JP2670306B2 - Musical tone synthesizing apparatus and musical tone synthesizing method - Google Patents

Description

The present invention relates to a tone synthesis system for synthesizing one tone from a plurality of tone components.

[Summary of the Invention] The present invention synthesizes a musical tone whose frequency is proportional to a pitch-dependent frequency and a musical tone which is not necessarily proportional to each other among one musical tone to synthesize a musical tone closer to a real musical tone. I was able to generate.

[Prior Art] With a conventional electronic musical instrument, in order to realize a musical sound that is closer to a real sound, an external musical instrument is recorded and played as a sound source. To record, record the tone waveform in the form of PCM data,
The method of reading and reproducing this was adopted. In performing the music, the frequency of the recorded external sound is changed according to the designated pitch, so that a low sound to a high sound can be realized. Also, recording external sound for each specific range,
For example, there is a type in which sampling is performed for each octave, and the frequency of a recorded external sound is changed in accordance with a pitch specified within this one octave.

[Problems to be Solved by the Invention] However, considering a case where a sound to be recorded as an external sound, for example, a real piano sound is recorded as an external sound and used as a sound source, the piano sound vibrates when a key is pressed. The sound of the string vibration generated from the string and emitted from the soundboard, and the sound of the soundboard itself generated by the impact applied by the hammer when the soundboard that strengthens the sound of this string is struck (striking sound or noise sound). ) There is. In this case, the frequency of the sound of the string increases in proportion to the pitch, but the frequency of the sound of the soundboard itself does not necessarily increase in proportion to the pitch and does not change so much. This is the same for other stringed instruments such as a violin.

Fig. 2 shows the frequency spectrum of the sound of the piano soundboard itself. The frequency spectrum of the sound of the soundboard when the first low-pitched key on the piano keyboard is struck and the 73rd key of the pitch are shown. In the frequency spectrum of the sound of the soundboard when struck, the sound pressure level slightly shifts in the low frequency range, but the sound pressure level in the middle and high frequency range is not so different, and the spectrum The formant shapes of the components are almost the same.

Therefore, if you change the frequency according to the specified pitch by PCM recording the piano sound including the sound of the soundboard itself, such frequency components are hardly influenced by the pitch or proportional to the pitch frequency, A sound different from the piano sound is generated and emitted.

Fig. 1 (1) shows the piano sound in PC for each octave
An example of M-recording and changing the frequency of the recorded piano sound according to the specified pitch is shown.Since the frequency of the string sound originally increases in proportion to the pitch, continuous frequency changes Obtainable. However, the sound of the soundboard itself is essentially constant regardless of the pitch. The change in sound according to will not be continuous.
Of course, it is sufficient to record the piano sound for each key, not for each octave, but this causes a problem that the memory capacity of the musical tone waveform becomes very large. This was the same even when the external sound was not recorded and various waveforms were combined to synthesize a musical sound similar to the external sound.

The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide a musical tone synthesizing method capable of generating a musical tone corresponding to each pitch extremely close to a real musical tone with a small memory capacity. There is.

[Means for Solving the Problem] In order to achieve the above object, in the present invention, the first generating means generates a tone component whose frequency is not necessarily proportional to the pitch of one tone, The second generating means generates a tone component whose frequency is substantially proportional to the pitch of the one tone. The pitch control means controls the generated musical tone component to have a frequency corresponding to the designated pitch for the second generating means, and the synthesizing means controls the musical tone component from the first generating means. And the musical tone components from the second generating means are combined.

[Operation] In the above configuration, like the sound of the soundboard of the piano itself,
If the frequency is not proportional to the pitch, the first generating means generates the musical tone component, and if the frequency is proportional to the pitch, like the sound of a piano string, the second generating means generates the musical tone. By generating the components and synthesizing them, the tone corresponding to each pitch becomes very close to the real tone. As shown in FIG. 1 (2), the tone component whose frequency is not proportional to the pitch has a constant frequency (A) regardless of the pitch or a stepwise change (B). It may increase little by little according to the pitch (C), may increase exponentially (D), or may be in a state where there is no part (E).

[Embodiment] FIG. 3 shows an overall circuit diagram of an electronic musical instrument according to the musical sound synthesizing method of the present invention. In this embodiment, a piano sound can be synthesized and output, and one piano musical sound is shown in FIG. The following four tone components as shown in FIG.

(1) A sounding component (attack component) that mainly occurs at the rising (attack) portion of the envelope waveform among the musical sound components that commonly occur when a piano string is struck and struck weakly.

This component has a musical tone waveform formed by a spectrum component common to the attack portion among frequency spectrum components of a strong tap sound and a weak tap sound.

(2) A musical tone component (attenuation component) mainly generated in the decay (decay, release) portion of the envelope waveform among the musical tone components commonly generated when a piano string is strongly and weakly struck.

This component has a tone waveform formed by a common spectrum component in the attenuation part among the frequency spectrum components of the strong tap sound and the weak tap sound.

(3) A tone component (striking component) generated only when the strings of the piano are struck.

This component has a tone waveform formed by comparing the frequency spectrum components of the strong tap sound and the weak tap sound, and the spectrum component that appears only in the strong tap sound. Since this waveform is only for a characteristic portion appearing in a strong tapping sound, for example, as shown in a frequency spectrum component diagram of FIG. 5 (3), it has no fundamental wave component.

(4) A tone component (noise component) of a striking sound generated when a sound board of a piano is hit.

This component is a sine wave having a frequency corresponding to the frequency value at each peak point of the formant of the soundboard shown in FIG. 2, and a sine wave having an amplitude corresponding to the sound pressure level (gain) at this peak point. It is a composite of several.

In the above (1) to (3), the frequency is proportional to the pitch, but in (4), the frequency is not proportional to the pitch and is almost constant. (1) to (4) are examples obtained by sine synthesis using a computer. (1) to (3) are collected by picking up vibrations of the strings, and (4) is collected by a microphone. You may.

In FIG. 3, a key code and a key-on signal and a key-off signal corresponding to an operation key from a key switch 1 provided on each key of the keyboard are supplied to the CPU 2, and the key-on signal and the key-off signal Touch data indicating a key pressing speed based on the time difference is generated by the touch sensor 3 and is provided to the CPU 2. The touch sensor 3 may detect a key pressing pressure by a pressure sensor. The CPU 2 reads out the frequency number corresponding to the given key code from the frequency number table 4, and together with the touch data and the like, each tone signal generation section 5,
Give to 6.

The frequency number table 4 stores data as shown in FIG. 1 (2), and a frequency number for realizing a frequency corresponding to a given key code (pitch) is read out. For the above-mentioned (1) to (3) piano string attack component, attenuation component, and bang component, a frequency number proportional to the key code is read.
For the noise component of the piano soundboard of (4), the conversion table of any of A, B, C, and D of FIG. 1 (2) is selected, and the frequency number not proportional to the key code is read. It should be noted that, for the pitches A _{3 and} below, the musical sound components of the noise components of the soundboard are not generated and emitted, but the sound components are not limited to this and may be generated in the entire range. However, when the pitch is low, the frequency of the sound of the soundboard itself and the frequency of the sound of the strings are close to each other, the sound of the soundboard is masked by the sound of the strings, and the sound of the soundboard is output synthetically. There is no problem and it may be omitted.

The tone signal generation units 5 and 6 constitute a tone generation system for 32 channels by time division processing, and the tone signal generation unit 5 processes the tone components of the attack component and the attenuation component described above. The musical tone components of the strong hit component and the noise component are processed by the musical tone signal generation unit 6. In this case, since two channels are used in the processing of each component, the tone signal generators 5 and 6 have a 16-tone polyphonic tone generation capability capable of simultaneously emitting up to 16 tones. The tone component signals synthesized by the tone signal generators 5 and 6 are added to an adder 7.
Finally, one tone signal is synthesized, and four tone components are synthesized into one. This synthesized tone signal is D-
The signal is converted into an analog signal by the A converter 8 and is emitted from the sound system 10 via the amplifier 9.

FIG. 4 shows a specific circuit configuration of the tone signal generators 5 and 6, and the waveform memory 11 stores the same in FIG. 5 (1) and (2).
(3) and (4) of FIG. 5 (3) and (4) are stored for each half wavelength. The envelope waveform data memory 12 stores the envelope of the tone component of the attack component or the attenuation component of FIG. 5 (1) (2) or FIG. 5 (3).
The data for forming the envelope of the tone component of the strong hit component or the noise component of (4) is stored. This envelope waveform data storage format stores coefficients that control the time and level of the attack portion, the decay portion, and the time of the release portion. One of the two waveforms and data stored in the waveform memory 11 and the envelope waveform data memory 12 is selected by the data stored in the waveform upper address register 13 and the tone generation component selection register 14, and the waveform upper address register 13 is selected. The data of the upper address of any waveform in the waveform memory 11 is set in
Tone component selection register 14 is set with specified data of a storage area for any of the waveform data of envelope waveform data memory 12, and these two data are set to the same channel of waveform upper address register 13 and tone component selection register 14. And channel synchronization is achieved.

The frequency members from the CPU 2 are set in a frequency member register 15 and are added to an address register 17 by an adder 16.
Is added to the previous frequency number accumulated value, which is set in the address register 17 and applied to the waveform memory 11 via the mode selector 18 as read address data. Only the upper integer part of the accumulated value from the address register 17 is given to the waveform memory 11 as the read address data, the whole integer part and the whole decimal part are sent to the adder 16, and are incremented in steps of the frequency number. Go. Therefore, as the frequency number set in the frequency number register 15 is larger, the read address data of the waveform memory 11 which is an integer part of the accumulated value of the address register 17 is incremented faster, and the waveform data of higher frequency is stored in the waveform memory 11 than the waveform memory 11. Will be output. The waveform data read from the waveform memory 11 is multiplied by the envelope by the multiplier 19, accumulated and synthesized by the accumulator 20, and output.

The mode selector 18 is composed of exclusive OR gates,
Each bit data of the integer part of the accumulated value from the address register 17 is input to each gate, and the most significant bit data of this integer part is input to all the gates. Therefore, since the most significant bit is changed from "0" to "1" when the reading of the final step of the waveform data in the waveform memory 11 is completed, the accumulated value is inverted, and the waveform data is not reversed from the beginning but from the end. Data is read out, resulting in 1
Waveform data for the wavelength will be output. In this case, the data for the latter half wavelength is output as a value that is inverted in the plus or minus based on the most significant bit.

The touch data from the CPU 2 is given to the touch data conversion circuit 21, and the touch data is converted so that the strength of the touch response can be changed according to the curve as shown in FIG. Given. The characteristics of strong (A), weak (C), normal (B), and strong key emphasis (D) of the touch response characteristics are set corresponding to each tone component. That is, the CPU 2 outputs the given touch data, for example, for the attack component in FIG.
, The attenuation component is converted by the characteristic shown in FIG. C, the hit component is converted by the characteristic shown in FIG. A, and the noise component is converted by the characteristic shown in FIG. It is stored in the conversion circuit 21. In the envelope generator 22, an envelope waveform is generated based on the data from the envelope waveform data memory 12, and this is set to a waveform level corresponding to the above-described converted touch data,
It is sent to the multiplier 19 and multiplied by the waveform data.

Waveform upper address register 13, tone component selection register 14, frequency number register 15, address register
17, the touch data conversion circuit 21, the memory in the envelope generator 22, etc. are composed of 32 registers, and 32
A tone generation system for the channels is formed. Attack components (odd components in the tone signal generation unit 6) for odd channels and attenuation components (sound components for the even channel).
Then, the noise component) is assigned, and the accumulator 20 synthesizes the two musical tone components and the musical tone signals of 16 tones. Note that each of the tone signal generation units 5 and 6 has one chip.
It is composed of LSI.

Next, the operation and effect of the electronic device having the above configuration will be described.

First, in order to generate and emit musical tones, a predetermined switch operation causes the frequency number table 4 shown in FIG.
Any one of the change characteristics A, B, C, and D of the noise component of the soundboard may be selected and the keyboard may be operated. here,
When the change characteristic of C is selected and the keys of arbitrary pitches such as A ₃ , A ₄ , A ₅ ... Are operated, the CPU 2 causes the attack component, the attenuation component, for swipe portion sets the 220 Hz, a 440 Hz, a each frequency number register 15 is read out from the frequency number table 4 the frequency number as a 880 Hz sound when the ₅ when the ₄ when the a ₃ . However, for a noise component having a frequency that is not proportional to the pitch, a frequency number that produces a sound much smaller than 220 Hz at A ₃ is read from the frequency number table 4 and set in the frequency number register 15, and A ₄ , A _{In case of 5} , the frequency numbers which are slightly larger than those in case of A ₃ are read out from the frequency number table 4 and set in the frequency number register 15.

As a result, it is possible to set the attack component, attenuation component, and hit component of the strings to a frequency that is proportional to the pitch, and the noise component of the soundboard to a substantially constant frequency that is not proportional to the pitch. It is possible to generate and emit a musical sound extremely close to. In this case, the noise component waveform stored in the waveform memory 11 is one waveform common to all keys of the keyboard, and since the string component sound may be in each range, the number of stored waveforms is small and the memory usage amount is accordingly. It can be small.

In the present embodiment, for the human hearing, the sound component of the soundboard itself is masked by the sound of the strings, and the sound components of the pitch A ₃ or less are not output the noise component of the soundboard. The storage capacity of the table 4 is small, and data processing is easy.

The present invention is not limited to the above embodiment, and can be variously modified without departing from the spirit of the present invention. For example, a musical tone component whose frequency is not proportional to the pitch is the sound of the tube itself, and a musical tone component whose frequency is proportional to the pitch is also applicable to wind instruments and percussion instruments as the vibration sound of the air column in the tube. Each tone component may be a component or waveform other than that shown in FIG. Further, by changing the waveform of the striking sound (noise component) in accordance with the change of the formant shown in FIG. 2, a sound closer to a true piano sound can be realized.

[Effects of the Invention] As described in detail above, according to the present invention, like an actual musical tone, a musical tone of one musical tone whose frequency is not necessarily proportional to a musical tone component whose frequency is proportional to the pitch. By synthesizing components separately, it is possible to generate a tone that is closer to a real tone. Further, unlike the conventional one, all the musical tone components are not made to have a frequency proportional to the pitch.Therefore, in order to bring the tone closer to the actual musical tone, the deviation state between the component whose frequency is proportional to the pitch and the component which is not proportional to it is In order to be the same as the actual musical tone, it is not necessary to store the musical tone waveform for each pitch, and the amount of memory used for storing the waveform can be reduced accordingly.

[Brief description of the drawings]

1 to 6 show an embodiment of the present invention.
FIG. 1 is a diagram showing the contents of the conventional example of the frequency variation characteristics with respect to the pitch of a musical tone component (sound of a string) whose frequency is proportional to the pitch and a musical tone component (sound of the soundboard itself) which is not proportional to the pitch, and the contents of the present invention. The contents of FIG. 1 (2) are also the contents stored in the frequency number table 4, FIG. 2 is a diagram showing the frequency spectrum of the sound at each position of the soundboard of the piano, and FIG. FIG. 4 is an overall circuit diagram of the musical instrument, FIG. 4 is a circuit diagram of a musical tone signal generation unit, FIG. 5 is a diagram showing the contents of each musical tone component of a piano sound, and FIG. 6 is a touch data conversion circuit 21.
5 is a diagram showing conversion characteristics of touch data in FIG. 2 ... CPU, 4 ... Frequency number table, 5, 6, 2
3, 24, 25, 26 ... tone signal generator, 7 ... adder, 11
...... Waveform memory, 12 …… Envelope waveform data memory, 15 …… Frequency number register 20 …… Accumulator, 27 ・ ・ ・
… Mixer.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Yasushi Sato 200 Terashimacho, Hamamatsu-shi, Shizuoka Prefecture Kawai Musical Instruments Co., Ltd. (56) References JP-A-59-75294 (JP, A) JP-A-56-146197 (JP, A) JP-A-1-193899 (JP, A)

Claims (4)

(57) [Claims] 1. A first tone component, of which frequency is substantially proportional to tone pitch, is generated from one tone, and the frequency of the tone is not proportional to tone pitch. A second musical tone component is generated, and the frequency of the second musical tone component changes with respect to the pitch different from the proportional change with respect to the pitch of the first musical tone component, The frequency of the generated first musical sound component is controlled so as to change substantially proportionally to the specified pitch, and the frequency of the generated second musical sound component becomes the specified pitch. On the other hand, it is controlled so as not to be proportional and different from the proportional change of the first musical sound component with respect to the pitch, and to synthesize these controlled first musical sound component and second musical sound component. Characteristic sound synthesizer. 2. The method according to claim 1, wherein the second tone component whose frequency is not proportional to the pitch is a noise sound of the soundboard itself, and the first tone component whose frequency is substantially proportional to the pitch. 2. The musical tone synthesizer according to claim 1, wherein the musical tone component is a sound of the string itself, and the sound of the string itself is an attack component, an attenuation component, and a bang component. 3. The second musical sound component is in a state where it is not generated at a part of the pitch, and the changing characteristic of the second musical sound component with respect to the pitch can be selected from a plurality of changing characteristics. The second musical tone component is a combination of signals having a frequency and an amplitude corresponding to each formant peak point of the frequency spectrum of the soundboard, and the change characteristic with respect to the touch has the first musical tone component and the second musical tone. 3. The musical tone synthesizer according to claim 1, wherein the musical tone is different from the component. 4. A first tone component, the frequency of which is substantially proportional to the pitch of one tone, is generated, and the frequency of the one tone is not compared with the tone pitch. Is generated, and the frequency of the second musical sound component changes with respect to the pitch different from the proportional change with respect to the pitch of the first musical sound component. The frequency of the first musical tone component is controlled to change substantially proportionally to the designated pitch, and the frequency of the second musical tone component generated above is proportional to the designated pitch. The musical tone characterized by being controlled so as not to be changed and different from the proportional change of the first musical tone component with respect to the pitch, and the controlled first musical tone component and the second musical tone component being synthesized. Synthesis method.