JP3414119B2 - Music synthesizer - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a musical tone synthesizing device used for a sound source of an electronic musical instrument, and more particularly to a musical tone synthesizing device based on a system (hereinafter referred to as a physical model system) simulating a sounding mechanism of a natural musical instrument with an electronic circuit. Regarding

[0002]

2. Description of the Related Art Various musical tone synthesizers based on a physical model system have been proposed (see, for example, Japanese Patent Application No. 6-683).

A conventional tone synthesizer as described above will be described below with reference to the drawings.

FIG. 5 is a block diagram of a conventional tone synthesizer. In FIG. 5, reference numeral 501 denotes a drive data generation circuit for reading drive data stored in advance in a memory or the like, 502,
508 is a delay device that delays data for a predetermined time, 503 is a low-pass filter, 504 and 509 are multipliers, 505 is an adder, and 506 and 511 are delay devices 502 and 50, respectively.
8 is a delay control circuit for controlling the delay time of 510, 510 is a subtractor, 507 is a delay device 502, a low-pass filter 503,
A recursive comb filter including a multiplier 504, an adder 505, and a delay control circuit 506 is a non-recursive comb filter including a delay unit 508, a multiplier 509, a subtractor 510, and a delay control circuit 511. Since the drive data generation circuit 501 can be easily realized by using a read-only memory in which drive data is stored in advance and a well-known circuit such as a counter that increments the memory address by 1 according to the sampling clock SCK, its internal configuration is simple. The description is omitted. Further, the delay units 502 and 508 are composed of a read / write memory in a ring memory format, and read / write the address ADR output from the delay unit control circuits 506 and 511.

FIG. 6 is a circuit diagram of the delay control circuits 506 and 511. In FIG. 6, 601 is a counter 603.
Of the delay stage 502 and the delay stage number M of the delay device 508 or the delay stage number N of the delay device 502, and 602, when the comparison match flag output from the comparator 601 is 1 or the sound generation start flag KON is When it is 1, an OR gate resets the count value of the counter 603 to 0, and 603 is a counter which increments the value by 1 according to the sampling clock SCK. The pronunciation start flag KO
N is a flag that has a value of 1 only during the first one sampling period Ts when the synthesis process is started in response to a sounding operation such as a key depression.

FIG. 7A is a schematic view showing a vibration mechanism of a string of a guitar. In FIG. 7A, the string is A,
It is assumed that the force is fixed at the point C and the force (corresponding to the driving data) given when the string is played with a finger or a pick is given to the point B. FIG. 7B is a graph showing a time axis waveform of a guitar sound. In FIG. 7B, the horizontal axis and the vertical axis are axes representing time and amplitude, respectively. FIG. 7C is a graph showing the amplitude frequency spectrum of the guitar sound. In FIG. 7C, the horizontal axis and the vertical axis are axes representing frequency and amplitude, respectively. FIG. 7D is a graph showing the amplitude frequency spectrum of the drive data of the guitar. In FIG. 7 (D),
The horizontal axis and the vertical axis are axes representing frequency and amplitude, respectively.

FIG. 8 is a transfer characteristic diagram of the non-recursive comb filter 512. In FIG. 5A, the number N of delay stages of the delay unit 508 in FIG. 5 is 1/4 of the number M of delay stages of the delay unit 502, and the dip level data D given to the multiplier 509 has a value of 0.5. 5B, the number N of delay stages of the delay device 508 in FIG.
02 is 1/2 of the number M of delay stages, and the multiplier 509
5 is based on the transfer function when the dip level data D given to is 0.5, and FIG.
The number of delay stages N of 8 is 1/4 of the number M of delay stages of the delay unit 502, and the dip level data D given to the multiplier 509 is based on the transfer function when the value is 0.0.
The horizontal axis and the vertical axis are axes that represent frequency and amplitude, respectively.

FIG. 9A is a graph showing the amplitude frequency spectrum of the string ensemble sound. FIG. 9 (B)
6 is a graph showing an amplitude frequency spectrum of drive data of a string ensemble. FIG. 9C is a graph showing the amplitude frequency spectrum of the synthetic sound data of the string ensemble. Note that in FIGS. 9A to 9C,
The horizontal axis and the vertical axis are axes representing frequency and amplitude, respectively.

The operation of the conventional tone synthesizing apparatus configured as described above will be described. First, the vibration mechanism of the strings of the guitar will be described with reference to FIG. A, C
In the string fixed by, the length ratio between AB and BC is 1: 3. A single force (corresponding to drive data) is applied to the string by playing point B of the string, and propagates in the direction of the broken line arrow. Then, it is reflected and inverted at points A and C, and then returns to point B again. By repeating this operation, vibration of the strings occurs and sound is generated through the body. For simplicity, the body is omitted and the guitar sound is regarded as the vibration of the strings. In addition, since the high frequency component is mainly radiated from the strings into the air with the passage of time and the vibration energy is gradually consumed at the fixed points A and C, the time-axis waveform of the guitar sound is shown in FIG.
As shown in (B), the shape gradually becomes gradually longer with time. The time interval TM between W1 and W3 is
The waveform propagated from point B corresponds to the time required to reflect C → A or A → C and return to point B, and the time interval TN between W1 and W2 is only the waveform propagated from point B is A. Corresponds to the time required to reflect and return to point B again.

An equivalent circuit simulating the above-mentioned sound generation mechanism is the musical sound synthesizer shown in FIG. In FIG. 5, the drive data generation circuit 501 supplies drive data to the cyclic comb filter 507 based on the tone generation start flag KON and the sampling clock SCK. The supplied drive data circulates in the recursive comb filter 507, and
A periodic waveform having a time interval TM such as W1, W3, W5, ... Of (A) is generated. The high frequency band of the periodic waveform is blocked by the low-pass filter 503 during the circulation, and is attenuated by multiplication of the gain G (G is less than 1) in the multiplier 504. The time interval TM is the number of delay stages M of the delay device 502.
Corresponding to. The relationship between the time interval TM and the number of delay stages M is shown in (Equation 1).

[0011]

[Equation 1]

The periodic waveform generated by the recursive comb filter 507 is sent to the non-recursive comb filter 512 and filtered, whereby synthesized sound data corresponding to the time axis waveform of the guitar sound shown in FIG. 7B. Is obtained. The synthesized sound data or the amplitude frequency spectrum of the guitar sound has a shape as shown in FIG. 7 (C). That is, the amplitude frequency spectrum of the sound when a quarter of the length of the string is played is the fourth and eighth harmonics (f4, f8).
There is a dip in ... Generally, when a sound is generated by driving 1 / n of the total length of a resonance part such as a string, the amplitude frequency spectrum of the sound is n, 2
n, 3n, ... Dips occur in the next overtone. Figure 7
Time interval TN of W1 and W2 in the time axis waveform of (B)
Corresponds to the number N of delay stages of the delay device 508. The relationship between TN and N is shown in (Equation 2).

[0013]

[Equation 2]

The dip level data D multiplied by the multiplier 509 is the data of W2 with respect to W1 of FIG. 7B.
Alternatively, it corresponds to the level ratio of W4 to W3. Here, the driving data is, for example, a circuit corresponding to the original sound of the guitar to the strings of the guitar, in this case, the cyclic comb filter 50.
7 and the non-recursive comb filter 512 are data obtained by conversely convoluting the circuit connected in series, and correspond to the force given by the finger or the pick to the strings.

Next, the operation of the delay control circuits 506 and 507 will be described with reference to FIG. Delay control circuits 506 and 5
Since both 07 operate basically the same, only the delay control circuit 506 will be described. In FIG. 6, the sounding start flag KON is set to the OR gate 602 in response to sounding operation such as key depression.
Is input to the counter 603 and the count value of the counter 603 is reset to the value 0. After that, the counter 603 increments the count value by 1 according to the generation timing of the sampling clock SCK. Then, when it becomes equal to M input to the comparator 601, the comparator 601 sends a comparison match flag to the OR gate 602 and resets the count value of the counter 603 to the value 0. Through the above operation, the read / write address ADR is sent from the counter 603 to the delay device 502 (read / write memory having a ring memory configuration). Note that the address ADR first acts as a read address within one sampling period Ts, and immediately after that acts as a write address.
2 functions as a delay device with delay time TM (delay stage number M).

Next, the transfer characteristic of the non-recursive comb filter 512 will be described. As described above, when a sound is generated by driving 1 / n of the entire length of the resonance part such as a string, the amplitude frequency spectrum of the sound is n, 2n,
3n, ... A dip occurs in the next harmonic. Similarly to this phenomenon, the delay stage number N of the delay device 508 in the non-recursive comb filter 512 is set to the delay device 50 in the recursive comb filter 507.
When the delay stage number M is set to 1 / n, the amplitude frequency spectrum of the synthesized sound data has dips in n, 2n, 3n, ... Next overtones. However, the non-recursive comb filter 512
The depth of the dip varies depending on the value of the dip level data D in the middle multiplier 509. For example, N = M
In the case of / 4, D = 0.5, as shown in FIG.
A dip of -6 dB occurs at 4, f8, ..., N = M
In the case of / 2, D = 0.5, as shown in FIG.
A dip of -6 dB occurs at 2, f4, .... Further, when N = M / 4 and D = 0.0, no dip occurs as shown in FIG. However, when D = 0.0, the dip does not occur regardless of the value of N.

Now, as shown in FIG. 7C, f4 and f8
When the drive data of the sound (guitar sound) when there is a dip of about 6 dB is obtained, the transfer characteristic of the non-recursive comb filter 512 is set in advance to the characteristic shown in FIG. 8A (N = M / 4, D = 0.5), by convolving the original sound of the guitar into the reverse transfer characteristic of this characteristic, the component levels of f4, f8, ... In FIG. 7C are almost the same as those of other components. Periodic waveform (hereinafter referred to as periodic waveform W)
Is obtained. This is the characteristic of the periodic waveform W to be output from the cyclic comb filter 507. Further, the number of delay stages M of the delay unit 502 in the cyclic comb filter 507 is set to the pitch of the periodic waveform W, and the characteristics of the low-pass filter 503 are adjusted so as to be the closest to the curve of the amplitude frequency spectrum envelope of the periodic waveform W. Performed, recursive comb filter 5
The reverse transfer characteristic of 07 is obtained. Then, by convolving the periodic waveform W into this inverse transfer characteristic, each component of the periodic waveform W is reduced to almost the same level, and the amplitude frequency spectrum of the obtained drive data is white as shown in FIG. 7 (D). Like noise, it has a characteristic with components in all bands. As the time axis waveform, a sporadic waveform as shown in FIG. 7 (A) is obtained.

The synthesis of guitar sounds has been described above.
Finally, I will supplement the musical tones of compound tones such as string ensembles. A polyphonic musical tone such as a string ensemble is a musical tone that is made by overlapping a plurality of stringed instrument sounds such as a violin, and each musical instrument is equivalent to 1/4 of the total length of the string like the guitar as described above. Even if the positions to be driven are driven, they overlap each other while interfering with each other, so that the amplitude frequency spectrum of the string-unstable sound does not have a clear dip in a specific component (see FIG. 9).
(See (A)).

[0019]

However, in the configuration as shown in FIG. 5, when the tone color is changed by controlling the number of delay stages N and the dip level data D, FIG.
As shown in (A), (B), and (C), since the level of the first-order component of the transfer characteristic of the non-recursive comb filter 512 is greatly different, the sense of volume of the synthetic sound data also changes. Most musical tones such as guitar sounds are shown in Figure 7.
This is because the level of the primary component is the highest as shown in (C).

Further, in the case of a musical tone of a polyphonic system such as a string ensemble, a clear dip does not occur in the amplitude frequency spectrum as shown in FIG. 9 (A), so that the driving position of the string is ¼ of the total length, that is, When the drive data is calculated with the delay stage number N of the delay device 508 set to M / 4, FIG.
A peak appears at f4, f8 ,. When synthesized using this drive data, when N = M / 4, the non-recursive comb filter 512 reduces the level of this peak component to a level equivalent to that of the original sound. Although the desired musical sound as shown is obtained, for example, when N = M / 6, the peaks of f4, f8, ... Are left and the levels of f6, f12 ,. (6dB)
Down), as shown in FIG. 9C, the synthesized sound data has a specific component, that is, a component (N = M) corresponding to the value of N when the inverse transfer function of the non-recursive comb filter 512 is convoluted.
At / 4, peaks occur at f4, f8, ...), which is not preferable in terms of hearing.

In order to prevent a peak from occurring in a specific component of the drive data, the reverse transfer characteristic of the non-recursive comb filter 512 is not convolved in the process of obtaining the drive data, and the reverse transfer of the recursive comb filter 507 is not performed. A method of only convolving the characteristics is conceivable. In that case, in order to reproduce the original sound, that is, the string unsable sound in which the reverse transfer characteristic of the recursive comb filter 507 should be convoluted, synthesis is performed using only the recursive comb filter 507. There is a need to. In other words, the non-recursive comb filter 512
The original sound cannot be reproduced if filtering is also performed. Therefore, the non-recursive comb filter 512 cannot be used for reproducing the original sound, and in this case, the tone color change corresponding to the change in the driving position of the string is realized as described in the guitar sound example. Can not do it.

[0022]

In order to solve the above problems, the present invention corrects the volume level of synthesized sound data according to the number of delay stages of a delay device in a non-recursive comb filter and dip level data. A conversion means and an amplitude control means for performing the above are provided. Then, the conversion means generates the volume level correction data generated according to the number of delay stages and the dip level data, and the amplitude control means performs the volume level correction of the synthetic sound data using the volume level correction data, thereby obtaining the number of delay stages. It is possible to keep the volume of the synthetic sound data constant when the tone color is controlled by controlling the dip level data.

Further, the present invention provides a circuit including a drive data generating circuit and a recursive comb filter, in addition to adding a converting means and an amplitude controlling means to the conventional musical tone synthesizing device (the above-mentioned solving means). , PCM data generation circuit. Then, with respect to the PCM data generated by the PCM data generation circuit and corresponding to the original sound of a natural musical instrument or the like, a timbre change corresponding to the number of delay stages of the delay device in the non-recursive comb filter, that is, a resonance part such as a string In the case of realizing the tone color change according to the drive position control,
Since the volume of the synthesized sound data can always be kept constant, there is an advantage that the sound source circuit based on the conventional PCM method can be utilized as the waveform generating means before filtering, in addition to the effect of the above-mentioned solving means.

Further, according to the present invention, the conversion means is provided for controlling the dip level of the transfer characteristic of the non-recursive comb filter according to the number N of delay stages of the delay device in the non-recursive comb filter. Then, when the conversion means reproduces the original sound of the string unsable (if the delay stage number N in this case is M / 4, where M is the delay stage number of the delay device in the cyclic comb filter), the dip level data is obtained. When synthesizing with 0 as the value and obtaining another timbre in which the drive position of the resonance part such as a string is controlled (N is other than M / 4)
Performs the combining process with the dip level data as a value other than 0. Furthermore, by controlling the dip level data to gradually increase from a value of 0 as the value of N deviates from M / 4, the original sound of the string unsable can also be reproduced, and the original sound of the string unsable can be centered. It is possible to continuously change the timbre according to the change in the driving position of the resonance part such as the strings.

Further, in the present invention, a drive data generating circuit and a recursive comb filter are combined in addition to the conversion means for controlling the dip level data added to the conventional musical sound synthesizer (the above-mentioned solving means). The circuit is to be replaced with a PCM data generation circuit. And PC
To the PCM data generated by the M data generation circuit and corresponding to the original sound of a natural musical instrument or the like, a timbre change corresponding to the delay stage number N of the delay unit in the non-recursive comb filter, that is, the driving position of the resonance part such as a string Since the timbre change according to the control can be continuously realized, there is an advantage that the tone generator circuit based on the conventional PCM method can be utilized as the waveform generating means before filtering, in addition to the effect of the above-mentioned solving means.

[0026]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention has drive data generating means for generating drive data and a closed loop circuit including at least a delay device, and makes the drive data supplied from the drive data generating means circulate. Data is extracted as waveform data, and the waveform data extracted from the cyclic comb filter is subtracted from the waveform data after delaying the waveform data for a predetermined time N and level-adjusted to obtain a subtraction result. A non-recursive comb filter that outputs, an amplitude control unit that outputs desired synthesized sound data by controlling the amplitude of the output of the non-recursive comb filter, and amplitude level data that is a control parameter in the amplitude control unit. Drive data, and a conversion unit that stores amplitude level data in advance according to the characteristics of the non-recursive comb filter. The raw means supplies drive data corresponding to the force applied to the resonance part (string etc.) of a natural musical instrument such as a guitar to the cyclic comb filter, and the cyclic comb filter cyclically drives the drive data to generate a periodic waveform, The non-recursive comb filter controls the delay time N of the delay device therein to give a dip to a specific component of the periodic waveform (a component corresponding to the delay time N), so that the driving position of the resonance part (string) is And the volume level correction data L stored in advance by the conversion means are adjusted on the characteristics of the non-recursive comb filter, such as the delay time N and the dip level data D (on the delay path of the non-recursive comb filter). (Corresponding to the level) corresponding to the volume level), and the amplitude control means performs the volume level correction of the synthesized sound data based on the read volume level correction data L. Accordingly, the keeping constant the volume of the synthesized speech data in the tone color change of controlling the number of delay stages N or dip level data D.

Further, the present invention is a PCM data generating means for generating PCM data corresponding to a musical tone of a natural musical instrument, and P
A non-recursive comb filter for subtracting the PCM data sent from the CM data generating means and the data after delaying the PCM data for a predetermined time N and adjusting the level, and outputting the subtraction result; and the non-recursive comb filter. Amplitude control means for outputting desired synthesized sound data by controlling the amplitude of the output, and amplitude level data which is a control parameter in the amplitude control means are stored in advance, and the amplitude level is adjusted according to the characteristics of the non-recursive comb filter. The non-recursive comb filter supplies the musical tone data corresponding to the musical instrument original sound generated by the PCM data generating means to the non-recursive comb filter, and the non-recursive comb filter determines the delay time N of the internal delay unit. By controlling and giving a dip to a specific component (component corresponding to the delay time N) of the musical sound data,
The tone color is changed according to the driving position of the resonance part (string), and the volume level correction data L stored in advance by the conversion means is
The delay is performed by selecting and reading according to the characteristics of the non-recursive comb filter, for example, the delay time N and the dip level data D, and the amplitude control means correcting the volume level of the synthetic sound data based on the volume level correction data L. The volume of the synthesized sound data is kept constant when the tone color is changed by controlling the number of stages N and the dip level data D. Further, the sound source circuit based on the conventional PCM system can be utilized as a waveform generating means before filtering.

Further, the present invention has a drive data generating means for generating drive data and a closed loop circuit including at least a delay device, and makes the drive data supplied from the drive data generating means circulate. Is obtained as waveform data, and the waveform data extracted from the cyclic comb filter is subtracted from the waveform data after delaying the waveform data for a predetermined time N and the level is adjusted to obtain a subtraction result. Of the non-recursive comb filter and the dip level data D corresponding to the level adjusted on the delay path in the non-recursive comb filter are stored in advance according to the characteristics of the non-recursive comb filter. And a conversion means for reading the dip level data D, and the drive data generation means is a compound tone system such as a string ensemble. The drive data is supplied to the cyclic comb filter, the cyclic comb filter cyclically drives the drive data to generate a periodic waveform, and the non-cyclic comb filter controls the delay time N of the delay device therein to control the cyclic waveform. By giving a dip to a specific component (a component corresponding to the delay time N), the tone color is changed according to the driving position of the resonance part (string), and the conversion means stores the dip level data D stored in advance in the delay time N. The level adjustment on the delay path of the non-recursive comb filter is performed based on the read and read dip level data D. At this time, when the original sound of the string unsavable is reproduced (if N in this case is set to 1/4 of the delay stage number M of the delay unit in the cyclic comb filter), the dip level data D is synthesized as a value 0. Further, when another tone color in which the driving position of the resonance part such as a string is controlled is obtained (N is other than M / 4), the dip level data D is synthesized as a value other than 0. Furthermore, N is M /
The dip level data D is stored in the converting means so as to gradually increase from the value 0 as it goes away from 4, and the converting means reads the dip level data D according to the number of delay stages N, and then on the delay path of the non-recursive comb filter. By adjusting the level of, the original sound of the string unsable is also reproduced, and the original sound of the string unsable is the center,
The tone color is continuously changed according to the change in the driving position of the resonance part such as the strings.

Further, the present invention is a PCM data generating means for generating PCM data corresponding to a musical tone of a natural musical instrument, and P
A non-recursive comb filter for subtracting the PCM data sent from the CM data generating means and the data after level-adjusting the PCM data by delaying it for a predetermined time N and outputting the subtraction result as desired synthesized sound data; PCM data generation is provided with conversion means for pre-storing the dip level data D corresponding to the level to be adjusted on the delay path in the non-recursive comb filter and reading the dip level data D according to the characteristics of the non-recursive comb filter. The musical tone data generated by the means and corresponding to the original sound of the musical instrument is supplied to the non-recursive comb filter, and the non-recursive comb filter controls the delay time N of the delay device inside the non-recursive comb filter to specify a specific component (delay time By giving a dip to the component corresponding to N, the tone color is changed according to the driving position of the resonance part (string), and further the conversion means Selected according dip level data D because stored in the delay time N, reading, performs level adjustment of the delay path of the non-recursive comb filter based on the dip level data D read. At this time, when the original sound of the string unsavable is reproduced (if N in this case is set to 1/4 of the delay stage number M of the delay unit in the cyclic comb filter), the dip level data D is synthesized as a value 0. Further, when another tone color in which the driving position of the resonance part such as a string is controlled is obtained (N is other than M / 4), the dip level data D is synthesized as a value other than 0. Furthermore, N is M /
The dip level data D is stored in the converting means so as to gradually increase from the value 0 as it goes away from 4, and the converting means reads the dip level data D according to the number of delay stages N, and then on the delay path of the non-recursive comb filter. By adjusting the level of, the original sound of the string unsable is also reproduced, and the original sound of the string unsable is the center,
The tone color is continuously changed according to the change in the driving position of the resonance part such as the strings. Further, the sound source circuit based on the conventional PCM system can be utilized as a waveform generating means before filtering.

(Embodiment 1) Hereinafter, Embodiment 1 of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of a musical sound synthesizing apparatus according to the first embodiment of the present invention. In FIG. 1, 101 is a drive data generation circuit for reading drive data stored in advance in a memory or the like, 102 and 108 are delay devices for delaying data for a predetermined time, 103 is a low-pass filter, and 104 and 1
09 and 114 are multipliers, 105 is an adder, and 106 and 11
1 is a delay control circuit for controlling the delay times of the delay devices 102 and 108, 110 is a subtracter, and 107 is a delay device 10.
2, low-pass filter 103, multiplier 104, adder 1
05, a cyclic comb filter including the delay control circuit 106, 112 is a delay unit 108, a multiplier 109, and a subtractor 11
0, a non-recursive comb filter that includes the delay control circuit 111, 113 is a sound volume adjustment table for reading the pre-stored sound volume level correction data L based on the number of delay stages N and dip level data D, and 114 is a non-recursive comb filter It is a multiplier that multiplies the waveform sent from 112 by the volume level correction data L read from the volume adjustment table.

Since the drive data generation circuit 101 can be easily realized by using a read-only memory in which drive data is stored in advance and a well-known circuit such as a counter which increments the memory address by 1 in response to the sampling clock SCK, The description of the internal configuration is omitted. Further, the delay devices 102 and 108 are composed of a read / write memory in a ring memory format, and read / write the address ADR output from the delay device control circuits 106 and 111.

The operation of the musical tone synthesizer configured as above will be described. First, in FIG. 1, the drive data generation circuit 101 supplies drive data corresponding to the force applied to the resonance portion (string or the like) of a natural musical instrument such as a guitar to the cyclic comb filter 107 based on the sound generation start flag KON and the sampling clock SCK. Supply. The supplied drive data is circulated through the cyclic comb filter 107,
A periodic waveform having a time interval TM such as W1, W3, W5, ... In FIG. 7A is generated. The high frequency band of the periodic waveform is cut off by the low-pass filter 103 during the circulation, and is attenuated by the multiplication of the gain G (G is less than 1) in the multiplier 104. Then, by the filtering process in the non-recursive comb filter 112, W1 in FIG.
Waveforms such as W2, W3, W4, W5, ... Are output from the non-recursive comb filter 112. The transfer characteristics of the non-recursive comb filter 112 are shown in FIGS. 8 (A), (B), and (C).
As shown in, the difference depends on the number of delay stages N and the dip level data D (as a result, the timbre of the synthetic sound data is different). Here, since most musical tones such as guitars have a primary component having a higher level than the other overtone components, the volume of the desired synthesized sound data is almost determined by the level of the primary component. Therefore, the number of delay stages N and the dip level data D are set in advance.
And clarify the relationship of the amplitude characteristics corresponding to the first-order component,
The relationship may be stored in the volume adjustment table 113, read based on the number of delay stages N and the dip level data D, and the volume level of the synthetic sound data may be adjusted. Number of delay stages N
In order to clarify the relationship between the dip level data D and the amplitude characteristic corresponding to the first-order component, first, the transfer function H (w) of the non-recursive comb filter 112 is obtained as shown in (Equation 3).

[0034]

[Equation 3]

Further, the amplitude characteristic of the non-recursive comb filter 112 is obtained by taking the absolute value of H (w). This is shown in (Equation 4).

[0036]

[Equation 4]

Further, based on (Equation 4), the value of the volume level correction data L is calculated so that the primary component of the synthesized sound data becomes constant. This is a table in which the number of delay stages N and the dip level data D are used as an address (Table 1).
Is.

[0038]

[Table 1]

In order to realize (Table 1) as the volume adjustment table 113, first, the dip level data D is divided into 64 steps between values 0 and 1, and the value of one dip level data D is delayed. The value of the number of stages N is 64 steps between M / 32 and M / 2 (32 to 2 are set at 64 at arbitrary intervals).
Divide into stages). Then, the value of the volume level correction data L calculated based on (Equation 4) is stored in the read-only memory in each division (address), and the dip level data D is stored in the upper address (6 bits) of the read-only memory. The volume adjustment table 113 is realized by causing the delay stage number N to be addressed as the lower address (6 bits) of the read-only memory. The number of delay stages N and the dip level data D are 64 stages, respectively, but with such resolution, sufficient continuity can be obtained in terms of hearing.

Although the value to be stored in the volume adjustment table is determined with reference to the level of the first-order component, there are some major tones such as a fourth-order harmonic component depending on the musical tone (for example, trumpet). In this case, the value to be stored in the volume adjustment table may be determined based on the level of the fourth-order component. Then, the volume adjustment table 113 may be selected according to the type of instrument sound. In addition, it is more accurate to combine the values stored in the volume adjustment table 113 by determining the values to be stored in the volume adjustment table 113 with reference to a specific component such as the first order or the fourth order and adjusting the values stored in the volume adjustment table 113. The volume level of sound data can be made uniform. Note that the connection order of the recursive comb filter 107 and the non-recursive comb filter 112 may be changed. Further, the connection order of the delay unit 108 and the multiplier 109 of the non-recursive comb filter 112 may be changed.

As described above, according to the present embodiment, the volume level correction data L as shown in (Table 1) is stored in the volume adjustment table 113, and the non-recursive comb filter 11 is stored.
The volume level correction data L in the volume adjustment table 113 is selected by using the characteristic of No. 2, that is, the delay stage number N and the dip level data D as an address, and the volume is adjusted with respect to the waveform sent from the non-recursive comb filter 112. Volume level correction data L read from the table 113
By multiplying by, it is possible to keep the volume level of the synthetic sound data constant regardless of the change of the tone color due to the characteristic change of the non-recursive comb filter 112.

(Second Embodiment) A second embodiment of the present invention will be described below with reference to the drawings.

FIG. 2 is a block diagram of a musical sound synthesizing apparatus according to the second embodiment of the present invention. In FIG. 2, 201 is a P that corresponds to the original sound of a natural musical instrument stored in advance in a memory or the like.
A PCM data generation circuit for reading CM data, 202 a delay device for delaying data for a predetermined time, 203 a multiplier,
205 is a delay control circuit for controlling the delay time of the delay unit 202, 204 is a subtractor, 206 is a delay unit 202, a multiplier 2
03, a subtractor 204, and a delay control circuit 205, a non-recursive comb filter, 207 is a volume adjustment table for reading out prestored volume level correction data L based on the number of delay stages N and dip level data D, and 208 is a non-volume adjustment table. It is a multiplier that multiplies the waveform sent from the cyclic comb filter 206 by the volume level correction data L read from the volume adjustment table 207.

The operation of the musical tone synthesizer configured as above will be described. This musical sound synthesizer is the drive data generation circuit 101 of the musical sound synthesizer in the first embodiment.
The cyclic comb filter 107 is replaced with a PCM data generation circuit 201. The pitch data input to the PCM data generation circuit 201 is inversely proportional to the number M of delay stages in the cyclic comb filter 107.
The PCM data generation circuit 201 reads the PCM data corresponding to the original sound of the natural musical instrument stored in advance with the address step width corresponding to the pitch data, and outputs the non-recursive comb filter 20.
Supply to 6. The subsequent operation is similar to that of the first embodiment. The delay device 202 of the non-recursive comb filter 206
It does not matter even if the connection order of the multiplier 203 and the multiplier 203 is changed.

As described above, according to this embodiment, the volume level correction data L as shown in (Table 1) is stored in the volume adjustment table 207, and the non-recursive comb filter 20 is stored.
6, the volume level correction data L in the volume adjustment table 208 is selected by using the delay stage number N and the dip level data D as addresses, and the volume is adjusted with respect to the waveform sent from the non-recursive comb filter 206. Volume level correction data L read from the table 207
By multiplying by, it is possible to keep the volume level of the synthesized sound data constant regardless of the change of the timbre based on the control of the characteristic of the non-recursive comb filter 206. In addition, PC
The PCM data generated by the M data generation circuit 201 is changed in tone color based on the control of the characteristic of the non-recursive comb filter 206. Therefore, the tone generator circuit based on the conventional PCM method generates a waveform before filtering. It can be used as a means.

(Third Embodiment) A third embodiment of the present invention will be described below with reference to the drawings.

FIG. 3 is a block diagram of a musical sound synthesizing apparatus according to the third embodiment of the present invention. In FIG. 3, 301 is a drive data generation circuit for reading drive data stored in advance in a memory or the like, 302 and 308 are delay devices for delaying data for a predetermined time, 303 is a low-pass filter, and 304 and 3
09 is a multiplier, 305 is an adder, 306 and 311 are delay control circuits that control the delay times of the delay units 302 and 308, 310 is a subtractor, 307 is a delay unit 302, a low-pass filter 303, a multiplier 304, and an adder. Circuit 305 and delay control circuit 306, which is a recursive comb filter, 31
Reference numeral 2 is a non-recursive comb filter that includes the delay device 308, the multiplier 309, the subtractor 310, and the delay control circuit 311.
Reference numeral 3 is a dip level adjustment table for reading the prestored dip level data D based on the number of delay stages N.

Since the drive data generation circuit 301 can be easily realized by using a read-only memory in which drive data is stored in advance and a well-known circuit such as a counter that advances the memory address by 1 in response to the sampling clock SCK, The description of the internal configuration is omitted. Further, the delay units 302 and 308 are constituted by a read / write memory in a ring memory format, and read / write the address ADR output from the delay unit control circuits 306 and 311.

The operation of the musical tone synthesizer configured as above will be described. This tone synthesis apparatus is the same as the tone synthesis apparatus according to the first embodiment except that the volume adjustment table 113 is changed to the dip level adjustment table 313 and the multiplier 114 is omitted. Therefore, only the difference will be described.

As shown in FIG. 9A, a musical tone of a polyphonic system such as a string ensemble does not have a clear dip in a specific component. Therefore, the driving data is obtained by previously circulating the original tone of the string ensemble or the like with a cyclic comb filter 307. It is obtained by convoluting the reverse transfer characteristic of ## EQU1 ## and stored in the read-only memory in the drive data generation circuit 301. When reproducing synthetic sound data corresponding to an original sound such as a string ensemble, at the time of synthesis, the same parameter value as the reverse transfer characteristic of the cyclic comb filter 307 (for example, the high cutoff frequency of the low-pass filter 303 and the gain in the multiplier 304). ) Is set, and further, the dip level data D of the non-recursive comb filter 312 is set to a value 0, and the synthesis processing is performed. At this time, since a musical tone such as a violin which normally constitutes a string ensemble sound drives the position of the string around 1/4 to 1/8 of the total length of the string, the number N of delay stages of the delay device 308 is set. For example, it is set to M / 4. Further, the value 0 is stored in the dip level adjustment table 313 at an address corresponding to the number of delay stages N of M / 4, and the dip level data D is gradually increased from the value 0 as N moves away from M / 4. To remember. The range of N is from M / 32 to M / 2 as shown in (Table 2).

[0051]

[Table 2]

The value of the dip level data D at the point where N is farthest from M / 4, that is, at N = M / 32 or N = M / 2, is close to a tone such as a violin because the value around 0.5 is close to a musical tone. , When the tone color is changed by controlling the delay stage number N,
A relatively natural timbre change feeling can be obtained. On the other hand, N is M
The closer the value of the point farthest from / 4 is to the value 1, the larger the range of the timbre change is perceptually large. Therefore, when it is desired to change the timbre extremely, the value 1 may be set. Since these values differ depending on the type of musical instrument, a plurality of dip level adjustment tables 313 may be provided for each musical instrument and used properly according to the type of musical instrument. Note that the connection order of the recursive comb filter 307 and the non-recursive comb filter 312 may be changed. Further, the connection order of the delay unit 308 and the multiplier 309 of the non-recursive comb filter 312 may be changed.

As described above, according to the present embodiment, the value of the number of delay stages N corresponding to the position at which the resonance part such as a string is driven in a musical instrument forming a complex tone such as a string ensemble. And the value 0 is stored in the memory address of the dip level adjustment table 313 corresponding to this delay stage number N, and the value of the dip level data D gradually increases from the value 0 as the distance from the address storing the value 0 increases. Dip level adjustment table 3
13, the dip level data D stored in the dip level adjustment table 313 according to the number of delay stages N is selected, read out, and the read dip level data D is stored in the non-recursive comb filter 312. By setting the multiplication value in the multiplier 309, the original sound of the string unsable can be reproduced, and the timbre can be continuously changed in accordance with the change of the driving position of the resonance part such as a string centering on the original sound of the string unsable. .

(Fourth Embodiment) A fourth embodiment of the present invention will be described below with reference to the drawings.

FIG. 4 is a block diagram of a musical sound synthesizing apparatus according to the fourth embodiment of the present invention. In FIG. 4, 401 is a P that corresponds to the original sound of a natural musical instrument stored in advance in a memory or the like.
A PCM data generation circuit for reading CM data, 402 a delay device for delaying data for a predetermined time, 403 a multiplier,
405 is a delay control circuit that controls the delay time of the delay unit 402, 404 is a subtractor, 406 is a delay unit 402, and a multiplier 4
03, the subtractor 404, and the delay control circuit 405, a non-recursive comb filter, and 407 is a dip level adjustment table for reading the prestored dip level data D based on the number of delay stages N.

The operation of the musical tone synthesizing apparatus configured as described above will be described. This tone synthesizer is the drive data generation circuit 301 of the tone synthesizer in the third embodiment.
And a circuit that combines the recursive comb filter 307
It is replaced with the data generation circuit 401. PCM
The data generation circuit 401 reads the PCM data corresponding to the original sound of the natural musical instrument stored in advance with the address step width corresponding to the pitch data and supplies it to the non-recursive comb filter 406. The subsequent operation is similar to that of the third embodiment. The delay unit 402 and the multiplier 4 of the cyclic comb filter 406
It does not matter if the connection order of 03 is changed.

As described above, according to the present embodiment, the value of the number of delay stages N corresponding to the position at which the resonance part such as a string is driven in a musical instrument that constitutes a complex tone such as a string ensemble. The value 0 is stored in the memory address of the dip level adjustment table 407 corresponding to the delay stage number N, and the value of the dip level data D gradually increases from the value 0 as the distance from the address in which the value 0 is stored is increased. Dip level adjustment table 4
07, and at the time of synthesis, the dip level data D stored in the dip level adjustment table 407 is selected according to the delay stage number N, and the read and read dip level data D is stored in the non-recursive comb filter 406. By setting the multiplication value in the multiplier 403, the original sound of the string unsable can be reproduced, and the timbre can be continuously changed in accordance with the change of the driving position of the resonance part such as a string centering on the original sound of the string unsable. .

Further, for the PCM data generated by the PCM data generation circuit 401, the non-recursive comb filter 4 is used.
Since the tone color control is performed based on the characteristic change of 06, the sound source circuit based on the conventional PCM method can be utilized as the waveform generating means before filtering.

[0059]

The present invention is carried out in the form as described above, and has the following effects.

The drive data generating means supplies drive data corresponding to the force applied to the resonance part (string or the like) of a natural musical instrument such as a guitar to the cyclic comb filter, and the cyclic comb filter cycles the drive data to cycle. By generating a waveform and applying a dip to a specific component of the periodic waveform while the non-recursive comb filter controls the delay time N of the delay device therein, a tone color according to the drive position of the resonance part (string) is obtained. The volume level correction data L that has been changed and further stored in advance by the conversion means is selected according to the characteristics of the non-recursive comb filter, for example, the delay time N and the dip level data D, and read out, and the amplitude control means corrects the volume level. By correcting the sound volume level of the synthetic sound data based on the data L, the synthetic sound data is changed in the timbre change in which the number of delay stages N and the dip level data D are controlled. It is possible to maintain the volume of data to be constant.

Further, the PCM data generating means generates
By supplying musical tone data corresponding to the original sound of the musical instrument to the non-recursive comb filter, and the non-recursive comb filter controls the delay time N of the delay device therein to give a dip to a specific component of the musical tone data, The tone color is changed according to the driving position of the resonance part (string), and the volume level correction data L stored in advance by the conversion means is changed according to the characteristics of the non-recursive comb filter, for example, the delay time N and the dip level data D. The number of delay stages N and the dip level data D are selected and read out, and the amplitude control means corrects the volume level of the synthesized sound data based on the volume level correction data L.
The tone volume of the synthesized sound data can be kept constant in the tone color change controlled by the above, and the sound source circuit based on the conventional PCM system can be utilized as the waveform generating means before filtering.

Further, the driving data generating means supplies driving data of a compound tone system such as a string ensemble to the cyclic comb filter, and the cyclic comb filter cyclically drives the driving data to generate a periodic waveform, and the non-cyclic comb filter. Controls the delay time N of the internal delay device to give a dip to a specific component of the periodic waveform, thereby changing the timbre according to the driving position of the resonance part (string), and further storing the conversion means in advance. The selected dip level data D is selected according to the delay time N, and the level adjustment after the delay of the delay device in the non-recursive comb filter is performed based on the read and read dip level data D. At this time, when the original sound of the string unsavable is reproduced (if N in this case is set to 1/4 of the delay stage number M of the delay unit in the cyclic comb filter), the dip level data D is synthesized as a value 0. Further, when another tone color in which the driving position of the resonance part such as a string is controlled is obtained (N is other than M / 4), the dip level data D is synthesized as a value other than 0. Further, as the N becomes further away from M / 4, the dip level data D is stored in the converting means so as to gradually increase from the value 0, and the converting means reads the dip level data D according to the delay stage number N, and the non-cyclic type. By adjusting the level on the delay path of the comb filter, the original sound of the string unsavable is also reproduced,
Further, the tone color can be continuously changed centering on the original sound of the string unsavable according to the change of the driving position of the resonance part such as a string.

Further, the PCM data generating means generates
By supplying musical tone data corresponding to the original sound of the musical instrument to the non-recursive comb filter, and the non-recursive comb filter controls the delay time N of the delay device therein to give a dip to a specific component of the musical tone data, The tone color is changed in accordance with the driving position of the resonance part (string), and the dip level data D stored in advance by the conversion means is selected according to the delay time N, read out, and non-cyclic based on the read dip level data D. Adjust the level after delay of the delay device in the comb filter. At this time, when the original sound of the string unsavable is reproduced (if N in this case is set to 1/4 of the delay stage number M of the delay unit in the cyclic comb filter), the dip level data D is synthesized as a value 0. , To obtain another timbre in which the drive position of the resonance part such as a string is controlled (N is M / 4
Other than) synthesizes the dip level data D as a value other than 0. Further, as the N becomes further away from M / 4, the dip level data D is stored in the converting means so as to gradually increase from the value 0, and the converting means reads the dip level data D according to the delay stage number N, and the non-cyclic type. By adjusting the level on the delay path of the comb filter,
The original sound of the string unsable is also reproduced, and the timbre can be continuously changed depending on the driving position change of the resonance part such as a string centering on the original sound of the string unsable.
Further, the sound source circuit based on the conventional PCM system can be utilized as a waveform generating means before filtering.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a musical sound synthesizer according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing the configuration of a musical sound synthesizer according to a second embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of a musical sound synthesizer according to a third embodiment of the present invention.

FIG. 4 is a block diagram showing the configuration of a musical sound synthesizer according to a fourth embodiment of the present invention.

FIG. 5 is a block diagram showing the configuration of a conventional musical sound synthesizer.

FIG. 6 is a circuit configuration diagram of a delay control circuit of the musical sound synthesizer.

7A is a schematic diagram showing a vibration mechanism of a guitar string, FIG. 7B is a graph showing a time axis waveform of a guitar sound, FIG. 7C is a graph showing an amplitude frequency spectrum of the guitar sound, and FIG. Graph showing the amplitude frequency spectrum of drive data

FIG. 8 is a transfer characteristic diagram of the non-recursive comb filter.

9A is a graph showing an amplitude frequency spectrum of a string ensemble sound, FIG. 9B is a graph showing an amplitude frequency spectrum of drive data of the string ensemble, and FIG. 9C is an amplitude frequency spectrum of synthesized sound data of the string ensemble. Graph

[Explanation of symbols]

101 Drive data generation circuit 102, 108 delay circuit 103 low-pass filter 104, 109, 114 multipliers 105 adder 106,111 Delay control circuit 107 Recursive comb filter 110 subtractor 112 Non-recursive comb filter 113 Volume adjustment table

─────────────────────────────────────────────────── ─── Continuation of front page (56) References JP-A-7-199936 (JP, A) JP-A-5-73062 (JP, A) JP-A-56-88198 (JP, A) (58) Field (Int.Cl. ⁷ , DB name) G10H 1/06-1/12 G10H ^7/ 00-7/12

Claims (6)

(57) [Claims] 1. A drive data generating means for generating drive data, and a closed loop circuit including at least a delay device, wherein the drive data supplied from the drive data generating means is circulated, and the data being circulated is waveform data. As a cyclic comb filter, and a non-cyclic circuit that subtracts the waveform data extracted from the cyclic comb filter from the waveform data after delaying the waveform data for a predetermined time N and adjusting the level. Type comb filter, amplitude control means for outputting desired synthesized sound data by controlling the amplitude of the output of the non-recursive comb filter, and amplitude level data which is a control parameter in the amplitude control means is stored in advance. Musical sound characterized by comprising conversion means for reading out amplitude level data according to the characteristics of the non-recursive comb filter Forming apparatus. Wherein conversion means, non-recursive comb according to the delay time N in a filter musical tone synthesizing apparatus according to claim 1, wherein the reading the amplitude level data. Wherein conversion means, non-recursive comb according to the level value to adjust the delay path in the filter musical tone synthesizing apparatus according to claim 1, wherein the reading the amplitude level data. 4. A PCM data generating means for generating PCM data corresponding to a musical tone of a natural musical instrument, PCM data sent from the PCM data generating means, and data after the PCM data is delayed by N for a predetermined time and its level is adjusted. In the amplitude control means, a non-recursive comb filter for performing subtraction with and outputting a subtraction result, an amplitude control means for outputting desired synthetic sound data by controlling the amplitude of the output of the non-recursive comb filter, A musical sound synthesizing apparatus comprising: a conversion unit that stores amplitude level data, which is a control parameter, in advance and reads the amplitude level data according to the characteristics of the non-recursive comb filter. 5. The tone synthesis apparatus according to claim 4 , wherein the converting means reads the amplitude level data according to the delay time N in the non-recursive comb filter. 6. The tone synthesis apparatus according to claim 4 , wherein the converting means reads the amplitude level data according to the level value adjusted on the delay path in the non-recursive comb filter.