JP2968455B2 - Method and apparatus for forming repetitive waveform of electronic musical instrument - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION The present invention relates to, for example, synthesizers,
The present invention relates to a method and apparatus for forming a repetitive waveform stored in an electronic musical instrument such as an electronic piano, an electronic organ, a single keyboard and used for repeated reading.

[0002]

2. Description of the Related Art Conventionally, in an electronic musical instrument, required musical tone waveform data is stored in a waveform memory in advance, and tone waveform data corresponding to panel settings and key operations are sequentially read out, and this waveform data is reproduced as a musical tone signal. A so-called read-out system that generates a musical tone by using such a method is known. Generally, a musical sound waveform is roughly divided into a rising portion showing a sharp waveform change at the beginning of musical tone generation and a continuous (envelope) portion showing a relatively smooth waveform change following the rising portion. The most prominent feature of a musical tone is a rising portion showing a sharp waveform change, and it is less in a subsequent envelope portion. Since the envelope portion of an actual musical tone lasts for a long time, an enormous memory capacity is required to store the entire tone as it is, and the memory price occupying the entire apparatus increases, which is impractical.

[0003] In consideration of such circumstances, mainly for the purpose of saving the memory capacity, a musical tone is converted into a predetermined section between a rising (attack) portion showing a sharp waveform change and a continuous (envelope) portion having a relatively small change. Is stored, and when playing a musical tone, the rising portion corresponding to the musical tone specified by the key operation is read out, and then the data of the predetermined section of the continuous portion is reduced to a required length. There is one in which a series of tone signals are read out repeatedly to generate a continuous series of tone signals. (For example, JP-A-2-146097)

In such a system, although the waveform of a portion to be repeatedly read (hereinafter referred to as a loop) has little change, there is a difference between the crest and the tail of the waveform. Are unnatural and unsuitable as a musical sound. Based on such circumstances, a cross-fade loop process for forming a natural loop waveform with respect to the original waveform as shown in FIG. FIG.
A loop top address LT corresponding to a position at which a loop is started in an original waveform of a musical tone as shown in (A) and a loop length LL which is a length for performing loop processing are designated. As a result, the loop end address LE becomes “LT + L
E "and the fade-in address FI are obtained as" LT-LL. "In such a state, first, the waveform in the section from the loop top address LT to the loop end address LE is faded out (as indicated by the broken line in the figure, In addition, the waveform of the section from the fade-in address FI to the loop top address LT is fade-in (as indicated by a broken line in the figure). (A process of gradually increasing the volume from 0 to the volume of the original signal).
As described above, there is known a cross-fade loop method in which a waveform obtained by adding a waveform that has been faded in and a waveform that has been faded out is used as a loop waveform. In the case of reproduction, as shown in FIG. 5B, the original waveform O indicated by a thick arrow extends from the beginning of the original waveform to the loop top address.
, And after this, after the loop top address,
The loop waveform obtained as described above is 1, 2, 3,
.. Are reproduced repeatedly a required number of times as shown in FIG.

[0005] However, in this method, if the waveforms of the cross-fade portions are not in phase, the tone connection becomes unnatural when repeatedly reading out the musical tone waveform data of the repetitive portions, and unnecessary higher harmonics are generated. Or
In addition, unnatural groans often occurred, and the quality of musical sounds was often significantly impaired.

Conventionally, in order to solve such a problem, the operator has to use a loop top address and a loop length (in addition, either a loop end address and a loop length or a loop top address and a loop end address may be used). Give some kind of points like this,
He created a cross-fade loop, listened to the reproduced sound of the obtained loop waveform, and searched for a combination that would produce as natural a loop as possible with less unnatural groans and higher harmonics. Therefore, it often takes time to find such a point, or the adjustment is delicate, and it is often difficult to completely match the phases.

Therefore, there has been a need for a method of forming a repetitive waveform that can easily generate a repetitive waveform and that can generate a high-quality repetitive waveform.

[0008]

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has been developed to quickly and efficiently produce high-quality repetitive waveforms.
It is another object of the present invention to provide a method for forming a repetitive waveform that can be easily created and a device for forming the waveform.

[0009]

According to the present invention, a loop top address (LT) and an initial value (LL) of a loop length are generated when a repetitive waveform used for an electronic musical instrument is created.
A loop length change width (WD) is given, and the loop length is changed from an initial value of the loop length to “(loop length initial value LL) + (loop length change width WD)”, and according to each loop length, Fade-in address “(loop top address LT) − (loop length L) of a waveform from which a repetitive waveform is created
L) ”to a waveform obtained by fading in the section from the loop top address, and from the loop top address to the loop end address“ (loop top address LT).
+ (Loop length LL) "is generated as a cross-faded repetitive waveform, and a waveform obtained by adding a waveform obtained by fading out a section of" + (loop length LL) "is obtained. The cross-faded repetitive waveform and the loop top address of the original waveform are generated. From the waveform in the section of the loop end address, a predetermined number of cross-fade repetitive waveforms are sequentially selected from the waveform having a high correlation coefficient, and the user actually selects the selected waveform from the selected waveforms. A method for forming a repetitive waveform in an electronic musical instrument, wherein a waveform to be used is determined.

Further, according to the present invention, the loop length can be changed from the loop length initial value to "(loop length initial value) + (loop length change width) in accordance with the designated loop top address, loop length initial value, and loop length change width. )) And an address calculator that calculates the loop top address, the loop end address “(loop top address) + (loop length)”, and the fade-in address “(loop top address) − (loop length)” at that time. And a waveform in which a section from the fade-in address to the loop top address of the waveform from which the repetitive waveform is to be created is faded in based on the output of the address calculator, and a section from the loop top address to the loop end address. Waveform faded out , A loop processor for generating a cross-fade repetitive waveform as a cross-fade repetitive waveform, and a waveform in a section from the loop top address to the loop end address of each repetitive waveform created by the loop processor and the original waveform A correlation coefficient calculator for calculating a correlation coefficient between the peak detector, a peak detector for detecting a point at which the correlation coefficient calculated by the correlation coefficient calculator reaches a maximum, A repetitive waveform forming apparatus for an electronic musical instrument, comprising: a selector for selecting a predetermined number of waveforms in descending order of number; and a display for displaying a selection result by the selector.

[0011]

According to the present invention, a loop top address LT, a loop length initial value LL and a length change width WD are given when a cross-faded repetitive waveform is created by using a general-purpose computer or a sampler of a synthesizer. The loop length of the waveform that is cross-fade within the range of the length change width can be automatically and automatically changed sequentially to create a repetitive waveform corresponding to each loop length.

Next, a correlation coefficient between each of the generated repetitive waveforms and the same portion of the original waveform as that of the original waveform is obtained, and a predetermined number of waveforms are sequentially selected from the repetitive waveforms having a high degree of correlation to determine the repetitive waveform. Things.

As a result, a cross-fade loop close to the original waveform is obtained, and the connection of musical tones becomes natural, and unnecessary high-order harmonics can be prevented from being generated, so that high-quality musical tones can be obtained.

Furthermore, according to the present invention, since a repetitive waveform is automatically created, the repetitive waveform can be created in a short time with certainty and ease of operation.

[0015]

FIG. 1 is a block diagram showing a schematic configuration of a repetitive waveform forming apparatus according to the present invention.

As shown in the figure, the repetitive waveform forming apparatus of the present invention comprises an address calculator 1, a loop processor 2, a correlation coefficient calculator 3, a peak detector 4, a selector 5, and a display 6. .

In such a configuration, when the operator inputs the loop top address LT, the initial value LL of the loop length and the loop length change width WD to be changed for simulation, these input values are read by the address calculator 1. Then, the address calculator 1 sequentially calculates each address value of the original waveform that is cross-fade as a repetitive waveform based on the respective input values, and sends it to the loop processor 2.

The length change width WD defines a range of the loop length LL to be simulated, and within such a range of the length change width, the computer automatically changes the loop length to change each loop length. It is configured to calculate an address and send the obtained calculation result to the loop processor 2.

On the other hand, since the original waveform is input to the loop processor 2, the loop processor 2 performs cross-fading waveforms corresponding to the respective loop lengths based on the data sent from the address calculator 1. And sends the result to the correlation coefficient calculator 3.

The correlation coefficient calculator 3 includes the loop processor 2
The correlation coefficient between each repetitive waveform created by the above and the same part of the original waveform is calculated.

The correlation coefficient calculated by the correlation coefficient calculator 3 is expressed by the following equation, where {Xn} is the data sequence of the original waveform and {Yn} is the data sequence of the cross-fade loop.

[0022]

Here, m: loop top address (= LT) L: loop length (= LL) σ _X : standard deviation of X _{m to} X _{m + L -1} σ _Y : Y _{m to} Y _{m + L -1} standard deviation

The peak detector 4 detects the peak value (maximum point) by comparing the magnitude of the correlation coefficient for each waveform sent from the correlation coefficient calculator 3.
Is sent to the selector 5.

The selector 5 selects a predetermined number of repetitive waveforms based on the result sent from the peak detector 4, and the selector 5 sequentially generates a predetermined number of repetitive waveforms in descending order of correlation. The selected result is displayed on the display 6
Will be displayed.

Therefore, the operator selects and determines a suitable repetitive waveform while watching the result displayed on the display 6.

The predetermined number defines the number of repetitive waveforms to be created, and is set in advance by the operator.

According to the present invention, according to the present invention, a computer or a sampler automatically creates a predetermined number of waveforms sequentially from a repetitive waveform having a high degree of correlation when an operator inputs a setting condition of the repetitive waveform.

Therefore, it is possible to quickly and easily create a repetitive waveform, and to provide a natural and high-quality repetitive waveform.

When the repetitive waveform of the present invention is created using, for example, a general-purpose computer, the address calculator 1, the loop processor 2, the correlation coefficient calculator 3, the peak detector 4, the selector 5, etc. This function can be realized by a program.

FIG. 2 is a main flowchart for explaining the operation of the repetitive waveform forming apparatus of the present invention. Hereinafter, the operation of this embodiment will be described with reference to the drawings.

First, parameters are read (step S11). This is the loop top address (LT) and the initial value of the loop length (LL) input by the operator.
And the process of taking the length change width (WD) into the address calculator 1.

Next, the original waveform is read into the loop processor 2 (step S12), and a loop length LE indicating the end condition of the simulation is set (step S13). The simulation end condition LE is obtained by adding the length change width WD to the loop length initial value LL.

Next, data for calculating the cross-fade waveform sent from the address calculator 1 is set (step S14). That is, the variables AD1 and AD2 include
The loop top address of each cross-fade waveform, that is, LT and LT-LL is set.

The loop processor 2 performs a loop process based on this (step S15). The loop process is a process of creating each cross-fade waveform based on the initial value set in step S14, and details will be described in detail with reference to FIG.

Subsequently, the correlation coefficient calculator 3 calculates a correlation coefficient between each created cross-fade waveform and the original waveform (step S16), and sends the calculated result to the peak detector 4. And stored. The calculation of the correlation coefficient will be described in detail in the description of the correlation coefficient calculation subroutine in FIG.

Next, the loop length LL is incremented for the next simulation (step S1).
7) It is checked whether or not to continue the simulation (step S18). This is a process for checking whether or not the loop length LL set in step S17 has reached the simulation end condition LE set in step S13.

In step S18, the loop length L
If L has not yet reached the end condition LE, the process returns to step S14, and the simulation is repeated with the next loop length incremented in step S17.

On the other hand, if the loop length LL has reached the termination condition LE in step S18, that is, if the simulation has been completed for all the loop lengths, the peak detector 4 determines the result of the simulation for all the waveforms. And a peak (maximum point) is detected (step S19).

Subsequently, the selector 5 selects a predetermined number of waveforms in descending order of correlation from the waveforms detected and stored by the peak detector 4 (step S 20), and sends them to the display 6. A predetermined number of waveforms are displayed on the display 6 (step S21).

Accordingly, the operator repeatedly selects and determines the waveform while viewing the result displayed on the display 6.

As described above, according to the present invention, using the given initial value as an initial condition, within the range of the length change width,
The correlation coefficient between the created cross-fade waveform and the original waveform is automatically simulated, and an optimal repetitive waveform is selected.

Therefore, it is possible to automatically create a highly accurate repetitive waveform.

Next, referring to FIG.
Will be described.

In the loop process, first, an initial value is set (step S31). In the initial value setting,
Variables PO1 and PO2 have head addresses AD1 and AD2 of two waveforms used for creating a cross-fade waveform.
2 is set to 0 and the counter CO is set to 0.

Subsequently, based on the amplitude values at the sample points of the two cross-fade waveforms, the amplitude values Y (P
O1) is obtained (step S32).

Next, the sample points PO1 and PO2 of the two cross-fade waveforms and the counter CO are incremented (step S33).

Further, it is checked whether or not the incremented sample point is within the loop length range of the waveform to be cross-fade (step S34).
If it is within the range, the process returns to step S32, and the calculation of the amplitude of the crossfade waveform at the next sample point is repeated.

On the other hand, if it is out of the range, the calculation of the cross-fade waveform related to the waveform is completed, and the process returns to step S16 of the main routine.

As described above, according to the present invention, a cross-faded waveform is formed while changing the loop length within the range of the length change width, and the correlation coefficient between the waveform and the original waveform is obtained. Is examined.

In step S32 of this embodiment, a calculation formula in the case of a linear crossfade is illustrated. In addition, as a method of crossfade, a logarithmic, exponential function,
There are methods such as Sin curve and Cos curve.

FIG. 4 is a diagram for explaining the operation of the subroutine for calculating the correlation coefficient in step S16 of the main flowchart. Hereinafter, the operation of calculating the correlation coefficient will be described with reference to the drawings.

In calculating the correlation coefficient, first, an initial value is set (step S41). In setting the initial value, 0 is set for the variable SUM for inputting the total value, and PO1 is set for the variable SUM.
, A loop top address (AD1) indicating the loop top of the original waveform and the cross-fade waveform is set.

Next, SUM = SUM + X (PO1) *
Y (PO1) is obtained (step S42). In addition,
X (PO1) is the amplitude value of the original waveform, and Y (PO1) is the amplitude value of the cross-fade waveform.

Next, the value of the sample point PO1 used for calculating the crossfade waveform is incremented (step S43).

Subsequently, it is checked whether or not PO1 <LT + LL (step S44). This is a process for checking whether or not the incremented sample point is within the loop length range of the waveform to be created.

If the value of the sample point PO1 has not yet reached the loop end address, step S42
And the same processing is repeated for the next data value.

On the other hand, if the value of PO1 exceeds the loop end value, the extraction of the sample points relating to the waveform has been completed, so the process proceeds to the next step to obtain the standard deviations σ _x and σ _y of X and Y. (Steps S45 and S46). The reason why the standard deviations σ _x and σ _y are obtained is to normalize waveforms having different loop lengths and make them comparable.

Next, using the standard deviations σ _x , σ _y and the above-described equation for calculating the correlation coefficient, A correlation coefficient is obtained based on the above (Step S47).

When the calculation of the correlation coefficient of the sample waveform is completed, the process returns to the main routine, and the calculation of the correlation coefficient of the next sample waveform is repeated.

[0061]

As described in detail above, according to the present invention,
The correlation coefficient of the same part of the cross-fade waveform and the original waveform is calculated in advance, and a waveform with a high degree of correlation is selected. At the same time, it is possible to prevent the generation of unnecessary higher harmonics and growls.

Further, since the repetitive waveform can be created automatically, the waveform can be created quickly and easily, and various loop lengths can be easily simulated.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of a waveform forming apparatus according to the present invention.

FIG. 2 is a main flowchart for explaining the operation of the waveform forming apparatus of the present invention.

FIG. 3 is a flowchart illustrating an operation of a loop process.

FIG. 4 is a flowchart illustrating an operation of calculating a correlation coefficient.

FIG. 5 is a diagram for explaining a crossfade loop creation method.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Address calculator 2 Loop processor 3 Correlation coefficient calculator 4 Peak detector 5 Selector 6 Display

Claims (2)

(57) [Claims] When a repetitive waveform used for an electronic musical instrument is created, a loop top address (LT) and an initial value (LL) of a loop length and a loop length change width (W) are set.
D), and the loop length is changed from the initial value of the loop length to “(initial value of loop length) + (loop length change width)”. According to each loop length, an original waveform is generated. A waveform in which the section from the fade-in address “(loop top address) − (loop length)” to the loop top address is faded in;
Then, from the loop top address to the loop end address “(loop top address) + (loop length)”
A waveform obtained by adding a waveform obtained by fading out the section of the section is generated as a cross-fade repetitive waveform, and the cross-faded repetitive waveform and the section of the section from the loop top address to the loop end address of the original waveform are generated. A correlation coefficient with a waveform is obtained, and a predetermined number of cross-fade repetitive waveforms are sequentially selected from a waveform having a high correlation coefficient, and a waveform actually used by a user is determined from the selected waveforms. A method for forming a repetitive waveform in an electronic musical instrument, characterized in that: 2. The loop length changes from the loop length initial value to “(loop length initial value) + (loop length change width)” according to the designated loop top address, loop length initial value, and loop length change width. An address calculator that calculates a loop top address, a loop end address “(loop top address) + (loop length)”, and a fade-in address “(loop top address) − (loop length)” at that time; Based on the output of the calculator, the waveform from which the section from the fade-in address to the loop top address of the waveform from which the repeated waveform is to be created is faded in, and the section from the loop top address to the loop end address is faded out. Add the waveform and A loop processor that generates the crossed waveform as a cross-fade repetitive waveform, and a phase relationship between each repetitive waveform created by the loop processor and a waveform in a section from the loop top address to the loop end address of the original waveform A correlation coefficient calculator that calculates the number, a peak detector that detects a point at which the correlation coefficient calculated by the correlation coefficient calculator reaches a maximum, and an ascending correlation coefficient from the output of the peak detector. A repetitive waveform forming device for an electronic musical instrument, comprising: a selector for selecting a predetermined number of waveforms; and a display for displaying a result of the selection by the selector.