JP6728843B2 - Electronic musical instrument, musical tone generating device, musical tone generating method and program - Google Patents

Description

The present invention relates to an electronic musical instrument, a musical tone generating device, a musical tone generating method and a program.

In an electronic musical instrument that implements a wind instrument by electronic technology, it absorbs the individual differences of the wind instrument player (performer) while giving a musical sound of the wind intensity of the wind instrument player and the strength of biting the mouthpiece of a traditional wind instrument (for example, a saxophone). There is known a conventional technique capable of performing a wind performance in accordance with its characteristic value as a parameter (for example, the technique described in Patent Document 1).

Further, in an electronic musical instrument, there is known a conventional technique of controlling the position and movement of the winder's tongue, a so-called tongue rendition style, and controlling the sound of the wind instrument being produced (for example, the technique described in Patent Document 2 or 3). ..

Japanese Patent No. 2605761 Japanese Patent No. 2712406 Japanese Patent No. 3389618

Here, for traditional wind instruments, there is a "growl method" as a special playing method that not only blows or tongs, but also makes a sound while actually uttering "Woooo" while giving a muddy sound. is there.

However, in the conventional electronic wind instrument, the normal tone, which is the tone of a normal wind instrument, and the above-described glowing tone are different tones, and it is necessary to use them separately by a tone color switching operation. Therefore, it was not possible to seamlessly change the timbre between the two timbres during performance/pronunciation.

Further, in the conventional acoustic wind instrument, skill is required such as the pitch relationship between the normal musical sound and the special sound by the growl playing method.
To solve this problem, it is possible to store the special performance sound by the growl method in the memory in advance and read it from the memory when needed according to the performance. There is a problem that only a single glowing sound can be produced.

Further, it may be possible to allow the wind performer to pronounce the growl sound while holding the mouthpiece, but in that case, a microphone is required for the mouthpiece to pick up the growl sound. There is.

INDUSTRIAL APPLICABILITY The present invention makes it possible to generate a more varied musical tone by generating and outputting a variety of human voice synthesized sounds based on the vocalization action of a wind performer without requiring a dedicated microphone or the like. With the goal.

In one example embodiment, to separate the output signal from the breath sensor for detecting the state of breath, to the human voice signal indicating the state of exhalation generated by blowing signal and human voice indicating the state of exhalation generated by blowing, the A musical instrument sound is output based on a wind signal, a designated pitch, and a designated tone color, and a pulse train signal having a pitch cycle corresponding to the designated pitch is associated with the designated tone color . it outputs human voice synthetic sound and filtered by the speech synthesis filter, and a processing unit.

According to the present invention, it is possible to generate a more varied musical tone by generating and outputting a variety of human voice synthesized sounds based on the vocalization action of the wind performer without requiring a dedicated microphone or the like. It will be possible.

It is sectional drawing which shows the structural example of the mouthpiece in embodiment. It is a figure showing an example of hardware constitutions of an embodiment of an electronic musical instrument. It is a figure which shows the block structural example of CPU. It is a figure which shows the block structural example of 2nd Wave Generator. It is a figure which shows the result of FFT analysis of the phoneme "u" of a human voice, and the example of the frequency characteristic of a formant filter. It is a figure which shows the result of FFT analysis of the phoneme "a" of a human voice, and the example of the frequency characteristic of a formant filter. 7 is a diagram showing a relationship between a time-domain waveform of a sound source pulse train uttered by a pulse generator 402 and a time-domain waveform of a human voice synthesis sound signal 310 generated by a formant filter 403. FIG. 9 is a table showing correspondences between mouthpieces, characteristics of pronunciation at the time of growl, and types of phonemes suitable for generation by the formant filter, for each musical instrument timbre. It is a figure which shows the example of a waveform of a sensor output signal and a human voice signal when it blows normally. It is a figure which shows the example of a waveform of a sensor output signal and a human voice signal at the time of consciously paying attention to a special performance style. It is a figure which shows the example of a waveform in this embodiment. It is a flow chart which shows an example of pronunciation control processing which a CPU performs. It is explanatory drawing of the reflected sound of a wind instrument. It is a block diagram which shows the function which adds a reflected sound effect to a synthetic voice signal.

Hereinafter, modes for carrying out the present invention will be described in detail with reference to the drawings. FIG. 1 is a sectional view showing a configuration example of a mouthpiece in an embodiment of an electronic musical instrument according to the present invention. The mouthpiece 100 includes a mouthpiece body 102, a blowing port 103, and an exhalation sensor 101. In the present embodiment, the exhalation sensor 101 is a pressure sensor that detects at least one of the pressure of the exhalation and the flow rate of the exhalation that accompany the blowing of the blower. The exhalation sensor 101 is protected by a protective wall 104 so as not to get wet with saliva or the like mixed from the insufflation port 103 at the time of blowing, and detects, for example, the pressure of the exhaled air flowing in through the hole 105.

FIG. 2 is a diagram illustrating a hardware configuration example of the embodiment of the electronic musical instrument 200. The breath of the wind blower is detected as a sensor output signal 208 by the exhalation sensor 101 in the mouthpiece 100 of FIG. The sensor output signal 208 is converted into a digital signal by the A/D (analog/digital) conversion unit 202, then converted into a blow pressure signal 209, and input to the CPU (central processing unit) 201.

It should be noted that an AC cut element that cuts a frequency component equal to or higher than the fundamental frequency of human voice may be arranged on the output side of the A/D conversion unit 202. Alternatively, an AC cut element may be arranged as an analog circuit before the A/D conversion unit 202.

Further, the sensor output signal 208 is converted into a digital signal by the A/D conversion unit 204 after a DC (direct current) cut element 203 such as a capacitor cuts off a direct current component sufficiently lower than the fundamental frequency of human voice. And is input to the CPU 201 as a human voice signal 210. In the present embodiment, the A/D conversion units 202 and 204 and the DC cut element 203 (and the AC cut element) are used to output the sensor output signal 208 by the blow pressure signal 209 indicating the pressure of the expiratory air accompanying the blowing and the human voice. A separation unit that separates the human voice signal 210 indicating the pressure of the exhaled air is formed.

The operation key 205 is a key for the wind player to specify a pitch and a tone color. When the wind player presses the operation key 205, the operation key 205 is fetched into the CPU 201 as the pitch information 211 and the tone color selection information 212, and the instrument sound is reproduced. It is a factor that determines the pitch and tone of the. That is, the operation key 205 functions as a pitch output section and a tone color selection information output section.

The CPU 201 generates a musical instrument sound signal based on the blow pressure signal 209, the pitch information 211, and the tone color selection information 212. Further, the CPU 201 selects a filter parameter corresponding to the tone color selection information 212, configures a voice synthesis filter based on the filter parameter, generates a sound source pulse train having a pitch cycle corresponding to the pitch information 211, and performs voice synthesis. It is input to the filter, a human voice synthesis sound is generated as an output from the voice synthesis filter, and the generated human voice synthesis sound is controlled based on the human voice signal 210. Then, the CPU 201 generates and outputs a musical sound signal 213 from the musical instrument sound signal and the human voice synthetic sound signal.

The output digital musical sound signal 213 is converted into an analog audio signal by the D/A conversion unit 206, amplified by the sound system 207 to a volume that can be heard by the wind performers, and is particularly emitted from a speaker (not shown). ..

FIG. 3 is a diagram showing a block configuration example of the CPU 201 of FIG. The CPU 201 is composed of, for example, a DSP (digital signal processor), and each functional portion shown in FIG. 3 is realized as a digital signal processing function of the DSP. Note that these functional parts may be realized as normal software processing in which the processor executes the control program.

The first Wave Generator 301 (musical instrument sound generation unit) generates a musical instrument sound signal 309 from the pitch information 211 and the tone color selection information 212 from the operation keys 205 of FIG. In the present embodiment, the first Wave Generator 301 generates the instrument sound signal 309 by, for example, the pulse modulation method or the sine wave synthesis method, but may be a sampling sound source.

The musical instrument sound signal 309 output from the first Wave Generator 301 is amplification-controlled by the AMP (amplification block) 304 based on the blow pressure envelope 311 generated from the blow pressure signal 209, and then input to the adder 306.

The second Wave Generator 302 selects a filter parameter corresponding to the tone color selection information 212, configures a voice synthesis filter based on the filter parameter, generates a sound source pulse train having a pitch period corresponding to the pitch information 211, and performs voice synthesis. It is input to the filter, and a human voice synthesis sound signal 310 is generated and output as the output from the voice synthesis filter.

The human voice synthesis sound signal 310 output from the second Wave Generator 302 is amplification-controlled by the AMP 305 based on the human voice envelope 312 generated from the human voice signal 210, and then input to the adder 306.

The adder (adding unit) 306 adds the instrument sound signal 309 output from the AMP 304 and the human voice synthesized sound signal 310 output from the AMP 305, mixes them, and outputs the mixed signal. Further, the clip unit 307 executes a clipping process, which will be described later, on the output of the adder 306 when the player specifies clipping with the operation key 205. As a result, the final tone signal 213 is output from the CPU 201 of FIG.

FIG. 4 is a diagram showing an example of a block configuration of the second Wave Generator 302 (human voice synthetic speech generation unit) of FIG. The second Wave Generator 302 includes a calculator 401, a pulse generator 402, and a formant filter 403. Further, the formant filter 403 includes a first filter 404, a second filter 405, and an adder 406.

As shown in FIG. 4, the formant filter 403 includes a first filter 404 and a second filter 405. The first filter 404 is a filter that maximizes the first formant of the phoneme corresponding to the synthesized human voice signal 310 to be generated. The filter parameter (hereinafter, referred to as “first filter parameter”) is given as a set of frequency and gain corresponding to the first formant of the phoneme. Similarly, the second filter 405 is a filter that maximizes the second formant of the phoneme corresponding to the synthesized human voice signal 310 to be generated. The filter parameter (hereinafter referred to as “second filter parameter”) is given as a set of frequency and gain corresponding to the second formant of the phoneme. Now, when the tone color selection information 212 is input, the formant filter 403 selects the first filter corresponding to the input tone color selection information 212 from the set of the first filter parameter and the second filter parameter stored for each tone color. The parameters and the second filter parameters are selected and set in the first filter 404 and the second filter 405.

For example, FIG. 5A shows the result of FFT (Fast Fourier Transform) analysis of the phoneme "U" of a human voice. In FIG. 5A and FIG. 6A described later, the horizontal axis represents frequency (Hz: Hertz) and the vertical axis represents amplitude (dB: decibel). In FIG. 5A, the first formant of a large peak of 400 Hz is made by the first filter 404, and the second formant of 2000 Hz which is much smaller than that is made by the second filter 405. The formant filter 403 adds the outputs of the first filter 404 and the second filter 405 by the adder 406 to synthesize the synthesized voice signal 310 of the phoneme “U”. The filter characteristic obtained by combining the first filter 404 and the second filter 405 of the formant filter 403 which operates based on the FFT analysis result of FIG. 5A is as shown in FIG. 5B. In FIG. 5B and FIG. 6B described later, the horizontal axis represents frequency (Hz) and the vertical axis represents amplitude (dB).

Although there are third formants, fourth formants, and the like in actual speech, the first filter 404 simulating only the first formant and the second formant is used for the purpose of creating the human voice synthesis sound signal 310 as a sound effect. Also, a simple calculation using the two filters of the second filter 405 is sufficient. As a result, the calculation performance required for the second Wave Generator 302 can be suppressed, and a cost reduction can be realized when the electronic musical instrument 200 is manufactured. Note that if the processor that implements the formant filter 403 has sufficient computing power, a sophisticated filter such as a linear prediction filter may be used instead of the formant filter 403. In that case, the filter parameter is a set of linear prediction parameters for each phoneme for forming the linear prediction filter.

When the first filter 404 and the second filter 405 in the formant filter 403 are implemented as digital signal processing, a resonance filter that can obtain a sharp peak even if the order is small is preferable.

FIG. 6A shows the result of the FFT analysis of the phoneme "A" of the human voice. In FIG. 6A, the first formant having a large mountain of 1000 Hz is formed by the first filter 404, and the second formant having a slightly smaller size of 1800 Hz is formed by the second filter 405. The outputs of the first filter 404 and the second filter 405 are added by the adder 406 to synthesize the synthesized voice signal 310 of the phoneme "O". The filter characteristic obtained by combining the first filter 404 and the second filter 405 of the formant filter 403 which operates based on the FFT analysis result of FIG. 6A is as shown in FIG. 6B.

As can be seen by comparing FIG. 5B and FIG. 6B, the phoneme “U” has a large amplitude at a low frequency near 400 Hz, and the phoneme “A” has a small amplitude at a low frequency. In the present embodiment, the filter parameter (first filter parameter+second filter) of the formant filter 403 having a larger vowel “U” or “A” effect according to the tone color selection information 212, that is, the musical instrument sound selected by the player. Allocate filter parameters).

On the other hand, in FIG. 4, the pulse generator 402 generates a sound source pulse train having a pitch period corresponding to the pitch information 211, for example, as shown in FIG. 7A, and sends the sound source pulse train to the formant filter 403. Input. The first filter 404 and the second filter 405 in the formant filter 403 respectively execute the filtering process on the input sound source pulse train, and the outputs are added by the adder 406. The synthesized human voice signal 310 having a frequency characteristic similar to that of 6(a) and having a time domain waveform as shown in FIG. 7(b) is generated.

Here, the pulse generator 402 does not determine the pitch period of the sound source pulse train only by the pitch frequency corresponding to the pitch information 211 output from the operation key 205 that specifies the pitch, but the calculator 401 determines The pitch information 211 is changed based on the human voice envelope 312 generated from the human voice signal 210, and the pitch information output from the calculator 401 after the change is input to the pulse generator 402. it can. Accordingly, when the adder 306 of FIG. 3 mixes the musical instrument sound signal 309 and the human voice synthesized sound signal 310, the pitch of the musical instrument sound signal 309 is determined by the pitch information 211, and the human voice synthesized sound signal 310 is obtained. The pitch of the pitch can be determined by the pitch frequency of the pitch information 211 which is slightly changed based on the human voice envelope 312 obtained by the winder playing the pitch. As a result, a subtle undulation can be generated between the instrument sound signal 309 and the human voice synthesis sound signal 310 according to the human voice envelope 312, and various changes can be made to the output musical sound signal 213. Become. In this case, the calculator 401 substantially follows the pitch information 211, but shifts the pitch frequency to either up or down in proportion to the magnitude of the value of the human voice envelope 312 by up to a quarter semitone. May act like. If there is a slight deviation, it is possible to obtain the effect that "swell" occurs in the output sound. However, if the difference between the pitch frequencies of the two is too large, the output tone signal 213 sounds like two separate sounds, so the calculator 401 further causes the pitch information 211 indicated by the player. By controlling the degree of shift by, it is possible to obtain a rough "swell" that a human blows.

FIG. 8 is a table showing, for each musical instrument tone color specified as the tone color selection information 212, the correspondence between the mouthpiece and the pronunciation characteristic at the time of growling, and the type of phoneme suitable for the formant filter 403 to generate. .. In this way, when, for example, the tone color of “trumpet” is specified as the tone color selection information 212, the formant filter 403 selects the filter parameter (first filter parameter+second filter parameter) corresponding to the phoneme “bu”. By doing so, it is possible to optimally reproduce the growling sound when the trumpet is played. Similarly, the vowel "u" for the horn, the phonological "vo" for the clarinet, the vowel "o" for the saxophone, the vowel "a" for the flute, and the vowel "i" for the oboe are preferred. is there.

As described above, the second Wave Generator 302 has the pitch information 211 output from the operation key 205 of the pitch designated by the wind instrument player and the tone color selection information 212 output from the operation key 205 of the tone color designated by the wind instrument player. And a human voice signal 210 separated from the breath sensor 101 of the mouthpiece 100, a human voice synthetic sound signal 310 is generated. As a result, even if the wind instrument player cannot utter the special rendition style sound well, it is possible to easily obtain the special rendition style sound that can reflect the intention of the wind instrument player as the human voice synthesized sound signal 310. ..

The operation of the embodiment of the electronic musical instrument 200 having the above configuration will be described below. First, the human voice waveform blown by the wind blower at the blowing opening 103 will be described. FIG. 9 is a diagram showing an example of waveforms of the sensor output signal 208 (FIG. 9A) and the human voice signal 210 (FIG. 9B) of FIG. 2 when the player blows normally. On the other hand, FIG. 10 shows an example of the waveforms of the sensor output signal 208 (FIG. 10A) and the human voice signal 210 (FIG. 10B) when the player is groaning “U-” in consideration of the special performance. FIG. In both figures, the vertical axis represents the magnitude of the amplitude waveform of each signal, and the horizontal axis represents time. As can be read from the relationship between FIGS. 9 and 10, the DC cut element 203 and the A/D converter 202 shown in FIG. 2 are combined without installing a dedicated microphone in the mouthpiece 100 of FIG. As shown in FIG. 10B, it can be seen that the circuit can detect the human voice signal 210 corresponding to the voice of the wind blower from the waveform of the exhalation sensor 101. Then, the second Wave Generator 302 of FIG. 3, which is realized as digital signal processing in the CPU 201, generates a human voice synthesized sound signal 310 based on the human voice signal 210, thereby reflecting the intention of the wind performer's special performance style. It is possible to generate the synthesized human voice signal 310.

FIG. 11 is a diagram showing a waveform example in the present embodiment. FIG. 11A is a diagram showing a waveform example of the musical instrument sound signal 309 generated by the first Wave Generator 301 of FIG. As described above, the first Wave Generator 301 can generate the instrument sound signal 309 as a pulse-modulated waveform as illustrated in FIG. 11A by generating the instrument sound signal 309 by, for example, the pulse modulation method. it can.

FIG. 11B is a diagram showing a waveform example of the human voice synthesis sound signal 310 generated by the second Wave Generator 302 of FIG. The second Wave Generator 302 synthesizes the calculation results of the first filter 404 and the second filter 405 in the formant filter 403 with the adder 406 to generate a phoneme waveform as illustrated in FIG. 11B.

11C and 11D are diagrams showing examples of processing waveforms in the clip unit 307 of FIG. For the waveform data output from the adder 306, the clipping unit 307 replaces the amplitude value larger than the positive threshold value with the positive threshold value and the amplitude value smaller than the negative threshold value with the negative threshold value, as shown in FIG. 11C. By replacing, the waveform data as shown in FIG. 11D is output. As a result, the output musical tone signal 213 can include a complicated signal component.

FIG. 12 is an overall flowchart showing a processing example of the control operation executed by the CPU 201 to realize the above control operation. This processing is realized as an operation in which the CPU 201 executes a control program stored in a memory such as a ROM (read only memory) which is not particularly shown. Each functional portion shown in FIG. 3 is realized by digital signal processing of the DSP. Has been realized. The cycle in which the series of steps S1201 to S1213 completes once is the sampling frequency (44.1 kilohertz) of the D/A conversion unit 206 in FIG. 2, and requires about 22 microseconds.

First, the CPU 201 reads the state of the tone color operation key 205 (step S1201).

Next, the CPU 201 determines the tone color from the value of the tone color operation key 205 and outputs the tone color selection information 212 (step S1202).

Next, the CPU 201 reads the state of the pitch operation key 205 (step S1203).

Next, the CPU 201 determines the pitch from the value of the tone color operation key 205 and outputs the pitch information 211 (step S1204).

Next, the CPU 201 reads the blow pressure signal 209 and the human voice signal 210 from the exhalation sensor 101 via the hardware circuit of FIG. 2 (step S1205).

Next, the CPU 201 generates the blow pressure envelope 311 from the blow pressure signal 209 (step S1206). This generation process is, for example, an appropriate smoothing process such as a moving average process for the blow pressure signal 209.

Next, the CPU 201 generates the musical instrument sound signal 309 by executing the processing corresponding to the above-mentioned first Wave Generator 301 and AMP 304 of FIG. 3 (step S1207).

Next, the CPU 201 generates a human voice envelope 312 from the human voice signal 210 obtained in step S1205 (step S1208). Similar to step S1206, this generation process is an appropriate smoothing process such as a moving average process for the human voice signal 210.

Next, the CPU 201 executes the processing corresponding to the arithmetic unit 401 and the pulse generator 402 in FIG. 4 in the second Wave Generator 302 in FIG. 3 to generate the sound source from the human voice envelope 312 and the pitch information 211. A pulse wave of a pulse train is created (step S1209).

Next, the CPU 201 receives the pulse wave of the sound source pulse train created in step S1209 as an input and executes the process corresponding to the above-described formant filter 403 in FIG. 4 in the second Wave Generator 302 in FIG. 3 (step S1210). As a result, the CPU 201 generates the human voice synthesized voice signal 310 (step S1211).

Next, the CPU 201 executes the process corresponding to the adder 306 of FIG. 3 described above to add the musical instrument sound signal 309 generated in step S1207 and the synthesized human voice signal 310 generated in step S1211 ( Step S1212).

Further, the CPU 201 executes the processing corresponding to the clipping unit 307 of FIG. 3 described above, thereby executing the clipping processing on the musical tone signal output from the adder 306 (step S1213).

Finally, the CPU 201 transmits the musical sound signal 213 obtained as a result of the clipping processing in step S1212 as digital audio to the D/A conversion unit 206 (step S1214).

As described above, the CPU 201 ends the process of one sampling period and returns to the process of step S1201 at the timing of the next sampling period.

According to the above-described embodiment, a musical tone signal in which the synthesized voice signal 310 of the human voice is easily mixed with the musical instrument sound signal 309 as intended based on the vocalization operation of the winder without requiring a dedicated microphone or the like. 213 can be obtained, and it becomes possible to pronounce a more varied musical tone. More specifically, since the blower's blowing pressure taken in by the exhalation sensor 101 is separated into a direct current component and an alternating current component and made into two elements, the exhalation sensor 101 is effective without requiring a dedicated microphone or the like. It is possible to contribute to the cost reduction of the electronic musical instrument 200.

Further, according to the above-described embodiment, the second Wave Generator 302 of FIG. 3 is realized by the formant filter 403 including the first filter 404 and the second filter 405 shown in FIG. In this case, the filter parameters corresponding to one timbre designated by the timbre selection information 212 are the first filter parameter (frequency data and gain data) corresponding to the first formant and the second filter parameter corresponding to the second formant. Since only (frequency data and gain data) is required, it is only necessary to store four data in total for each tone color. For this reason, the data capacity required to generate the human voice synthesis sound signal 310 corresponding to various tones can be made extremely small, and the cost of the electronic musical instrument 200 can be reduced.

Further, according to the above-described embodiment, in the electronic musical instrument 200, the winder makes a voice while blowing his breath, and the special playing method “growl” unique to the wind instrument is effective without depending on the frequency of the voice of the winder. It has the effect of sounding. Further, in the electronic musical instrument 200, since the human voice synthesis sound is made to correspond to the selected musical instrument sound, there is an effect that the special playing method "growl" peculiar to the wind instrument can be effectively sounded without depending on the words emitted by the wind instrument player. is there.

Although the above embodiment is a method of producing a human voice synthesis sound for the growl playing method, it is necessary to reflect the effect in consideration of the reflection of the voice in the live wind instrument in the musical tone signal 213 output from the electronic musical instrument 200. Other possible embodiments will be described below.

For example, in the saxophone shown in FIG. 13(a), the voice uttered from the mouthpiece held in the mouth comes back with a part thereof reflected at the U-shaped portion of the tube. Therefore, a voice transmitted over a distance twice the length 1301 (for example, 50 cm) from the mouthpiece to the U-shaped portion is added to the voice uttered from the mouthpiece. In the air, the speed at which the sound propagates is about 333 m per second. Therefore, as shown in FIG. 13B, a reflected sound component 1303 delayed by 3 ms with respect to the original synthesized human voice 1302 is generated. It is considered that reflected sounds can be simulated by adding them.

Therefore, in another embodiment, a delay processing unit 1100 and an adder 1403 as shown in FIG. 14 to which the synthesized human voice signal 310 output from the AMP 305 of FIG. 3 is input are provided. The delay processing unit 1100 includes a delay unit 1401 and an attenuator 1402. The delay unit 1401 delays the synthesized human voice signal 310 by, for example, 3 ms, and the attenuator 1402 attenuates the delayed signal, whereby one of the reflected sounds 1303 shown in FIG. 13B is reflected. Generates a reflected sound signal corresponding to. At this time, since the delay time in the delay unit 1401 and the attenuation amount in the attenuator 1402 differ depending on the type of wind instrument, these parameters can be controlled according to the tone color selection information 212 in FIG. 2, for example. In this way, the reflected sound signal generated by the delay processing unit 1100 is mixed with the original human voice synthesized sound signal 310 and output by the adder 1403.

With the above-described configuration, in the electronic musical instrument 200, the human voice synthesis sound is delayed and added to the original human voice synthesis sound. Therefore, in the special playing method “growl” peculiar to a wind instrument, a sound effect like that played by a wind pipe is obtained. It is possible.

Regarding the above embodiment, the following supplementary notes will be further disclosed.
(Appendix 1)
An exhalation sensor that detects a state of exhalation and outputs a signal corresponding to the detected state of exhalation,
A pitch specification key that specifies the pitch of the musical sound to be generated,
A tone designating key that designates the tone color of the musical sound to be generated,
An output signal from the exhalation sensor, a separation unit that separates a blowing signal indicating a state of exhalation generated by a blow and a human voice signal indicating a state of exhalation generated by a human voice,
A tone color selection output process that outputs tone color selection information based on an operation state of the tone color designation key, a tone pitch output process that outputs tone pitch information based on an operation state of the tone pitch designation key, the wind signal and the Musical instrument sound output processing for outputting a musical instrument sound based on pitch information and the tone color selection information, pulse train generation processing for generating a pulse train signal having a pitch cycle corresponding to the pitch information, and the pulse train signal, A processing unit that executes a filtering process for outputting a human voice synthesized voice by filtering with a voice synthesis filter corresponding to the tone color selection information;
An electronic musical instrument equipped with.
(Appendix 2)
The separating unit uses the blowing signal as a signal obtained by cutting a frequency component equal to or higher than the fundamental frequency of the human voice from a frequency component included in a sensor output signal from the exhalation sensor, and a frequency component lower than the fundamental frequency of the human voice. The electronic musical instrument according to Appendix 1, wherein the cut signal is used as the human voice signal.
(Appendix 3)
In the filter process, the processing unit acquires a filter parameter corresponding to the phonetic color selection information, filters the pulse train signal by a voice synthesis filter based on the acquired filter parameter, and outputs a human voice synthesized voice. The electronic musical instrument according to appendix 1 or 2, which executes processing.
(Appendix 4)
The filter parameter holds a set of frequency and gain of the formants for each of a plurality of formants corresponding to the phonemes corresponding to the tone color selection information,
The electronic musical instrument according to appendix 3, wherein the voice synthesis filter is a formant filter provided with a number of resonance filters that resonate according to a set of frequencies and gains of the formants, the number corresponding to the plurality of formants.
(Appendix 5)
The filter parameter corresponding to the timbre selection information is a linear prediction parameter corresponding to the phoneme corresponding to the timbre selection information,
The electronic musical instrument according to Appendix 3, wherein the voice synthesis filter is a linear prediction filter configured by the linear prediction parameter.
(Appendix 6)
In the pitch output process, the processing unit changes pitch information determined based on an operation state of the pitch designation key, based on a human voice envelope generated based on the human voice signal, and outputs the pitch information. 6. The electronic musical instrument according to any one of appendices 1 to 5, which executes processing.
(Appendix 7)
The processing unit executes a process of controlling the amplitude of the synthesized human voice based on a human voice envelope generated from the human voice signal in the synthetic voice output process. Or electronic musical instrument described.
(Appendix 8)
8. The electronic musical instrument according to any one of appendices 1 to 7, wherein the processing unit further executes an addition process of adding the musical instrument sound and the human voice synthesis sound and outputting the resultant as a musical tone signal.
(Appendix 9)
In the addition processing, the processing unit adds the musical instrument sound and the human voice synthesis sound, clips the addition result with a predetermined threshold, and outputs the clipped signal as the musical tone signal. The electronic musical instrument according to appendix 8, which is executed.
(Appendix 10)
The processing unit further delays the human voice synthesis sound signal, and executes a delay process of adding and outputting the delayed human voice synthesis sound signal and the original human voice signal. The electronic musical instrument according to any one of 1.
(Appendix 11)
A breath sensor for detecting the state of exhalation and outputting a signal corresponding to the detected state of exhalation, a pitch designation key for designating the pitch of a musical tone to be generated, and a tone color of the musical tone to be generated. An electronic musical instrument having a tone color designating key and a separating unit for separating an output signal from the exhalation sensor into a blowing signal indicating a state of exhalation generated by a blow and a human voice signal indicating a state of exhalation generated by a human voice. A method for generating a musical sound used in, wherein the electronic musical instrument comprises:
Outputs tone color selection information based on the operation state of the tone color designation key,
Outputs pitch information based on the operation state of the pitch designation key,
Output a musical instrument sound based on the wind signal, the pitch information, and the tone color selection information,
Generating a pulse train signal having a pitch period corresponding to the pitch information,
A method for generating a musical tone, wherein the pulse train signal is filtered by a voice synthesis filter corresponding to the tone color selection information to output a human voice synthesized voice.
(Appendix 12)
A breath sensor that detects the state of exhalation and outputs a signal corresponding to the detected state of exhalation, a pitch designation key that designates the pitch of a musical tone that should be generated, and a tone color of the musical tone that should be generated An electronic musical instrument having a tone color designating key and a separating unit for separating an output signal from the exhalation sensor into a blowing signal indicating a state of exhalation generated by a blow and a human voice signal indicating a state of exhalation generated by a human voice. To the computer used as
Outputting tone color selection information based on the operation state of the tone color designation key,
Outputting pitch information based on the operation state of the pitch designation key,
Outputting a musical instrument sound based on the wind signal, the pitch information, and the tone color selection information,
Generating a pulse train signal having a pitch period corresponding to the pitch information;
A step of filtering the pulse train signal with a voice synthesis filter corresponding to the tone color selection information to output a human voice synthesis voice;
A program to execute.
(Appendix 13)
An output signal from an exhalation sensor that detects a state of exhalation and outputs a signal corresponding to the detected state of exhalation, indicates a state of exhalation generated by a blowing signal and a human voice that indicates the state of exhalation generated by blowing. A separation unit that separates the human voice signal shown,
Based on the tone selection output process that outputs tone selection information based on the tone selection key operation state that specifies the tone color of the tone that should be generated, and the operation state of the tone pitch specification key that specifies the tone pitch of the tone that should be generated. Pitch output processing for outputting pitch information, instrument sound output processing for outputting a musical instrument sound based on the wind signal, the pitch information, and the tone color selection information, and a pitch cycle corresponding to the pitch information. A pulse train generation process for generating a pulse train signal having, and a processing unit for executing a filter process for filtering the pulse train signal by a voice synthesis filter corresponding to the tone color selection information and outputting a human voice synthesized voice,
Musical tone generator equipped with.

100 Mouthpiece 101 Exhalation Sensor 102 Mouthpiece Body 103 Blow-in Port 104 Protective Wall 105 Hole 200 Electronic Musical Instrument 201 CPU201
202, 204 A/D conversion unit 203 DC cut element 205 Operation key 206 D/A conversion unit 207 Acoustic system 208 Sensor output signal 209 Blowing pressure signal 210 Human voice signal 211 Pitch information 212 Sound color selection information 213 Musical sound signal 301 First Wave Generator
302 Second Wave Generator
303 arithmetic unit 306, 406, 1403 adder 307 clip section 308 human voice selection information 309 musical instrument sound signal 310 human voice synthesized sound signal 311 blowing pressure envelope 312 human voice envelope 401 arithmetic unit 402 pulse generator 403 formant filter 404 first filter 405 2nd filter 1100 Delay processing part 1401 Delay part 1402 Attenuator

Claims (13)

From the output signal of the exhalation sensor that detects the state of exhalation, extract the blowing signal indicating the state of exhalation generated by the blowing,
The instrument signal is output based on the wind signal, the specified pitch, and the specified tone color,
An electronic musical instrument comprising a processing unit for filtering a pulse train signal having a pitch period corresponding to the specified pitch by a voice synthesis filter corresponding to the designated tone color and outputting a human voice synthesis sound. The processing unit acquires a filter parameter corresponding to the specified tone color, executes a process of filtering the pulse train signal with a voice synthesis filter based on the acquired filter parameter, and outputting the human voice synthesis voice. The electronic musical instrument according to claim 1 . The filter parameter holds a set of frequency and gain of the formants for each of a plurality of formants corresponding to the phonemes corresponding to the designated timbre,
The electronic musical instrument according to claim 2 , wherein the voice synthesis filter is a formant filter including a number of resonance filters that resonate according to a set of frequencies and gains of the formants, the number corresponding to the plurality of formants. The filter parameter corresponding to the specified tone color is a linear prediction parameter corresponding to the phoneme corresponding to the specified tone color,
The electronic musical instrument according to claim 2 , wherein the voice synthesis filter is a linear prediction filter configured by the linear prediction parameter. Wherein the processing unit, after adding the said person voice synthesized speech and the musical instrument sound, the result of adding the said person voice synthesized speech and the musical instrument sound is clipped at a predetermined threshold value, comfortable sound signal after the clipping executing a process of outputting a signal, an electronic musical instrument according to any one of claims 1 to 4. The processing unit separates the output signal of the exhalation sensor into the blowing signal indicating the state of exhalation generated by blowing and the human voice signal indicating the state of exhalation generated by human voice, and based on the human voice signal. The electronic musical instrument according to any one of claims 1 to 5, wherein the electronic musical instrument controls the amplitude or pitch of the synthetic voice. The processing unit, the blowing signal is a signal obtained by cutting a frequency component having a fundamental frequency of the human voice or higher from a frequency component included in a sensor output signal from the breath sensor, and a frequency component lower than the fundamental frequency of the human voice. The electronic musical instrument according to claim 6 , wherein the cut signal is used as the human voice signal. The electronic musical instrument according to claim 6 or 7 , wherein the processing unit executes a process of controlling an amplitude or a pitch of the synthesized human voice based on a human voice envelope generated from the human voice signal. The processing unit further delays the human voice synthetic sound signal which is the signal of the human voice synthetic voice, and adds the delayed human voice synthetic sound signal and the original human voice signal to output the delay processing. The electronic musical instrument according to any one of claims 6 to 8, which executes the following. An exhalation sensor that detects a state of exhalation and outputs a signal corresponding to the detected state of exhalation;
And Ruoto high-specified key to specify the pitch of a tone to be generated,
And Ruoto color specifying key to specify the tone of the tone to be generated,
Equipped with
The processing unit is
An output signal from the exhalation sensor, a separation unit that separates a blowing signal indicating a state of exhalation generated by a blow and a human voice signal indicating a state of exhalation generated by a human voice,
A tone color selection output process that outputs tone color selection information based on an operation state of the tone color designation key, a tone pitch output process that outputs tone pitch information based on an operation state of the tone pitch designation key, the wind signal and the Musical instrument sound output processing for outputting a musical instrument sound based on pitch information and the tone color selection information, pulse train generation processing for generating a pulse train signal having a pitch cycle corresponding to the pitch information, and the pulse train signal, The electronic musical instrument according to any one of claims 1 to 9 , wherein the electronic musical instrument is filtered by a voice synthesis filter corresponding to the tone color selection information and outputs a human voice synthesized voice. Electronic musical instruments
From the output signal of the exhalation sensor that detects the state of exhalation, extract the blowing signal indicating the state of exhalation generated by the blowing,
The instrument signal is output based on the wind signal, the specified pitch, and the specified tone color,
A musical sound generating method in which a pulse train signal having a pitch period corresponding to the designated pitch is filtered by a voice synthesis filter corresponding to the designated tone color, and a human voice synthesis sound is output. On the computer,
From the output signal of the exhalation sensor that detects the state of exhalation, extract the blowing signal indicating the state of exhalation generated by the blowing,
A process of outputting a musical instrument sound based on the wind signal, a specified pitch, and a specified tone color;
And processing you outputs human voice synthesized sound a pulse train signal, and filtered by the speech synthesis filter corresponding to the tone color of the specified having a pitch period corresponding to the pitch to be the designated,
A program to execute. An output signal from an exhalation sensor that detects a state of exhalation and outputs a signal corresponding to the detected state of exhalation, indicates a state of exhalation generated by a blow signal and a human voice that indicates the state of exhalation generated by blowing. A separation unit that separates the human voice signal shown,
Based on the tone color selection output process that outputs tone color selection information based on the operation state of the tone color specification key that specifies the tone color of the tone that should be generated, and the operation state of the tone pitch specification key that specifies the pitch of the tone that should be generated Pitch output processing for outputting pitch information, instrument sound output processing for outputting a musical instrument sound based on the wind signal, the pitch information, and the tone color selection information, and a pitch period corresponding to the pitch information. A pulse train generation process for generating a pulse train signal having, and a processing unit for executing a filter process for filtering the pulse train signal by a voice synthesis filter corresponding to the tone color selection information and outputting a human voice synthesized voice,
Tone generator equipped with.