JPH1063266A - Sound processor and voice processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute a chorus processing suitable for input signals by determining a read-out address in accordance with the result of the frequency analysis of an input signal. SOLUTION: A frequency analyzing section 104 analyzes the frequency of the input signal and outputs a parameter which is the result of the analysis and is used for forming modulation to a modulation section 103. This modulation section 103 calculates the read-out offset address in accordance with the parameter inputted from frequency analyzing blocks 104, 105 and outputs the address to chorus sections 110, 111. These chorus sections 110, 111 are inputted and written with the input data to respective input terminals 101, 102 and output the data assigned by the read-out offset addresses as chorus data to mixers 114, 115. The respective mixers 114, 115 mix the output data from the chorus sections 110, 111 and the input data to the respective input terminals 101, 102 and output the data to respective output terminals 116, 117.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice processing apparatus, and more particularly to a voice processing apparatus for performing chorus characteristic processing.

[0002]

2. Description of the Related Art Audio processing is widely performed in which digitally encoded audio signals are subjected to various processes so that reproduced audio has characteristics of original sounds. Among them, the process of giving the chorus characteristic is performed by reading out the digital signal corresponding to the original sound stored in the memory with a delay.

[0003] The characteristic of the chorus is that, when reading with a delay, the sound is heard as fluctuating by raising and lowering the read address, and the sound is thickened. Also, when a chorus with large fluctuation (increase the width of raising or lowering the read address) is applied to the low frequency, the low frequency original sound component is affected, and a chorus with a slow and small fluctuation (the width of raising or lowering the read address is small) ), The thickness is further increased, giving a sense of weight.

Furthermore, it is known that a clear sound can be heard when a chorus with large fluctuation is applied to a high frequency component.

As described above, the chorus characteristics have different effects depending on the frequency components of the original sound. Therefore, the processing is performed by a low-pass filter (hereinafter referred to as LPF) or a high-pass filter (hereinafter referred to as HPF). There is a related art in which a frequency band to be performed is separated, and a process for giving a chorus characteristic is changed for each frequency band.

FIG. 4 is a block diagram showing a configuration of a conventional example.

In this conventional example, an L channel input terminal 301 and an R channel input terminal 3 for inputting an L channel and an R channel of a signal whose original sound is digitally encoded, respectively.
02, split sections 304 and 305 for inputting input signals to the respective input terminals, a modulation section 303 for generating a signal for determining the chorus characteristic, a modulation section 303 and each of the split sections 304 and 30.
5 and chorus sections 310 and 311 for generating a chorus signal, and inputs to respective input terminals 301 and 302, respective outputs of split sections 304 and 305, and respective outputs of chorus sections 310 and 311 as inputs. Mixers 314, 315, and each mixer 314, 315
It comprises an L channel output terminal 316 and an R channel output terminal 317 for outputting an output.

The input signal is supplied to a split unit 304,
The data input to 305 is divided into LPFs constituting split units 304 and 305 in order to separate the data for each frequency band.
And HPF, the filter characteristics of each filter are applied, and output to the chorus units 310 and 311 and the mixers 314 and 315. The output of the LPF is output to mixers 314 and 315 as low-frequency component data.
PF output is chorus block 3 as high frequency component data
It is output to 10, 311. Note that the characteristics (cutoff frequency) of the LPF and the HPF are set in advance.

[0009] The modulation section 303 receives the read offset address calculated by a predetermined procedure so as to generate a chorus characteristic matching the HPF characteristic set in the split section 304, and outputs the read offset address to the chorus section 31.
0,311. Each of the chorus units 310 and 311 includes a chorus delay buffer for storing an audio signal, and the contents stored in the buffer are read according to a read offset address.

Here, the relationship between the characteristics of the chorus and the read address will be described. In addition, it is assumed that one process is executed for every chorus every sampling period,
In addition, three types of chorus characteristics of TYPEA, TYPEB, and TYPEC will be described as examples.

The characteristics of the chorus include initial delay, DEP
It is determined by three kinds of parameters, TH and LFO speed, and can be expressed by the following equation (1).

NRadr = (n−1) adr + LFO (1) where, Radr = read offset address nRadr = read offset address of nfs n = current fs cycle LFO = LFO speed, Iadr = initial delay address Dsize is the Max address value of the chorus delay buffer, and the triangular wave is expressed as a function in which the Max value is ± DEPTH and the gradient is ± LFO.
adr = Iadr + LFO.

In equation (1), the modulation section 303
Shows the correlation between the read offset address output by the parameter and the above parameter.

Here, the value of nRadr is Iadr + D
If EPTH is exceeded, loopback is performed and the value of LFO changes to a minus value from the next processing.

When the value of nRadr is Iadr-DEP
When the value falls below the threshold value TH, a loopback is performed in the same manner, and the value of the LFO turns positive again.

Thus, Iadr-DEPTH changes I
It is calculated by the slope LFO / fs between adr + DEPTH, and the result is output as a read offset address.

For example, if DEPTH is 10 and LFO is 2,
When the initial delay address is 15.

Fs read offset address (output of the modulation unit 303) 1 15 2 17 3 19 ··· 625 723 ··· 16 5 17 7

In an actual system, Iadr ± D
There are few cases where the EPTH is just a turning point, and interpolation processing is performed by a known method. However, in the present invention, the method of generating the read offset processing address is not important, and a description thereof will be omitted.

The generated Radr is an output of the Mod block 303.

FIG. 5 is a graph showing the value of the read offset address Radr in the horizontal direction and fs in the vertical direction, and shows a triangular wave (output value of the modulation section 303) for generating three types of chorus characteristics. .

Here, the relationship between the characteristics of the chorus actually generated and the DEPTH and LFO will be described with reference to FIG. The description of the initial delay value is omitted because the influence of the frequency component on the characteristic is small.

The chorus TYPEA shown in FIG.
Means shallow (small) DEPTH, slow (small) LFO
The chorus characteristic generated by the read-out offset address, which is the output of the triangular wave, causes a gentle sound fluctuation.

The chorus TYPEC shown in FIG.
Is a deep (large) DEP, contrary to Chorus TYPEA
TH is a triangular wave generated by a fast (large) LFO, and the chorus characteristic generated by the readout offset address that is the output of the triangular wave is a vibrating sound.

The chorus TYPEB shown in FIG.
Is a triangular wave with shallow DEPTH and fast LFO,
The chorus characteristic generated by the read-out offset address as the output is a small fluctuation.

The read offset address, which is the output of the modulation unit 303 calculated as described above, is input to the chorus delay buffers in the chorus units 310 and 311.

FIG. 6 is a diagram showing the internal configuration of the chorus units 310 and 311. The operation of the chorus delay buffer will be described below with reference to FIG.

The chorus sections 310 and 311 include a chorus delay buffer 603 for storing audio data, a write address generation section 601 for generating an address for writing output data of the split sections 304 and 305 to the chorus delay buffer 603, and a write address generation section 601. And a read address generation unit 602 for inputting a write address, which is the output of the chorus delay buffer 603, for correcting the write address, and outputting the corrected address as a read address of the chorus delay buffer 603.

The chorus delay buffer 603 has a write address input terminal 605 and a write input data terminal 6
06, a read data output terminal 607, and a read address input terminal 608. The content input to the write input data terminal 606 is written to the address input to the write address input terminal 605, and the read address input terminal The storage content of the address input to 608 is read and output from the data output terminal 607.

The write address generator 601 sends the write address to the write address input terminal 605 of the chorus delay buffer 603 and the read address generator 6.
02, the write address is output, and then the write address is decremented.

The chorus delay buffer 603 is used as a ring buffer. Therefore, when the write address is decremented after “0”, the Max address of the chorus delay buffer 603 is set.

The read address generation block 602 includes:
A read / address / offset value 312 and a write address 604 output from the write address generation block 601 are input, and the following processing is performed to output a read address 608.

Processing of Read Address Generation Block 602 Read Address = Write Address 4 + Read Address Offset (Output of Modulation Unit 303) Since the chorus delay buffer 310 is in a ring buffer format, the read address exceeds the Max address. In such a case, the following correction processing is required. When the value is less than or equal to “0” The read address after correction = read address before correction−Max address The calculated read address is input to the read address input terminal 608 of the chorus delay buffer 603.

The chorus delay buffer 603 stores the audio data output from the split units 304 and 305 at the address specified by the write address. The chorus delay buffer 603 has a read address 608
The contents of the address specified in step 3 as data
14 is output.

The mixers 314 and 315 output the data input to the L channel input terminal 301 and the R channel input terminal 302, the output data of the LPF forming the split units 304 and 305, and the chorus units 310 and 31 respectively.
The chorus data, which is the output of No. 1, is input, mixed, and output to the L channel output terminal 316 and the R channel output terminal 317.

Here, in the example shown in FIG. 4, the stereo input and stereo output of the L and R channels are used, but the processing is the same except that the input and output data of L and R are different.

That is, the split units 304 and 305,
Chorus block 310, 311, mixer 314,
Reference numeral 315 denotes a block having the same function. The outputs of the blocks having the same function have different values, but the data flows are the same.

The configuration of this conventional example is based on the reference page 5-20 of the MIDI sound source keyboard i4S sold by Korg Corporation, harmonic chorus. It is a guess of the operation that can be realized with the configuration shown in the figure.

FIG. 7 is a block diagram showing the configuration of another conventional example.

The basic configuration of this conventional example is the same as that of the conventional example shown in FIG. 4, so only the differences will be described.

In this conventional example, chorus units 518 and 519 and a modulation unit 520 for generating a read offset address are added to the conventional example shown in FIG.

The functions and operations of the chorus units 518 and 519 are the same as those of the chorus units 310 and 3 described in the conventional example shown in FIG.
It is equivalent to 11. The basic function and operation of the modulation unit 520 are the same as those of the modulation unit 303 described in the conventional example shown in FIG. 4, but the read offset address that has a chorus characteristic adapted to the sound of the low frequency component is set as described above. The difference is that the signal is output to the chorus units 518 and 519.

With the addition of the chorus units 310 and 311 and the modulation unit 520, a sound having a chorus characteristic adapted to the low frequency component is added to the low frequency component as compared with the conventional example shown in FIG. The high frequency component can output a sound having chorus characteristics adapted to the high frequency component.

[0044]

In the chorus processing performed by the conventional voice processing apparatus, a chorus characteristic is imparted to any one of a high frequency component, a low frequency component, and a middle frequency component.

When listening to music, the tune of the same song changes, such as a portion having many high-frequency components and a portion having many low-frequency components. There is a problem that the chorus processing by the conventional voice processing device cannot cope with such a change in the tune, and cannot provide a chorus characteristic adapted to the sound component.

The present invention has been made in view of the above-mentioned problems of the prior art, and is intended to analyze the frequency of an input sound and dynamically change the chorus characteristics based on the result. Thus, a sound processing device that can have different chorus characteristics corresponding to sound components, such as giving a thickness without changing the original component for the low range and making the sound clearer for the high range, etc. It is intended to realize a voice processing method.

[0047]

According to the present invention, there is provided an audio processing apparatus for performing audio processing on an original sound by changing a read address of a memory for storing a digital audio signal corresponding to the original sound. A frequency analysis unit that performs a frequency analysis of the digital audio signal,
A modulation unit that receives the analysis result of the frequency analysis unit and generates read address information according to the analysis result; and a chorus delay buffer as a memory that stores the input digital audio signal. A chorus section for reading and outputting data from a chorus delay buffer based on read address information to be read, and a mixer for adding and outputting an input digital audio signal and the output of the chorus section.

In this case, the frequency analysis unit includes a low-pass filter that allows only a low-frequency range of the input audio signal to pass therethrough;
A high-pass filter that passes only the high-frequency range, a comparison unit that compares the low-pass filter pass signal with the high-pass filter pass signal, and a parameter that is necessary when the modulation unit generates read address information. It may be configured to include a modulation table stored in correspondence with the comparison result, and a modulation value calculation unit that reads out the parameter corresponding to the comparison result of the comparison unit from the modulation table and outputs the parameter as an analysis result.

The frequency band may be divided into a plurality of frequency bands by an FFT analysis unit instead of the low-pass filter and the high-pass filter.

Further, instead of the low-pass filter and the high-pass filter, a plurality of frequency bands may be divided by band-pass filters having different pass-band widths.

The voice processing method of the present invention is a voice processing method for performing voice processing on an original sound by changing a read address of a memory for storing a digital voice signal corresponding to the original sound. Performs analysis, generates read address information according to the analysis result, and adds an output obtained when the content of the memory storing the input digital audio signal is read based on the read address information to the input digital audio signal. It is characterized by the following.

[Operation] In the present invention configured as described above, since the read address is determined based on the frequency analysis result of the input signal, it is possible to perform chorus processing suitable for the input signal. I have.

Performing a process suitable for an input signal by dividing such a sound component and changing the chorus characteristic for each frequency component has been conventionally performed as in the second conventional example shown in FIG. Although illustrated, many chorus sections are required in this case.

Since the chorus section is specifically configured using a random access memory, the memory size increases and the cost increases.

In the present invention, the read address is determined based on the result of the frequency analysis, and the chorus processing in accordance with the frequency band of the input signal is performed from the start of the memory read, so that the cost is low and the cost is high. Quality can be realized.

[0056]

Next, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of one embodiment of the present invention.

In this embodiment, the L channel input terminal 101 and the R channel input terminal 1 for inputting the L channel and the R channel of a signal whose original sound is digitally coded, respectively.
02, frequency analysis sections 104 and 105 for inputting input signals to the input terminals, a modulation section 103 for generating a signal for determining the chorus characteristic, an output of the modulation section 103, and input terminals 101 and 10 respectively.
2 and chorus sections 110 and 111 for generating chorus signals and input terminals 101 and 102, respectively.
Mixers 114 and 115 having inputs to the respective inputs and outputs of the chorus sections 110 and 111 as input, and an L channel output terminal 1 for outputting the outputs of the respective mixers 114 and 115
16 and an R channel output terminal 117.

As in the conventional example, the digitally input sound is input from the input terminals 101 and 102 and input to the frequency analysis units 104 and 105, the chorus units 110 and 111, and the mixers 114 and 115.

Frequency analysis section 104 analyzes the frequency of the input signal, and outputs the analysis result, ie, a parameter for generating modulation, to modulation section 103.

The modulation section 103 calculates a read offset address based on the parameters input from the frequency analysis blocks 104 and 105 in the same manner as in the conventional example, and outputs the read offset address to the chorus sections 110 and 111. The chorus units 110 and 111 input and write input data to the input terminals 101 and 102, and output data specified by the read offset address to the mixers 114 and 115 as chorus data.

Each of the mixers 114 and 115 mixes the output data from the chorus units 110 and 111 with the input data to the data input terminals 101 and 102 and outputs them to the output terminals 116 and 117.

Next, the detailed operation of each block will be described with reference to the drawings.

FIG. 2 is a block diagram showing the configuration of the frequency analysis units 104 and 105.

The frequency analysis units 104 and 105 provide the LPF 4
01, HPF 402, comparison section 403, modulation value calculation block 404, modulation table 40
5 is comprised.

The input data is LPF 401 and HPF 402
, And signals having different frequency bands after being processed by the respective filters are output to the comparing section 403. The comparison unit 403 compares the dynamic range (output level) of the input LPF 401 and the input HPF 402 and outputs the comparison result to the modulation value calculation block 40.
Output to 4.

The comparison in the comparing section 403 is performed by, for example, LP
If the output of F401 is large, '0' is set, and HPF40
If the output of 2 is large, '1' is output.

The modulation table 405 is shown in FIG.
The LFO, DEPTH and initial delay corresponding to the comparison result of the comparison unit 403 are stored as shown in FIG.
Is LF stored in the modulation table 405 corresponding to the comparison result indicated by the output of the comparison unit 403.
O, DEPTH and initial delay are read, and the modulation unit 3 shown in FIG.
03, the read offset address 107 is generated and output.

The functions of the chorus units 110 and 111 are the same as those of the conventional chorus units 310 and 311 shown in FIG. 4, and the operation is the same except that the input / output data is different. Since data input / output has already been described, it is not specified here.

In this embodiment constructed as described above, the frequency components of the input signal are analyzed, and the read address is determined by the LFO, DEPTH and initial delay according to the component. Different processing can be performed, and chorus processing suitable for input signal contents can be performed.

In the above-described embodiment, the level detection using the LPF and the HPF is taken as an example of the frequency component analysis method. However, other frequency analysis methods include a level detection method for each frequency band by FFT and a band detection method. A level detection method using a pass filter (hereinafter, referred to as a BPF) is applicable, and the frequency analysis method is not particularly limited.

When performing frequency analysis by FFT and BPF, an FFT analysis unit for detecting each frequency band and a plurality of BPFs each having a different passing frequency band
May be provided so as to be divided into different frequency bands and the respective levels are compared. Further, in this case, the contents stored in the modulation table 405 can be divided into more cases, and in such a configuration, more detailed chorus processing can be performed.

The modulation value calculation block 4
In 04, the output of the comparison unit 403 is input for each input, and the content read from the modulation table 405 is updated. However, when the frequency component of the input signal changes in a short time, the chorus characteristic also changes accordingly. Would. Therefore, the comparison method of the comparison unit 403 may be configured to compare the average value over a certain period of time so that the chorus characteristic is not changed during the certain period of time. With such a configuration, it is possible to prevent the chorus characteristic from changing in a short time, and it is more effective in hearing.

In order to listen to the output from each of the output terminals 116 and 117 as a normal sound, a D / A converter, a speaker, and the like are required. However, the description is omitted because it is not particularly important in the present invention.

The processing is the same except that the input and output data of L and R are different. That is, the frequency detection block blocks 104 and 105, the chorus delay blocks 110 and 111, and the mixing blocks 114 and 115 are blocks having the same function.
Although the values of data 107 and 108 and data 112 and 113 are different, the data flow is the same as shown in FIG.

[0075]

Since the present invention is configured as described above, it has the following effects.

Since the chorus processing is performed according to the frequency component of the input signal, the characteristics of the chorus are dynamically changed according to the input signal. For this reason, for example, there is an effect that it is possible to apply an arus characteristic corresponding to the sound component, such as giving a thickness without changing the original component in the low range and making the sound clearer in the high range. .

The chorus processing according to the frequency components as described above is performed by determining a read address of a memory for storing an original sound based on a frequency analysis result. As described above, since the chorus processing is performed in accordance with the input signal from the start of the voice reading, the number of required chorus portions can be reduced, and the cost can be reduced.
In addition, there is an effect that a high quality device can be realized.

[Brief description of the drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

FIG. 2 is a diagram showing a configuration of frequency analysis units 104 and 105 in FIG.

FIG. 3 is a diagram showing a configuration of a modulation table 405 in FIG.

FIG. 4 is a block diagram showing a configuration of a conventional example.

5A to 5C are diagrams illustrating output values of a modulation unit 303 in FIG.

6 is a diagram showing a configuration of chorus units 310 and 311 in FIG.

FIG. 7 is a block diagram showing a configuration of a conventional example.

[Explanation of symbols]

 101 L channel input terminal 102 R channel input terminal 103 Modulation unit 104, 105 Frequency analysis unit 110, 111 Chorus unit 114, 115 Mixer 116 L channel output terminal 117 R channel output terminal 401 Low pass filter 402 High pass filter 403 Comparison unit 404 Modulation value Calculation unit 405 Modulation table

Claims (5)

[Claims] 1. An audio processing apparatus for performing audio processing on an original sound by changing a read address of a memory for storing a digital audio signal corresponding to an original sound, wherein a frequency analysis unit performs a frequency analysis of the input digital audio signal. A modulation unit that receives an analysis result from the frequency analysis unit and generates read address information according to the analysis result; and a chorus delay buffer as a memory that stores the input digital audio signal. A chorus section for reading and outputting data from a chorus delay buffer in accordance with read address information output by the output section, and a mixer for adding and outputting an input digital audio signal and the output of the chorus section. apparatus. 2. The audio processing device according to claim 1, wherein the frequency analysis unit comprises: a low-pass filter that allows only a low-frequency range of the input audio signal to pass; a high-pass filter that allows only a high-frequency range to pass; A comparison unit that compares a pass signal with a pass signal of the high-pass filter; a modulation table that stores parameters necessary for generating read address information in a modulation unit corresponding to the comparison result of the comparison unit; A modulation value calculation unit that reads out the parameter corresponding to the comparison result of the unit from the modulation table and outputs the result as an analysis result. 3. The audio processing apparatus according to claim 2, wherein the FF is used instead of the low-pass filter and the high-pass filter.
An audio processing apparatus, wherein a frequency band is divided into a plurality of parts by a T analysis unit. 4. The audio processing apparatus according to claim 2, wherein the frequency band is divided into a plurality of bands by band-pass filters having different pass-band widths instead of the low-pass filter and the high-pass filter. apparatus. 5. An audio processing method for performing audio processing on an original sound by changing a read address of a memory for storing a digital audio signal corresponding to an original sound, wherein a frequency analysis of the input digital audio signal is performed. Generating a read address information in accordance with the result, and adding an output obtained when the content of the memory storing the input digital audio signal is read by the read address information to the input digital audio signal; Processing equipment.