JP2002023750A - Audio waveform data reproducing device, time base compression/expansion processor for audio waveform data, and pitch conversion processor for audio waveform data - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain high sound quality by eliminating factors for deteriorating the sound quality. SOLUTION: This device is provided with a storage means storing audio waveform data, a note information input means for inputting note information, a compression/ expansion information input means for inputting compression/expansion information, a time function generation means for generating a time function, a pitch information generation means for generating pitch information, a read address generation means for generating a read address to be changed at a speed corresponding to the note information, a control means for performing control so as to jump the read address for a value corresponding to the pitch information so as to turn the change of the read address to be within a change range corresponding to the pitch information set with the time function as a reference in the case that the read address is changed exceeding the change range, a read means for reading the audio waveform data from the storage means according to the read address controlled by the control means and a synthesis means for synthesizing reproducing waveform signals based on the audio waveform data read by the read means.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio waveform data reproducing apparatus, an audio waveform data time axis compression / expansion processing apparatus, and an audio waveform data pitch conversion processing apparatus. In the present specification, "compression or expansion of the time axis" of audio waveform data is referred to as "time axis compression / expansion" of audio waveform data as appropriate. The present invention relates to an audio waveform data reproducing apparatus capable of converting and reproducing the pitch of audio waveform data, a time axis compression / expansion processing apparatus for audio waveform data, and a pitch conversion processing apparatus for audio waveform data.

[0002]

2. Description of the Related Art Conventionally, in the field of electronic musical instruments,
There are known techniques for reproducing the audio waveform data by compressing or expanding the time axis, and for converting the pitch of the audio waveform data for reproduction.
JP-A-352970 discloses such a method.

That is, paragraph numbers “0013” to “0015” of JP-A-11-352970 include:
A technique described as "second compression / expansion method" in Japanese Patent Application Laid-Open No. 11-352970 is described with reference to FIG.

Here, the "second compression / expansion method" refers to reading audio waveform data corresponding to a reproduced pitch in accordance with position information indicating a position on a waveform that changes at a speed corresponding to the compression / expansion of the waveform. These are cut out at a speed, and they are connected by cross-fade processing to obtain a reproduced waveform signal that has been subjected to time axis compression / expansion and pitch conversion.

[0005] In addition, paragraphs [0016] to [0019] of JP-A-11-352970 include:
A method described as “third compression / expansion method” in JP-A-11-352970 is described with reference to FIG.

Here, the "third compression / expansion method" performs the time axis compression / expansion and pitch conversion of audio waveform data by the following method.

That is, in the waveform data, it is assumed that a mark section is assigned to a waveform portion in which a reproduced sound becomes natural even in a loop reproduction, and a waveform section is set. It is assumed that position information indicating a position on a changing waveform has been generated. In such a device, the waveform section is read at a read speed corresponding to the playback pitch, and when the read reaches the mark address at the end of the section, the waveform section indicated by the position information is reproduced. A reproduced waveform signal obtained by compressing / expanding the time axis of the waveform data and converting the pitch of the waveform data in accordance with the change speed of the position information has been obtained.

[0008] However, the above-mentioned conventional method has a problem that it is difficult to obtain high quality sound quality.

That is, in the "second compression / expansion method" disclosed in Japanese Patent Application Laid-Open No. H11-352970, a high-quality tone signal having a large noise component such as a cymbal tone and whose pitch is difficult to recognize is provided. Although it is possible to obtain sound quality, in the case of a tone signal in which the pitch is recognized, the phases of the audio waveform data synthesized in the cross-fade portion are not always matched, so that the sound signals cancel each other out. Alternatively, the sound quality may be degraded due to emphasis or the like.

On the other hand, in the "third compression / expansion method" disclosed in Japanese Patent Application Laid-Open No. 11-352970, when the change speed of the position information is slow, that is, when the expansion of the time axis is large, the same waveform section is considerably separated. It will be reproduced many times. For this reason, when the audio waveform data to be repeatedly reproduced in the next waveform section changes, the sound quality becomes such that the listener can clearly recognize the change.

[0011]

SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems of the prior art, and has as its object to compress and expand the time axis of audio waveform data and to adjust the pitch. Audio waveform data reproducing apparatus and audio waveform data that perform conversion processing and that can remove factors that degrade the sound quality of the conventional technology by simple processing, thereby obtaining high-quality sound quality And a pitch conversion processor for audio waveforms.

[0012]

In order to achieve the above object, according to the first aspect of the present invention, there is provided a storage means for storing a plurality of periods of audio waveform data, and a pitch representing a reproduced pitch. Pitch information input means for inputting information, companding information input means for inputting companding information representing the amount of compression / expansion on the time axis, and a speed corresponding to the companding information input by the companding information input means. A time function generating means for generating a time function that changes and represents an address position on the audio waveform data stored in the storage means; and periodic waveform data based on the address position indicated by the time function generated by the time function generating means. Pitch information generating means for generating pitch information representing the pitch of the pitch information, and a read address for generating a read address which changes at a speed corresponding to the pitch information input by the pitch information input means. And a read address generated by the read address generating means changes over a change range corresponding to pitch information generated by the pitch information generating means based on a time function generated by the time function generating means. Control means for controlling the jump of the read address by a value corresponding to the pitch information generated by the pitch information generating means so that the change of the read address falls within the change range; and Reading means for reading audio waveform data from the storage means in accordance with a read address controlled by the means; and synthesizing means for synthesizing a reproduced waveform signal based on the audio waveform data read by the reading means. Things.

According to the first aspect of the present invention, a reproduced waveform signal is synthesized based on audio waveform data read in accordance with a "read address" controlled to jump by the control means. In this case, the waveform data before the jump and the waveform data after the jump may be cross-fade and synthesized in order to remove the effect of the discontinuity of the waveform data generated at the jump portion.

The above-mentioned "read address" corresponds to a "read pointer PR" in an embodiment described later.

Further, the "change range" corresponding to the "pitch information" in the first aspect of the present invention is:
One pitch or a plurality of pitches may be used.

Further, the jump amount of the address in the "control means" according to the first aspect of the present invention may be one pitch or a plurality of pitches.

The plurality of pitches include an integer multiple of the pitch information, and also includes adding pitch information for an integer period.

According to a second aspect of the present invention, there is provided a storage means for storing a plurality of periods of audio waveform data, a pitch information input means for inputting pitch information representing a reproduced pitch, Companding information input means for inputting companding information representing the amount of compression / expansion of the shaft, and an audio waveform which changes at a speed corresponding to the companding information input by the companding information input means and is stored in the storage means A time function generating means for generating a time function representing an address position on the data; and a pitch information generating means for generating pitch information representing a pitch of the periodic waveform data based on the address position indicated by the time function generated by the time function generating means. Means, and changing at a speed corresponding to the pitch information input by the pitch information input means, separated from each other in accordance with the pitch information generated by the pitch information generating means. Read address generating means for generating a plurality of read addresses; and at least one of the plurality of read addresses generated by the read address generating means is generated by the pitch information generating means based on a time function generated by the time function generating means. When the read address changes beyond the change range corresponding to the pitch information to be output, the read address keeps the pitch separated state represented by the pitch information generated by the pitch information generating means, and Control means for controlling at least one of the plurality of read addresses to jump so as to fall within the range; read means for reading audio waveform data from the storage means in accordance with the read addresses controlled by the control means; Combining the reproduced waveform signal based on the audio waveform data read by the reading means. It is obtained to have a synthesizing means for.

According to the configuration of the invention described in claim 2 of the present invention, the jump by the "control means" in the invention described in claim 2 of the present invention generates, for example, two read addresses. If one of the read addresses changes beyond the change range corresponding to the pitch information, the read address that changes beyond that one pitch is separated from the other read addresses by one pitch, and a subsequent change in the read address occurs. A jump to a position falling within the change range corresponding to the pitch information may be performed, or a jump to a position falling within the change range at the same time while two read addresses are maintained for one pitch. There may be.

However, the "change range" corresponding to the "pitch information" in the second aspect of the present invention is:
One pitch or a plurality of pitches may be used.

Further, the jump amount of the address in the "control means" in the second aspect of the present invention may be one pitch or a plurality of pitches.

The above-mentioned plurality of pitches include a pitch that is an integral multiple of the pitch information, and also includes adding pitch information for an integer period.

[0023] The invention described in claim 3 of the present invention relates to any one of claim 1 and claim 2 of the present invention.
In the invention described in the paragraph, the control means compares an address represented by the time function generated by the time function generation means with a read address generated by the read address generation means, and the read address is determined by the pitch information generation means. Is controlled so as to jump to the read address when it is detected that the read address exceeds a change range corresponding to the pitch information generated by.

Further, the invention described in claim 4 of the present invention relates to any one of claim 1 and claim 2 of the present invention.
In the invention described in the paragraph, the control means includes a time function changing speed generated by the time function generating means, a read address changing speed generated by the read address generating means, and a pitch information generating means. The jump timing of the next read address is calculated based on the change range corresponding to the pitch information, and the read address generated by the read address generating means is jumped when the jump timing comes. It is something to do.

[0025] In the invention according to claim 5 of the present invention, in the invention according to any one of claims 1, 2, 3, or 4, the synthesizing means includes: A window function providing means for providing a window function to the audio waveform data read by the reading means; providing the window function having a minimum value at the time of controlling the read address by the control means; This is to prevent occurrence.

According to a sixth aspect of the present invention, there is provided a storage means for storing a plurality of periods of audio waveform data, and a companding information inputting means for inputting companding information representing the amount of compression / expansion on the time axis. Time function generating means that changes at a speed corresponding to the companding information input by the companding information input means and generates a time function representing an address position on the audio waveform data stored in the storage means; Pitch information generating means for generating pitch information representing the pitch of periodic waveform data based on the address position indicated by the time function generated by the time function generating means, and read address generating means for generating a read address that changes at a predetermined speed The read address generated by the read address generating means is controlled by the pitch information generating means based on the time function generated by the time function generating means. In the case where the read address changes within the change range corresponding to the change range corresponding to the generated pitch information, only the value corresponding to the pitch information generated by the pitch information generating means is changed so that the change of the read address is within the change range. Control means for controlling the read address to jump, read means for reading audio waveform data from the storage means in accordance with the read address controlled by the control means, and control means based on the audio waveform data read by the read means. And synthesizing means for synthesizing the reproduced waveform signal. It is an extension processing device.

According to a seventh aspect of the present invention, there is provided a storage means for storing a plurality of periods of audio waveform data, and a companding information input means for inputting companding information indicating the amount of compression / expansion on the time axis. Time function generating means that changes at a speed corresponding to the companding information input by the companding information input means and generates a time function representing an address position on the audio waveform data stored in the storage means; Pitch information generating means for generating pitch information representing the pitch of the periodic waveform data based on the address position indicated by the time function generated by the time function generating means, and pitch information generating means which change at a predetermined speed and are generated by the pitch information generating means. Address generating means for generating a plurality of read addresses separated by pitches represented by pitch information, and a plurality of read addresses generated by the read address generating means. When at least one of the dresses changes over a change range corresponding to pitch information generated by the pitch information generating means based on a time function generated by the time function generating means, the read address is changed to the pitch information. Control for jumping at least one of the plurality of read addresses so that the change is within the change range while maintaining the state of pitch separation represented by the pitch information generated by the generating means. Means, according to the read address controlled by the control means,
Reading means for reading audio waveform data from the storage means; and synthesizing means for synthesizing a reproduced waveform signal based on the audio waveform data read by the reading means, so as to correspond to the compression / decompression information. This is a time axis compression / expansion processing apparatus for audio waveform data which performs time axis compression / expansion processing on audio waveform data stored in the storage means with high sound quality. In this configuration, a plurality of read addresses separated by a pitch representing pitch information are always generated.

[0028] In the invention according to claim 8 of the present invention, pitch conversion information input means for inputting pitch conversion information representing an amount of pitch to be changed, and audio waveform data are input and changed at a predetermined speed. Storage means for storing the audio waveform data in accordance with the address; time function generating means for changing at the predetermined speed to generate a time function representing an address position on the audio waveform data stored in the storage means; Pitch detection means for detecting pitch information of audio waveform data, read address generation means for generating a read address that changes at a speed corresponding to the pitch conversion information, and a read address exceeding a change range corresponding to the pitch information When the read address changes, the read address corresponds to the pitch information so that the change is within the change range. That the said read out address by a value as a control means for controlling to jump,
An audio waveform data pitch conversion apparatus for performing a pitch conversion process on audio waveform data with high sound quality in accordance with the pitch conversion information.

According to a ninth aspect of the present invention, a pitch conversion information input means for inputting pitch conversion information representing an amount of a pitch to be changed, and audio waveform data are input, and change at a predetermined speed. Storage means for storing the audio waveform data according to the address; time function generating means for changing at the predetermined speed to generate a time function representing an address position on the audio waveform data stored in the storage means; Pitch detection means for detecting pitch information of audio waveform data, read address generation means for generating a plurality of read addresses which change at a predetermined speed and are separated by pitches represented by the pitch information, and at least one of the read addresses If the read address changes beyond the change range corresponding to the pitch information, the read address becomes While maintaining a minute away from the state, and,
Control means for performing control so as to jump at least one of the plurality of read addresses so that the change falls within the change range, so that the audio waveform data is increased in accordance with the pitch conversion information. This is a pitch conversion processing device for audio waveform data that undergoes pitch conversion processing with sound quality. In this configuration, a plurality of read addresses separated by a pitch representing pitch information are always generated.

In the pitch conversion processing device according to the ninth aspect of the present invention, the input audio waveform data can be pitch-converted and output in real time.

It is to be noted that the present invention can be applied to a case where one audio waveform data can be simultaneously reproduced at different pitches, in which case the pitch and the reproduction position are controlled independently. It becomes possible.

Further, the present invention can be applied to an audio waveform reproduction technique using a time axis compression / expansion technique.

The present invention removes the cause of the sound quality deterioration by controlling the jump of the read address between the waveform data portions having substantially the same phase, so that a desired effect can be obtained in the audio waveform data which is a periodic signal. Obtainable.

Further, in the present invention, the method of reproducing the audio waveform data is not limited, and various methods can be used.

[0035]

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a block diagram showing an embodiment of an audio waveform data reproducing apparatus, an audio waveform data time axis compression / expansion processing apparatus, and an audio waveform pitch conversion processing apparatus according to the present invention; Will be described in detail.

FIG. 1 is a block diagram showing an embodiment of an audio waveform data reproducing apparatus according to the present invention.

This audio waveform data reproducing apparatus includes a waveform data storage means 10 for storing audio waveform data.
And an audio waveform reproduction processing unit 12 for performing a reproduction process of reading and reproducing the audio waveform data stored in the waveform data storage unit 10, and a control parameter when the audio waveform reproduction processing unit 12 performs the reproduction process. And a parameter input means 14 for performing the operation.

Here, the audio waveform reproduction processing means 12
Is composed of a microcomputer,
The audio waveform reproduction processing means 12 has the parameter input means 14 as a control parameter to read out the audio waveform data stored in the waveform data storage means 10 and reproduce the audio waveform data. Rate) and compression / expansion information (T) indicating the amount of compression / expansion when the audio waveform data is compressed / expanded on the time axis.
i) is input.

That is, the audio waveform reproduction processing means 12
Reproduces the audio waveform data stored in the waveform data storage means 10 using the pitch information and the companding information input from the parameter input means 14, and reproduces the audio waveform data based on the processed audio waveform data. A desired waveform signal (reproduced waveform signal) is synthesized and output.

FIG. 2 is a waveform diagram showing an example of audio waveform data stored in the waveform data storage means 10.

As shown in FIG. 2, the audio waveform data stored in the waveform data storage means 10 includes mark information (Mark0) representing each waveform section for each waveform cycle of the audio waveform data. , Mark1,
Mark2,...) Are stored.

Next, FIGS. 3 and 4 show an outline of audio waveform data reproduction processing executed by the audio waveform reproduction processing means 12 described above.

More specifically, when the reproduced pitch is the same as the original reproduced waveform (the "original reproduced waveform" is a waveform represented by the audio waveform data stored in the waveform data storage means 10). 3 shows a case where the audio waveform data indicating the original playback waveform is expanded on the time axis, and FIG. 4 shows a case where the audio waveform data indicating the original playback waveform is compressed on the time axis. Is shown in

In FIGS. 3 and 4, the dashed line indicating the change of the original address indicates the range from the start address (start address) to the end address (end address) of the audio waveform data indicating the original playback waveform. An original address change indicating an address change is indicated.
0 indicates a read address change indicating an address change when the audio waveform reproduction processing means 12 reads out and reproduces the audio waveform data stored in 0, and a solid line indicates a time function and is drawn in parallel with the solid line. The dashed line indicates an address change separated from the time function by an address corresponding to one cycle of the original waveform.

When the reproduced pitch is the same as the pitch of the original reproduced waveform, the slope of the change of the read address matches the slope of the change of the original address in FIGS.

In the example shown in FIGS. 3 and 4, when the audio waveform data stored in the waveform data storage means 10 is read out and reproduced, the audio data stored in the waveform data storage means 10 is read. The waveform data is read alternately by the first processing channel and the second processing channel, and cross-fade processing is performed when the reading is switched. 3 and 4, a portion indicated by a broken line on the extension of the alternate long and short dash line represents a reading section (cross-fade section) of audio waveform data for cross-fade processing.

Here, when changing the reproduction pitch,
The slope of the read address change may be changed so as not to match the slope of the original address change, that is, the read speed may be changed in accordance with the pitch information.

As shown in FIGS. 3 and 4, the read address jumps to a portion one cycle before or one cycle after the original reproduction waveform. Hereinafter, this point will be described in detail.

First, in the original reproduced waveform shown in FIG. 3, there is one cycle between the address A1 and the address A2, and between the address A3 and the address A4, hereinafter, the address A5 and the address A6. Address A7
The control is performed so that the interval between the respective jumping portions, such as between the address and the address A8, is one cycle.

In the case of the expansion on the time axis shown in FIG. 3, reading of the audio waveform data is started from the start address by the first processing channel, and the read address is changed to the address A2 which is one cycle away from the time function. When the processing advances, the processing jumps to the address A1 one cycle before the address A2 and returns.
From 1, audio waveform data is read. When the read address advances to the address A4 one cycle away from the time function, the read address jumps back to the address A3 one cycle before the address A4, and the audio waveform data is read from the address A3 by the first processing channel. By repeating such processing, audio waveform data is read out to the end address.

Also, in the original reproduction waveform shown in FIG. 4, there is one cycle between the address B1 and the address B2, as in the original reproduction waveform shown in FIG.
In the following, control is performed so that the interval between each of the jumps between the address B3 and the address B4 and between the address B5 and the address B6 is one cycle.

In the case of the compression on the time axis shown in FIG. 4, the reading of the audio waveform data is started from the start address by the first processing channel, and the read address is changed to the address B1 which is one cycle away from the time function. When the process proceeds, the process jumps to the address B2 one cycle after the address B1 and proceeds, and the address B is set by the second processing channel.
2 to read audio waveform data. When the read address advances to the address B3 one cycle away from the time function, the read address jumps to the address B4 one cycle after the address B3, and the audio waveform data is read from the address B4 by the first processing channel. By repeating such processing, audio waveform data is read out to the end address.

As described above, in the case of expansion on the time axis, the read address is jumped back to a portion one cycle before the original reproduction waveform, and in the case of compression on the time axis, the read address is expanded. Is jumped to the part after one cycle of the original reproduction waveform, so that when the reproduced waveform signal is synthesized based on the read audio waveform data, cancellation or emphasis due to a phase difference does not occur.

Then, the reproduced waveform signal thus synthesized is output from the audio waveform reproduction processing means 12.

Next, FIG. 5 shows a flowchart of a reproduction processing routine executed by the audio waveform reproduction processing means 12 in order to realize the processing contents described with reference to FIGS. FIGS. 6 and 7 show a flowchart of a subroutine of the reproduction processing routine shown in FIG.

Here, the reproduction processing routine shown in FIG.
It is started as an interruption process for each sampling period, and controls audio waveform data by pitch information (Rate) and companding information (Ti) input from the parameter input means 14. As a result, a reproduced waveform signal is obtained.

In the processing of the reproduction processing routine shown in FIG. 5, a read pointer RP indicated by a chain line in FIGS. 3 and 4 is used.

The actual read address includes the one-dot chain line and the broken line in FIGS.

Here, as an initial setting when the reproduction processing routine shown in FIG.
The start address of the audio waveform data stored in the waveform data storage means 10 is set, and the pitch of the first waveform of the audio waveform data is set in the pitch information CL.

When the reproduction processing routine shown in FIG. 5 is started, first, audio waveform data stored in the waveform data storage means 10 is read according to the read pointer RP (step S502).

Next, the audio waveform data is output from the audio waveform reproduction processing means 12 based on the audio waveform data read in the previous step, including performing the cross-fade processing in the cross-fade section shown in FIGS. Power waveform data (output waveform data) to be calculated is calculated (step S504).

Further, the time function CtrAdr is updated by adding the companding information Ti to the time function CtrAdr (step S506), and the read pointer RP is updated by adding the pitch information Rate to the read pointer RP (step S506). S508).

Next, steps S506 to S5
It is determined whether or not the absolute value of the difference between the time function CtrAdr updated in 08 and the read pointer RP is equal to or greater than the pitch information CL (step S510).

Here, when it is determined that the absolute value of the difference between the time function CtrAdr and the read pointer RP is equal to or greater than the pitch information CL, a routine for updating the pitch information CL, which is a subroutine of this reproduction processing routine, is performed. (FIG. 6) is activated and executed (step S512).
The control routine (FIG. 7) of the read pointer RP, which is a subroutine of the reproduction processing routine, is started and executed (step S514), and the reproduction processing routine ends.

On the other hand, when it is determined that the absolute value of the difference between the time function CtrAdr and the read pointer RP is not greater than the pitch information CL, the reproduction processing routine is terminated.

Next, a routine for updating the pitch information CL executed in step S512 of the above-described reproduction processing routine will be described with reference to FIG.

When the pitch information CL updating routine is started, first, based on the time function CtrAdr,
Mark information corresponding to Mark0, Mark1, and Mark2 is extracted from the audio waveform data (step S602). At this time, the time function CtrAdr
Mark information of the waveform section to which Mark belongs and Mark 0 and Mark
1, and mark information of the next waveform section is Mark1 and Mark2.

Next, the pitch of each waveform section is calculated from the mark information corresponding to Mark 0, Mark 1, and Mark 2 (step S604). Specifically, the pitch information C is obtained by subtracting Mark0 from Mark1.
L0 is calculated, and Mark1 is subtracted from Mark2 to calculate pitch information CL1.

The interval of the mark information and the time function Ct
The pitch information CL is calculated by interpolation from the position of rAdr (step S606).

Specifically, CL0 + ((CtrAdr-Mark0) / CL0) ×
The calculation of (CL1-CL0) is performed, and the calculation result is set as pitch information CL.

When the processing for calculating the pitch information CL ends, the routine for updating the pitch information CL ends, and the process returns to the reproduction processing routine.

Next, a routine for controlling the read pointer RP executed in step S514 of the above-described reproduction processing routine will be described with reference to FIG.

When the control routine of the read pointer RP is started, first, the companding information Ti and the pitch information Rate are read.
Are compared (step S702). Then, the read pointer RP is controlled according to the comparison result.

That is, if the result of the comparison between the companding information Ti and the pitch information Rate is equal to or greater than the pitch information Rate, the readout is performed by adding the pitch information CL to the read pointer RP. The pointer RP is updated (step S704).

On the other hand, if the result of the comparison between the companding information Ti and the pitch information Rate indicates that the companding information Ti is less than the pitch information Rate, the pitch information CL is read from the read pointer RP.
Is updated to update the read pointer RP (step S706).

When the process of updating the read pointer RP is completed as described above, the read pointer R
The control processing routine of P is terminated, and the process returns to the reproduction processing routine.

Next, FIGS. 8 and 9 show an outline of another method of reproducing the audio waveform data executed by the audio waveform reproduction processing means 12 described above.

In the cases of FIGS. 8 and 9, FIG.
4 to 4, the reproduced pitch is the same as the pitch of the original reproduced waveform. That is, the slope of the read address change and the slope of the original address change match.

In the examples shown in FIGS. 8 and 9, when the audio waveform data stored in the waveform data storage means 10 is read out and reproduced, the audio data stored in the waveform data storage means 10 is read. Waveform data is read out simultaneously by the first processing channel and the second processing channel.

FIG. 8 shows a case where the audio waveform data indicating the original reproduced waveform is expanded on the time axis, and a case where the audio waveform data indicating the original reproduced waveform is compressed on the time axis. Is shown in FIG.

In FIGS. 8 and 9, a broken line described as an original address change indicates an original address change indicating an address change from a start address to an end address of audio waveform data indicating an original reproduction waveform. The dashed line indicates the address (first address) when the audio waveform reproduction processing means 12 reads out the audio waveform data stored in the waveform data storage means 10 through the first processing channel and performs the reproduction processing.
A first read address change indicating a change of the read address is shown. A two-dot chain line indicates that the audio waveform data stored in the waveform data storage means 10 is read by the audio waveform reproduction processing means 12 through the second processing channel. A second read address change representing a change in an address (second read address) at the time of reproduction processing is shown. A solid line indicates a time function, and two dashed lines drawn in parallel with the solid line indicate the time function from the time function. A change of each address separated by an address corresponding to one cycle before and after the waveform is shown.

Here, when changing the reproduction pitch,
The slope of the first read address change and the slope of the second address change may be changed so as not to coincide with the slope of the original address change, that is, the read speed may be changed in accordance with the pitch information.

In the decompression process on the time axis shown in FIG. 8, a first window function is added to the waveform data read at the first read address, and the first read address is one cycle from the time function. When advancing to a distant address, a jump is made so as to return from the advanced address to the address two cycles before by sandwiching the time function. Similarly, a second window function is added to the waveform data read at the second read address having a phase different from the first read address by one cycle address, and the second read address is an address one cycle away from the time function. When the process has progressed to, the jump is made so as to return from the advanced address to the address two cycles before by sandwiching the time function.

In the compression processing on the time axis shown in FIG. 9, a first window function is added to the waveform data read at the first read address, and the first read address is delayed from the time function. When advancing to an address one cycle away, the jump jumps from the advanced address to the address two cycles later with the time function interposed. Similarly, a second window function is added to the waveform data read at the second read address having a phase different from the first read address by one cycle address, and the second read address is delayed by 1 from the time function.
When advancing to an address separated by a cycle, the jump is made so as to advance from the advanced address to an address two cycles later by sandwiching the time function.

Then, a reproduced waveform signal is synthesized based on the audio waveform data read by the first read address and the second read address as described above, and the reproduced waveform signal is output from the audio waveform reproduction processing means 12. What is output.

Next, FIGS. 10 and 11 show an outline of another method of reproducing the audio waveform data executed by the audio waveform reproduction processing means 12 described above.

FIG. 10 shows a case where the audio waveform data representing the original reproduced waveform is expanded on the time axis.
FIG. 1 shows a case where audio waveform data indicating an original reproduced waveform is compressed on the time axis.
It is shown in FIG.

FIGS. 10 and 11 show only the differences between the shapes of the first window function and the second window function and the respective change ranges of the first read address and the second read address. It is different from FIG.

Next, FIG. 12 shows a flowchart of a reproduction processing routine executed by the audio waveform reproduction processing means 12 in order to realize the processing contents shown in FIGS. 13 and 14 show a flowchart of a subroutine of the reproduction processing routine shown in FIG.

Here, the reproduction processing routine shown in FIG. 12 is started as interruption processing for each sampling period, and the pitch information (Rate) and companding information (Ti) input from the parameter input means 14 are input. Controls the audio waveform data. As a result, a reproduced waveform signal is obtained.

In the processing of the reproduction processing routine shown in FIG. 12, as described later, the companding information Ti and the pitch information Rate are updated every window function cycle and are reflected in the reproduction processing routine. However, even with such a period, there is no problem in practical use.

Here, as an initial setting when the reproduction processing routine shown in FIG. 12 is started, the time function CtrAdr
Is set to the start address of the audio waveform data stored in the waveform data storage means 10, and the address control position Pw is set to the waveform data storage means 1.
0 is set to the start address of the audio waveform data stored in the first read address RAdr.
In 1, the start address of the audio waveform data stored in the waveform data storage means 10 is set.

When the reproduction processing routine shown in FIG. 12 is started, it is first determined whether or not the time function CtrAdr is equal to or greater than the address control position Pw (step S1202).

If it is determined that the time function CtrAdr is equal to or greater than the address control position Pw, the process proceeds to step S1204, where the companding information Ti and the pitch information R
and the companding information Ti is set as companding information MTi, and the pitch information Rate is the pitch information MRa.
Set as te.

Next, a routine for updating the pitch information CL which is a subroutine of the reproduction processing routine (FIG. 13)
Is started and executed (step S1206). This subroutine will be described later.

Then, a process of updating the next address control position Pw is performed (step S1208). Specifically, an operation of Pw + CL / | MTi-Mrate | is performed, and the operation result is set as the address control position Pw.

Next, a control processing routine (FIG. 14) for the first read address RAdr1 and the second read address RAdr2, which is a subroutine of the reproduction processing routine, is started and executed (step S1210). This subroutine will be described later.

When the processing in step S1210 is completed, the flow advances to step S1212.

If it is determined in step S1202 that the time function CtrAdr is not equal to or greater than the address control position Pw, the process proceeds to step S1212.
Jump to the processing of and proceed.

Here, in the processing of step S1212, the audio waveform data stored in the waveform data storage means 10 is read according to the first read address RAdr1 and the second read address RAdr2.

Next, processing for calculating the window function W is performed with reference to the window function table shown on the right side of the flowchart in FIG. 12 (step S1214). Specifically, WindowValue ((Pw-CtrAdr) / C
L) is calculated, and the calculation result is set as a window function W.

Further, audio waveform reproduction processing means 12
For calculating output waveform data Wave, which is audio waveform data to be output from the device (step S121)
6). Specifically, a calculation of W × Data1 + (1−W) × Data2 is performed, and the calculation result is output waveform data Wave.
And

Then, the time function CtrAdr is updated by adding the companding information MTi to the time function CtrAdr, and the pitch information MRate is added to the first read address RAdr1 to obtain the first read address RA.
dr1 is updated, and the pitch information MRate is added to the second read address RAdr2 to update the second read address RAdr1 (step S1218), and the reproduction processing routine ends.

Next, a routine for updating the pitch information CL executed in step S1206 of the above-described reproduction processing routine will be described with reference to FIG.

When the pitch information CL update processing routine is started, first, based on the time function CtrAdr,
Mark information corresponding to Mark0, Mark1, and Mark2 is extracted from the audio waveform data (step S1302). At this time, the time function CtrAd
The mark information of the waveform section to which r belongs is Mark 0 and Mark
k1 and mark information of the next waveform section is Mark1
And Mark2.

Next, the pitch of each waveform section is calculated from the mark information corresponding to Mark 0, Mark 1, and Mark 2 (step S1304). Specifically, the pitch information C is obtained by subtracting Mark0 from Mark1.
L0 is calculated, and Mark1 is subtracted from Mark2 to calculate pitch information CL1.

The interval of the mark information and the time function Ct
The pitch information CL is calculated by interpolation from the position of rAdr (step S1306).

More specifically, CL0 + ((CtrAdr-Mark0) / CL0) ×
The calculation of (CL1-CL0) is performed, and the calculation result is set as pitch information CL.

When the processing for calculating the pitch information CL ends, the routine for updating the pitch information CL ends, and the process returns to the reproduction processing routine.

Next, with reference to FIG. 14, the first read address RAdr1 and the second read address RA executed in step S1210 of the above-described reproduction processing routine.
The routine of the control process with dr2 will be described.

When the control processing routine for the first read address RAdr1 and the second read address RAdr2 is started, first, the companding information MTi is compared with the pitch information MRate (step S1402). Then, the first read address RAdr1 and the second read address RAdr2 are controlled according to the comparison result.

That is, the companding information MTi and the pitch information MRat
e, the companding information MTi is the pitch information MRat
e, the second read address RAdr2
Is substituted for the first read address RAdr1 to update the first read address RAdr1, and the pitch information CL is added to the first read address RAdr1 to update the second read address RAdr2 (step S1404).

On the other hand, companding information MTi and pitch information MRat
e, the companding information MTi is the pitch information MRat
e, the second read address RAdr2
Is substituted for the first read address RAdr1 to update the first read address RAdr1, and the second read address RAdr2 is updated by subtracting the pitch information CL from the first read address RAdr1 (step S1406).

When the processing for updating the first read address RAdr1 and the second read address RAdr2 is completed as described above, the first read address RAdr is updated.
The control routine for the first and second read addresses RAdr2 ends, and the process returns to the reproduction processing routine.

Next, FIG. 15 shows a flowchart of a reproduction processing routine according to another method executed by the audio waveform reproduction processing means 12. FIGS. 16 and 17 show a flowchart of a subroutine of the reproduction processing routine shown in FIG.

Here, the reproduction processing routine shown in FIG. 15 is started as an interruption processing for each sampling cycle, and the pitch information (Rate) and companding information (Ti) input from the parameter input means 14 are input. Controls the audio waveform data. As a result, a reproduced waveform signal is obtained.

In the processing of the reproduction processing routine shown in FIG. 15, as will be described later, the companding information Ti and the pitch information Rate are updated at a sampling cycle and reflected in the reproduction processing routine. Here, as an initial setting when the reproduction processing routine shown in FIG. 15 is started, the start address of the audio waveform data stored in the waveform data storage means 10 is set in the time function CtrAdr. In addition, the start address of the audio waveform data stored in the waveform data storage unit 10 is set in the first read address RAdr1, and the start address of the audio waveform data stored in the waveform data storage unit 10 is set in the second read address RAdr2. The start address of audio waveform data is set.
An appropriate constant is set in the pitch information CL.

When the reproduction processing routine shown in FIG. 15 is started, first, the first read address RAdr1 is read.
Is larger than the time function CtrAdr, or whether the second read address RAdr2 is smaller than the time function CtrAdr (step S1502).

Here, if it is determined that the first read address RAdr1 is larger than the time function CtrAdr, or if the second read address RAdr2 is larger than the time function CtrAdr.
If it is determined that it is smaller than dr, step S15
04, and a routine for updating the pitch information CL, which is a subroutine of the reproduction processing routine (FIG. 16).
Is started and executed (step S1504). This subroutine will be described later.

Then, a control processing routine (FIG. 17) for the first read address RAdr1 and the second read address RAdr2, which is a subroutine of the reproduction processing routine, is started and executed (step S1506). This subroutine will be described later.

When the processing in step S1506 ends, the flow advances to step S1508.

On the other hand, in the process of step S1502, the first read address RAdr1 is set to the time function CtrAd.
If it is not determined that the value is greater than r, and if it is not determined that the second read address RAdr2 is smaller than the time function CtrAdr, the process jumps to step S1508 and proceeds.

Here, in the process of step S1508, audio waveform data stored in waveform data storage means 10 is read according to first read address RAdr1 and second read address RAdr2.

Next, referring to a predetermined window function table,
A process of calculating the window function W is performed (step S151)
0). Specifically, WindowValue (1- (CtrAdr-RAd)
r1) / CL) is calculated, and the calculation result is set as a window function W.

Further, the audio waveform reproduction processing means 12
For calculating the output waveform data Wave which is the audio waveform data to be output from the device (step S151)
2). Specifically, a calculation of W × Data1 + (1−W) × Data2 is performed, and the calculation result is output waveform data Wave.
And

Then, the time function CtrAdr is updated by adding the companding information Ti to the time function CtrAdr, and the pitch information R is added to the first read address RAdr1.
a, the first read address RAdr
1 is updated, and the pitch information Rate is added to the second read address RAdr2, thereby obtaining the second read address Rdr2.
Adr1 is updated (step S1514), and this reproduction processing routine ends.

Next, a routine for updating the pitch information CL executed in step S1504 of the above-described reproduction processing routine will be described with reference to FIG.

When the routine for updating the pitch information CL is started, first, based on the time function CtrAdr,
Mark information corresponding to Mark0, Mark1, and Mark2 is extracted from the audio waveform data (step S1602). At this time, the time function CtrAd
The mark information of the waveform section to which r belongs is Mark 0 and Mark
k1 and mark information of the next waveform section is Mark1
And Mark2.

Next, the pitch of each waveform section is calculated from the mark information corresponding to Mark 0, Mark 1, and Mark 2 (step S1604). Specifically, the pitch information C is obtained by subtracting Mark0 from Mark1.
L0 is calculated, and Mark1 is subtracted from Mark2 to calculate pitch information CL1.

Then, the interval of the mark information and the time function Ct
The pitch information CL is calculated by interpolation from the position of rAdr (step S1606).

Specifically, CL0 + ((CtrAdr-Mark0) / CL0) ×
The calculation of (CL1-CL0) is performed, and the calculation result is set as pitch information CL.

When the processing for calculating the pitch information CL ends, the routine for updating the pitch information CL ends, and the process returns to the reproduction processing routine.

Next, referring to FIG. 17, the first read address RAdr1 and the second read address RA executed in step S1506 of the above-described reproduction processing routine will be described.
The routine of the control process with dr2 will be described.

When the control processing routine for the first read address RAdr1 and the second read address RAdr2 is started, first, the time function CtrAdr is changed to the second read address Rdr.
It is determined whether it is larger than Adr2 or whether the first read address RAdr1 is larger than the time function CtrAdr (step S1702).

Here, when it is determined that the time function CtrAdr is larger than the second read address RAdr2,
The second read address RAdr2 is changed to the first read address RAd.
r1 to the first read address RAdr1
Is updated, and the pitch information CL is added to the first read address RAdr1 to thereby obtain the second read address Rdr1.
Adr2 is updated (step S1704).

On the other hand, if it is determined that the first read address RAdr1 is larger than the time function CtrAdr, the second read address RAdr2 is changed to the first read address RAdr.
1 to update the first read address RAdr1 and subtract the pitch information CL from the first read address RAdr1 to update the second read address Rdr1.
Adr2 is updated (step S1706).

When the processing for updating the first read address RAdr1 and the second read address RAdr2 is completed as described above, the first read address RAdr is updated.
The control routine for the first and second read addresses RAdr2 ends, and the process returns to the reproduction processing routine.

FIGS. 18 and 19 are waveform diagrams showing audio waveform data reproduction processing executed by the audio waveform reproduction processing means 12 by the method of the flowcharts shown in FIGS.

FIG. 18 shows a case where audio waveform data representing the original playback waveform is expanded on the time axis.
FIG. 1 shows a case where audio waveform data indicating an original reproduced waveform is compressed on the time axis.
9.

FIGS. 18 to 19 show only the differences between the shapes of the first window function and the second window function and the respective change ranges of the first read address and the second read address. It is different from FIG.

FIG. 20 is a block diagram showing an embodiment of an audio waveform pitch conversion processing apparatus according to the present invention.

This audio waveform pitch conversion processing device includes an audio waveform reproduction processing means 20, an analog / digital converter (A / D) 22, a random access memory (RAM) 24, a pitch detection means 26, Child 2
8 is provided.

Here, the audio waveform reproduction processing means 20
The same as the audio waveform reproduction processing means 12 described above can be used. However, in this audio waveform pitch conversion processing device, since the audio waveform data is not compressed and expanded on the time axis, the expansion / compression information (Ti) is fixed at “Ti = 1”.

In the audio waveform pitch conversion processing device, the time function CtrAdr is stored in the RAM.
It is preferable that a predetermined delay (DAdr) is added from the input waveform data write address (WAdr) to the H.24.

That is, the time function CtrAdr may be obtained by the following equation: CtrAdr = WAdr-DAdr.

The above-mentioned predetermined delay (DAdr)
Is set to a value that does not exceed the write address (Wdr) even if the speed of the read address changes due to the pitch conversion processing.

The above-described embodiment can be modified as shown in the following (1) to (3).

(1) In the above embodiment, the change range of the read address is one cycle (one pitch). However, the present invention is not limited to this. Good. Here, the multiple cycle is
The pitch information may be an integral multiple of the pitch information, or pitch information for a plurality of cycles may be added.

(2) In the above embodiment, a triangular window is used as the window function. However, the present invention is not limited to this, and the shape of the window function may be a Hamming window, a Hanning window, or the like. Alternatively, a trapezoidal window or the like may be appropriately set according to the purpose.

(3) The above-described embodiments and the above-described modifications (1) and (2) may be appropriately combined.

[0151]

Since the present invention is configured as described above, it performs time axis compression / expansion and pitch conversion processing of audio waveform data and deteriorates the sound quality of the prior art by simple processing. This makes it possible to eliminate the factors and achieve an excellent effect that high-quality sound quality can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an example of an embodiment of an audio waveform data reproducing apparatus according to the present invention.

FIG. 2 is a waveform diagram of an example of audio waveform data stored in a waveform data storage unit.

FIG. 3 is a waveform diagram showing an outline of audio waveform data reproduction processing executed by audio waveform reproduction processing means, showing a case where audio waveform data is expanded on a time axis.

FIG. 4 is a waveform diagram showing an outline of audio waveform data reproduction processing executed by audio waveform reproduction processing means, showing a case where audio waveform data is compressed on a time axis.

FIG. 5 is a flowchart of a reproduction processing routine executed by an audio waveform reproduction processing means.

FIG. 6 is a flowchart of a routine for updating pitch information CL.

FIG. 7 is a flowchart of a control processing routine of a read pointer RP.

FIG. 8 is a waveform chart showing an outline of another method of audio waveform data reproduction processing executed by audio waveform reproduction processing means, and shows a case where audio waveform data is expanded on a time axis.

FIG. 9 is a waveform chart showing an outline of another method of reproducing audio waveform data executed by the audio waveform reproduction processing means, and shows a case where audio waveform data is compressed on a time axis.

FIG. 10 is a waveform diagram showing an outline of still another method of audio waveform data reproduction processing executed by audio waveform reproduction processing means, showing a case where audio waveform data is expanded on a time axis.

FIG. 11 is a waveform diagram showing an outline of still another method of audio waveform data reproduction processing executed by audio waveform reproduction processing means, showing a case where audio waveform data is compressed on a time axis.

FIG. 12 is a flowchart of a reproduction processing routine executed by audio waveform reproduction processing means.

FIG. 13 is a flowchart of a routine of a process of updating pitch information CL.

FIG. 14 is a flowchart of a control processing routine for a first read address RAdr1 and a second read address RAdr2.

FIG. 15 is a flowchart of a reproduction processing routine executed by an audio waveform reproduction processing means.

FIG. 16 is a flowchart of a routine for updating pitch information CL.

FIG. 17 is a flowchart of a control processing routine for a first read address RAdr1 and a second read address RAdr2.

FIG. 18 is a waveform chart showing audio waveform data reproduction processing executed by the audio waveform reproduction processing means by the method of the flowcharts shown in FIGS. 15 to 17, showing a case where audio waveform data is expanded on a time axis. ing.

FIG. 19 is a waveform diagram showing audio waveform data reproduction processing executed by the audio waveform reproduction processing means by the method of the flowcharts shown in FIGS. 15 to 17, showing a case where audio waveform data is compressed on a time axis. ing.

FIG. 20 is a block diagram showing an example of an embodiment of an audio waveform pitch conversion processing device according to the present invention.

[Explanation of symbols]

 Reference Signs List 10 waveform data storage means 12 audio waveform reproduction processing means 14 parameter input means 20 audio waveform reproduction processing means 22 analog / digital converter (A / D) 24 random access memory (RAM) 26 pitch detection means 28 operator

 ──────────────────────────────────────────────────続 き Continued on the front page F term (reference) 5D045 AC02 5D378 AD22 AD45 AD51 AD70 BB12 FF11 5J064 AA01 BB10 BC01 BC02 BC06 BC24 BD03

Claims (9)

[Claims] 1. A storage means for storing a plurality of periods of audio waveform data, a pitch information input means for inputting pitch information representing a reproduced pitch, and a companding information representing a time axis compression / expansion amount. Companding information input means; and a time which changes at a speed corresponding to the companding information input by the companding information input means and generates a time function representing an address position on audio waveform data stored in the storage means. Function generating means, pitch information generating means for generating pitch information representing the pitch of the periodic waveform data based on the address position indicated by the time function generated by the time function generating means, and input by the pitch information input means. Read address generating means for generating a read address that changes at a speed corresponding to pitch information; read address generated by the read address generating means;
When the read address changes within the change range when the change exceeds the change range corresponding to the pitch information generated by the pitch information generation means based on the time function generated by the time function generation means. Control means for controlling the read address to jump by a value corresponding to the pitch information generated by the pitch information generating means; and audio waveform data from the storage means in accordance with the read address controlled by the control means. An audio waveform data reproducing apparatus, comprising: reading means for reading the audio waveform data; and synthesizing means for synthesizing a reproduced waveform signal based on the audio waveform data read by the reading means. 2. A storage means for storing a plurality of periods of audio waveform data, a pitch information input means for inputting pitch information representing a reproduced pitch, and a compression / expansion information representing a time axis compression / expansion amount. Companding information input means; and a time which changes at a speed corresponding to the companding information input by the companding information input means and generates a time function representing an address position on audio waveform data stored in the storage means. Function generating means, pitch information generating means for generating pitch information representing the pitch of the periodic waveform data based on the address position indicated by the time function generated by the time function generating means, and input by the pitch information input means. A read address that changes at a speed corresponding to pitch information and generates a plurality of read addresses separated from each other in accordance with pitch information generated by the pitch information generating means; And at least one of the plurality of read addresses generated by the read address generating means changes corresponding to pitch information generated by the pitch information generating means based on a time function generated by the time function generating means. In the case where the read address changes beyond the range, the read address keeps a state of pitch separation represented by pitch information generated by the pitch information generating means, and the change falls within the change range.
Control means for controlling at least one of the plurality of read addresses to jump; read means for reading audio waveform data from the storage means in accordance with the read addresses controlled by the control means; read by the read means Audio waveform data reproducing apparatus comprising: synthesizing means for synthesizing a reproduced waveform signal based on the obtained audio waveform data. 3. The audio waveform data reproducing apparatus according to claim 1, wherein said control means includes an address represented by a time function generated by said time function generating means and said read address generation. The read address is compared with a read address generated by the means, and when it is detected that the read address exceeds a change range corresponding to the pitch information generated by the pitch information generating means, the read address is controlled to jump. Audio waveform data reproducing apparatus. 4. The audio waveform data reproducing apparatus according to claim 1, wherein said control means includes: a changing speed of a time function generated by said time function generating means; A jump timing of the next read address is calculated based on a change speed of the read address generated by the generating means and a change range corresponding to the pitch information generated by the pitch information generating means. An audio waveform data reproducing apparatus for controlling a jump of a read address generated by the read address generating means when the read address is generated. 5. The audio waveform data reproducing device according to claim 1, wherein the synthesizing unit includes an audio waveform read by the reading unit. An audio waveform data reproducing apparatus, comprising: a window function assigning means for assigning a window function to data, wherein the control means prevents occurrence of noise generated when controlling a read address. 6. A storage means for storing a plurality of periods of audio waveform data, a compression / expansion information input means for inputting compression / expansion information indicating a time axis compression / expansion amount, and a compression / decompression information input means. A time function generating means that changes at a speed corresponding to the expansion information and generates a time function indicating an address position on the audio waveform data stored in the storage means; and an address indicated by the time function generated by the time function generating means. Pitch information generating means for generating pitch information representing the pitch of the periodic waveform data based on the position; read address generating means for generating a read address changing at a predetermined speed; and read address generated by the read address generating means. ,
When the read address changes within the change range when the change exceeds the change range corresponding to the pitch information generated by the pitch information generation means based on the time function generated by the time function generation means. Control means for controlling the read address to jump by a value corresponding to the pitch information generated by the pitch information generating means; and audio waveform data from the storage means in accordance with the read address controlled by the control means. And a synthesizing means for synthesizing a reproduced waveform signal based on the audio waveform data read by the reading means. 7. A storage means for storing a plurality of periods of audio waveform data, a compression / expansion information input means for inputting compression / expansion information indicating a time axis compression / expansion amount, A time function generating means that changes at a speed corresponding to the expansion information and generates a time function indicating an address position on the audio waveform data stored in the storage means; and an address indicated by the time function generated by the time function generating means. Pitch information generating means for generating pitch information representing the pitch of the periodic waveform data based on the position; and a plurality of read addresses separated by pitches which change at a predetermined speed and are represented by pitch information generated by the pitch information generating means. And at least one of the plurality of read addresses generated by the read address generating means, When the read address changes over a change range corresponding to the pitch information generated by the pitch information generating means based on the time function generated by the generating means, the read address of the pitch information generated by the pitch information generating means is changed. While maintaining the pitch-separated state to represent, and so that the change is within the change range,
Control means for controlling at least one of the plurality of read addresses to jump; read means for reading audio waveform data from the storage means in accordance with the read addresses controlled by the control means; read by the read means Synthesizing means for synthesizing a reproduced waveform signal based on the obtained audio waveform data. 8. Pitch conversion information input means for inputting pitch conversion information representing an amount of pitch to be changed, storage means for inputting audio waveform data and storing the audio waveform data according to an address changing at a predetermined speed; A time function generating unit that changes at the predetermined speed and generates a time function indicating an address position on the audio waveform data stored in the storage unit; and a pitch detection unit that detects pitch information of the input audio waveform data. Read address generating means for generating a read address that changes at a speed corresponding to the pitch conversion information; and when the read address changes beyond a change range corresponding to the pitch information, the change in the read address is The read address is jumped by a value corresponding to the pitch information so as to be within the change range. Pitch conversion processing device of the audio waveform data and a control means for performing controlled so. 9. Pitch conversion information input means for inputting pitch conversion information indicating an amount of pitch to be changed, storage means for inputting audio waveform data and storing the audio waveform data according to an address changing at a predetermined speed; A time function generating unit that changes at the predetermined speed and generates a time function representing an address position on the audio waveform data stored in the storage unit; a pitch detection unit that detects pitch information of the input audio waveform data; Read address generating means for generating a plurality of read addresses separated by a pitch represented by the pitch information and changing at a predetermined speed, at least one of the read addresses changes beyond a change range corresponding to the pitch information In the case, the read address keeps the state separated by the pitch, and Pitch conversion processing device of the audio waveform data and a control means for controlling to jump at least one of said plurality of read address to be within the variation range.