the high-quality teacher audio data D30 supplied from the input terminal T _IN .

That is, the levels of n samples of the student audio data D37 are x ₁ , x _{2 ,.}
As a result of the quantized data subjected to ADRC of p bits each q _1, ......, and q _n. At this time, the class code class of this area is defined as in the above equation (2). Then, each level of the student audio data D37 as described above, x _1, x _2, ......, and x _n, the level of teacher audio data D30 of the high-quality sound y
, The prediction coefficients w ₁ , w ₂ ,
.., And a linear estimation equation of n taps by w _n is set. This is given by the following equation:

[0073]

(Equation 8)

Assume that Before learning, W _n is an undetermined coefficient.

The learning circuit 30 learns a plurality of audio data for each class code. When the number of data samples is M, the following equation is obtained according to the above equation (8).

[0076]

(Equation 9)

Is set. However, k = 1, 2,..., M.

When M> n, the prediction coefficients w ₁ ,..., W _n are not uniquely determined.

[0079]

(Equation 10)

(Where k = 1, 2,...)
..., M),

[0081]

[Equation 11]

A prediction coefficient for minimizing is calculated. This is a so-called least squares solution.

Here, the partial differential coefficient of w _n is obtained by the equation (11). In this case,

[0084]

(Equation 12)

Each W _n (n = 1 to 1) is set so that
6) may be obtained.

Then, the following equation:

[0087]

(Equation 13)

[0088]

[Equation 14]

When X _ij and Y _i are defined as follows, (1)
Expression 2) is obtained by using a matrix as follows:

[0090]

(Equation 15)

Is represented as

This equation is generally called a normal equation. Here, n = 6.

After the input of all the learning data (teacher audio data D30, class code class, audio waveform data D33) is completed, the prediction coefficient calculation unit 36 sets each class code class as shown in the above equation (15). A normal equation is established, and the normal equation is solved for each W _n using a general matrix solution method such as a sweeping method, and a prediction coefficient is calculated for each class code. Prediction coefficient calculation unit 36
Writes the calculated prediction coefficients (D36) into the prediction coefficient memory 15.

As a result of such learning, the prediction coefficient memory 15 stores a prediction for estimating high-quality audio data y for each pattern defined by the quantized data q ₁ ,..., Q _6. A coefficient is stored for each class code. This prediction coefficient memory 15 is used in the audio signal processing device 10 described above with reference to FIG. With this processing, the learning of the prediction coefficient for creating the high-quality audio data from the normal audio data in accordance with the linear estimation formula ends.

As described above, the learning circuit 30 performs the thinning process of the high-quality teacher audio data by the student signal generation filter 37 in consideration of the degree of performing the interpolation process in the audio signal processing device 10, thereby obtaining the audio signal. A prediction coefficient for the interpolation processing in the processing device 10 can be generated.

In the above configuration, the audio signal processing apparatus 10 calculates a power spectrum on the frequency axis by performing a fast Fourier transform on the input audio data D10. Frequency analysis (fast Fourier transform)
Can detect a subtle difference that cannot be known from the time axis waveform data, so that the audio signal processing device 10 can find a subtle feature that cannot find a feature in the time axis region.

In a state where a delicate feature can be found (that is, a state in which a power spectrum is calculated), the audio signal processing device 10 responds to selection range setting means (selection setting manually performed by the user from the input means 18). Only the significant power spectrum data is extracted (that is, N / 2−m).

Thus, the audio signal processing device 10
Can further reduce the processing load and improve the processing speed.

Further, the audio signal processing device 10
Log data is generated by performing normalization with the maximum amplitude and logarithmic conversion of the amplitude for the minimum required power spectrum data that has been regarded as significant. In this logarithmic conversion, a characteristic portion (significant small waveform portion) is also found, and as a result, logarithmic data is generated so that a person who is to hear the sound can comfortably hear it.

As described above, the audio signal processing device 10
Is to extract only the power spectrum data that is significant from the power spectrum data that was able to find delicate features by performing frequency analysis. The class is specified based on logarithmic data obtained by performing logarithmic conversion of amplitude and amplitude.

The audio signal processing apparatus 10 performs a prediction operation on the input audio data D10 using a prediction coefficient based on the class specified based on the extracted significant power spectrum data, thereby further increasing the input audio data D10. Audio data D1 of sound quality
6 can be converted.

Also, at the time of learning for generating prediction coefficients for each class, by obtaining prediction coefficients corresponding to a large number of teacher audio data having different phases, the input audio data D10 of the audio signal processing apparatus 10 can be obtained. Even if a phase change occurs during the classification adaptive processing, a process corresponding to the phase change can be performed.

According to the above configuration, by performing the frequency analysis, only the significant power spectrum data is extracted from the power spectrum data in which delicate features can be found, and the extracted power spectrum data is further extracted. Input data D1 using a prediction coefficient based on the result of classifying log data obtained by performing normalization with the maximum amplitude and logarithmic conversion of the amplitude.
0, the input audio data D10 is converted to the higher-quality audio data D1.
6 can be converted.

In the above-described embodiment, multiplication is performed using a Hamming window as a window function. However, the present invention is not limited to this. For example, instead of the Hamming window, a Hanning window, a Blackman window, or the like may be used. , Or by using various window functions (such as a Hamming window, a Hanning window, and a Blackman window) in advance in the spectrum processing unit. The spectrum processing unit may perform the multiplication using a desired window function according to the frequency characteristics.

Incidentally, when the spectrum processing unit performs the multiplication using the Hanning window, the spectrum processing unit applies the following equation to the class tap supplied from the cutout unit.

[0106]

(Equation 16)

Is multiplied by a Hanning window, to calculate multiplied data.

When the spectrum processing unit performs the multiplication using the Blackman window, the spectrum processing unit applies the following equation to the class tap supplied from the cutout unit.

[0109]

[Equation 17]

Is multiplied by a Blackman window, thereby calculating multiplied data.

In the above-described embodiment, the case where the fast Fourier transform is used has been described. However, the present invention is not limited to this. For example, the discrete Fourier transform (DFT: Discrete
te Fourier Transformer) and discrete cosine transform (DC
T: Discrete Cosine Transform), a maximum entropy method, a method based on linear prediction analysis, and other various frequency analysis means can be applied.

Further, in the above-described embodiment, the case has been described where the spectrum processing section 11 extracts only the left power spectrum data group AR1 (FIG. 5) from zero value to N / 2. Is not limited thereto, and only the power spectrum data group AR2 on the right side (FIG. 5) may be extracted.

In this case, the processing load on the audio signal processing device 10 can be further reduced, and the processing speed can be further improved.

Further, in the above-described embodiment, A is used as a pattern generating means for generating a compressed data pattern.
Although the case of performing DRC has been described, the present invention is not limited to this, and for example, lossless coding (DPCM: Differential Pull
se Code Modulation and vector quantization (VQ: Vector)
Compressing means such as Quantize) may be used.
In short, any compression means that can represent a signal waveform pattern with a small number of classes may be used.

Further, in the above-described embodiment, human voice and voice are selected (that is, 500 Hz to 4 kHz or 20 Hz to 20 kHz as frequency components to be extracted) as selection range setting means that can be manually selected and operated by the user.
z), the present invention is not limited to this. For example, as shown in FIG. 7, any one of high-frequency (UPP), middle-frequency (MID), and low-frequency (LOW) frequency components is selected. Alternatively, various other selection range setting means, such as sparsely selecting frequency components as shown in FIG. 8, and non-uniform frequency components as shown in FIG. 9, can be applied.

In this case, the audio signal processing device includes:
A program corresponding to the newly provided selection range setting means is created and stored in a predetermined storage means such as a hard disk drive or a ROM. Thus, even when the user manually selects the newly provided selection range setting means via the input means 18, the control data corresponding to the selected selection range setting means at this time is transmitted from the input means to the spectrum processing unit. The spectrum processing unit extracts power spectrum data from desired frequency components by using a program corresponding to the newly provided selection range setting means.

In this way, various other selection range setting means can be applied, and significant power spectrum data according to the user's intention can be extracted.

Further, in the above-described embodiment, a case has been described where the audio signal processing device 10 (FIG. 2) executes a class code generation processing procedure by a program. PCM (Pulse Code Modulation) used for various digital signal processing devices (for example, rate converter, oversampling processing device, BS (Broadcasting Satellite) broadcasting, etc.) to realize these functions and perform PCM error correction Correction device, etc.) or a program storage medium (floppy (registered trademark) disk,
These programs may be loaded from an optical disk or the like) into various digital signal processing devices to realize the respective functional units.

[0119]

As described above, according to the present invention, power spectrum data is calculated from a digital audio signal, and the calculated power spectrum data is normalized by a maximum value width to calculate normalized data. By classifying the class based on the digitized data and converting the digital audio signal by a prediction method corresponding to the classified class, it is possible to perform conversion that is more adapted to the characteristics of the digital audio signal. In addition, the digital audio signal can be converted to a high-quality digital audio signal with further improved waveform reproducibility.

[Brief description of the drawings]

FIG. 1 is a functional block diagram showing an audio signal processing device according to the present invention.

FIG. 2 is a block diagram showing an audio signal processing device according to the present invention.

FIG. 3 is a flowchart illustrating a procedure of audio data conversion processing;

FIG. 4 is a flowchart illustrating a logarithmic data calculation processing procedure.

FIG. 5 is a schematic diagram illustrating an example of calculating power spectrum data.

FIG. 6 is a block diagram illustrating a configuration of a learning circuit.

FIG. 7 is a schematic diagram illustrating an example of power spectrum data selection.

FIG. 8 is a schematic diagram illustrating an example of power spectrum data selection.

FIG. 9 is a schematic diagram illustrating an example of power spectrum data selection.

[Explanation of symbols]

10 audio signal processing device, 11 spectral processing unit, 22 ROM, 15 RAM, 24 communication interface, 25 hard disk drive, 26 input means, 27 data input / output unit , 28
...... Removable drive.

 ────────────────────────────────────────────────── ─── Continuing on the front page (72) Inventor Tsutomu Watanabe 6-7-35 Kita-Shinagawa, Shinagawa-ku, Tokyo Sony Corporation (72) Inventor Hiroto Kimura 6-35-35 Kita-Shinagawa, Shinagawa-ku, Tokyo Sony F term in reference (reference) 5D045 CC07 5J064 AA01 BB03 BC01 BC19 BC21 BD03 5K041 CC01 EE23 EE35 HH11 HH43

Claims (26)

[Claims] 1. A digital signal processing method for converting a digital audio signal, comprising: a frequency analysis step of calculating power spectrum data from the digital audio signal; and normalizing the power spectrum data by a maximum value width to calculate normalized data. A normalization step; a class classification step of classifying the class based on the normalized data; and a prediction operation of the digital audio signal by a prediction method corresponding to the classified class, thereby converting the digital audio signal. A prediction operation step of generating a new digital audio signal. 2. The method according to claim 1, wherein said frequency analysis step includes various arithmetic processing methods of a window function.
2. The digital signal processing method according to claim 1, wherein a desired arithmetic processing method is used. 3. The digital signal processing method according to claim 1, wherein in the spectrum data extracting step, when extracting the part of the power spectrum data, power spectrum data of a DC component is removed. 4. The digital signal processing method according to claim 1, wherein in the prediction calculation step, a prediction coefficient generated by learning based on a desired digital audio signal in advance is used. 5. The power spectrum data comprises substantially symmetrical components, and in the spectrum data extracting step, one of right and left components of the power spectrum data is to be extracted. 2. The digital signal processing method according to 1. 6. A digital signal processing device for converting a digital audio signal, comprising: frequency analysis means for calculating power spectrum data from the digital audio signal; and normalizing the power spectrum data by a maximum value width to calculate normalized data. Normalizing means, class classifying means for classifying the class based on the normalized data, and converting the digital audio signal by performing a prediction operation on the digital audio signal by a prediction method corresponding to the classified class. A digital signal processing apparatus for generating a new digital audio signal. 7. The frequency analysis means includes various operation processing means for a window function, and according to frequency characteristics of the digital audio signal,
7. The digital signal processing apparatus according to claim 6, wherein said arithmetic processing means is used as desired. 8. The digital signal processing apparatus according to claim 6, wherein said spectrum data extracting means excludes DC component power spectrum data when extracting said partial power spectrum data. 9. The digital signal processing apparatus according to claim 6, wherein said prediction operation means uses a prediction coefficient generated by learning based on a desired digital audio signal in advance. 10. The power spectrum data comprises substantially symmetrical components, and the spectrum data extracting means extracts one of right and left components from the power spectrum data. 2. The digital signal processing device according to claim 1. 11. A frequency analysis step of calculating power spectrum data from a digital audio signal; a normalization step of normalizing the power spectrum data by a maximum value width to calculate normalized data; A class classification step of classifying a class; a prediction calculation step of generating a new digital audio signal by converting the digital audio signal by performing a prediction calculation on the digital audio signal by a prediction method corresponding to the classified class And a program storage medium for causing a digital signal processing device to execute a program including the following. 12. The frequency analysis step includes various arithmetic processing methods of a window function, and according to a frequency characteristic of the digital audio signal,
12. The program storage medium according to claim 11, wherein a desired arithmetic processing method is used. 13. The program storage medium according to claim 11, wherein in the spectrum data extracting step, when extracting the part of the power spectrum data, power spectrum data of a DC component is removed. 14. The power spectrum data comprises substantially symmetric components, and in the spectrum data extracting step, one of right and left components of the power spectrum data is to be extracted. 12. The program storage medium according to claim 11. 15. A learning method for generating a prediction coefficient used for predicting the conversion processing of a digital signal processing device for converting a digital audio signal, wherein a student digital signal obtained by deteriorating a digital audio signal from a desired digital audio signal is provided. A student digital audio signal generation step of generating an audio signal; a frequency analysis step of calculating power spectrum data from the student digital audio signal; and a normalization step of normalizing the power spectrum data by a maximum value width to calculate normalized data. A classifying step of classifying the class based on the normalized data; and a predictor calculating a prediction coefficient corresponding to the class based on the digital audio signal and the student digital audio signal. Learning method characterized by comprising a calculation step. 16. The frequency analysis step includes various arithmetic processing methods of a window function, and according to a frequency characteristic of the digital audio signal,
16. The learning method according to claim 15, wherein a desired arithmetic processing method is used. 17. The learning method according to claim 15, wherein in the spectrum data extracting step, when extracting the part of the power spectrum data, power spectrum data of a DC component is removed. 18. The power spectrum data comprises substantially symmetric components, and in the spectrum data extracting step, one of right and left components of the power spectrum data is to be extracted. 15. The learning method according to 15. 19. A learning apparatus for generating a prediction coefficient used in a prediction operation of the above-mentioned conversion processing of a digital signal processing apparatus for converting a digital audio signal, wherein a student who has degraded the digital audio signal from a desired digital audio signal is provided. Student digital audio signal generation means for generating a digital audio signal; frequency analysis means for calculating power spectrum data from the student digital audio signal; normalization for normalizing the power spectrum data with a maximum value width to calculate normalized data Means, class classification means for classifying the class based on the normalized data, and prediction coefficient calculation means for calculating a prediction coefficient corresponding to the class based on the digital audio signal and the student digital audio signal. Learning device characterized by comprising. 20. The frequency analysis means includes various operation processing means for a window function, and according to frequency characteristics of the digital audio signal,
20. The learning device according to claim 19, wherein a desired one of the arithmetic processing units is used. 21. The learning apparatus according to claim 19, wherein said spectrum data extracting means excludes DC component power spectrum data when extracting said part of the power spectrum data. 22. The power spectrum data according to claim 19, wherein said spectrum data extracting means extracts either the left or right component of said power spectrum data. The learning device according to claim 1. 23. A student digital audio signal generating step of generating a student digital audio signal in which the digital audio signal is degraded from a desired digital audio signal, and a frequency analyzing step of calculating power spectrum data from the student digital audio signal A normalization step of normalizing the power spectrum data with a maximum value width to calculate normalized data; a class classification step of classifying the class based on the normalized data; a digital audio signal and the student digital audio A prediction coefficient calculating step of calculating a prediction coefficient corresponding to the class based on the signal and the digital signal processing device. 24. In the frequency analysis step, there are provided various arithmetic processing methods of a window function, and in accordance with a frequency characteristic of the digital audio signal,
24. The program storage medium according to claim 23, wherein a desired arithmetic processing method is used. 25. The program storage medium according to claim 23, wherein, in the spectrum data extracting step, when extracting the part of the power spectrum data, power spectrum data of a DC component is removed. 26. The power spectrum data comprises substantially symmetric components, and in the spectrum data extracting step, one of right and left components of the power spectrum data is to be extracted. 24. The program storage medium according to claim 23.