JPS5952433B2 - electronic musical instruments - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an electronic musical instrument that generates human voice sounds, and more particularly to an electronic musical instrument that automatically switches the phoneme of the generated human voice sounds.

Recently, in electronic musical instruments, musical tones of human voices, that is, human voice sounds, have been produced in addition to various musical instrument sounds. This type of electronic musical instrument has the desired phoneme “a”, “i”,
Fixed formant special j that corresponds to...
The musical tones of the gender are formed and pronounced using various methods.

By the way, in such electronic musical instruments, when human voices with the same phoneme are repeatedly produced, it feels monotonous to the listener.
Moreover, as you listen, the phonology becomes difficult to understand.

Therefore, it is conceivable to switch the phoneme of the human voice sound to be produced during the performance (in the middle of the song), but this requires operating controls such as a phoneme selection switch each time, which is cumbersome and difficult to perform. It has the disadvantage that it is not pleasant to listen to.

This invention was made in view of the above-mentioned drawbacks, and provides an electronic musical instrument that automatically switches the phonology of human voice sounds without the need for troublesome switch operations, making it possible to produce a wide variety of human voice sounds, and improving playability. The purpose is to provide 7 services.

To this end, the present invention generates phoneme instruction data that sequentially instructs the pronunciation of a plurality of phonemes, and switches the phonemes of human voice sounds to be pronounced in a crossfade manner in accordance with this data.

Hereinafter, the electronic musical instrument according to the present invention will be explained in detail using the drawings.

DESCRIPTION OF CONFIGURATION AND OPERATION OF THE INVENTION (1) Explanation of the Configuration FIG. 1 is a block diagram showing an embodiment of an electronic musical instrument according to the present invention.

In this embodiment, human voice sounds are formed using a harmonic synthesis method. In the figure, 1 is a key switch circuit provided in the keyboard section, which has a key switch corresponding to each key of the keyboard section. When a certain key is pressed, the corresponding key switch operates, and its output line is logic n
It outputs the key information KD of ", and also outputs a key-on signal KON indicating that any key has been pressed. 3 stores the frequency number R corresponding to the pitch of each key at each address. This frequency number memory 3 is a memory.
is addressed by the key information KD of the key switch circuit 1, and the frequency number R corresponding to the pitch of the pressed key is read from its output.

4 is a one-shot circuit that is triggered by the rising edge of the key-on signal KON and outputs a narrow one-shot pulse as the key-on pulse KONP; 5 is a clock oscillator that outputs a clock pulse Tc of a constant period; and 6 is a clock pulse Tc divided by W (w: A counter 7 outputs a calculation interval timing signal Tx (total number of harmonics to be synthesized at each sample point), and 7 counts clock pulses Tc and uses the output to obtain an order number n (total number of harmonics to be synthesized at each sample point)
It is a modulo (modulo) w counter that outputs n1 to Nw), and can also be used as the counter 6.

8 is a harmonic component generation circuit that generates each harmonic component including the fundamental wave based on the frequency number R output from the frequency number memory 3, and includes a calculation interval timing signal Tx.
A cumulative value QR (q=1, 2, 3... ...
) and a clock pulse T.
Accumulated value QR input through gate 8C every time c occurs
are added sequentially and the cumulative value NqR (n=1, 2, 3...
. . . W) as a signal indicating the phase of the n-th harmonic at each sample point.
d and the accumulated value NqR are decoded by the memory address decoder 8e, and the decoded output is used to determine the value corresponding to the accumulated value NqR among the amplitude values of each sample point of one cycle of the sine waveform stored at each address. A sine function memory 8f is provided in which sample point amplitude values are read out as π sine amplitude values SinwnqR of each harmonic.

9 multiplies the order number n output from the counter 7 and the frequency number R output from the frequency number memory 3, and the multiplied value NR indicates the frequency of each component of the high frequency corresponding to the pitch of the pressed key. This is a multiplier that outputs a frequency signal NR, and together with the counter 7 constitutes a frequency signal generation circuit 10.

Reference numeral 11 denotes an amplitude information generating circuit which outputs in parallel amplitude information FnR indicating the amplitude level of each of the seven formant frequencies of a plurality of vowels based on the frequency signal NR; in this embodiment, five types of vowels (a) , (1), (u), (e
), (0) amplitude information FnR(a), FnR(1
), FnR(u), FnR(e), and FnR(0) are output in parallel.

The amplitude information generating circuit 11 is composed of five amplitude information memories 11a to 11e corresponding to five types of vowels (a) to (0), as shown in the detailed circuit in FIG.
At each address of each of these memorandums 1a to 11e, the amplitude level of each formant frequency constituting each vowel is stored as a digital value, as shown in FIG. By being addressed by the frequency signal NR, the stored contents become amplitude information FnR(a), FnR(1).
, FnR(u), FnR(e), and FnR(0). 12 is a switching signal generating circuit that outputs a switching signal E as phoneme instruction data indicating the pronunciation order of a plurality of vowels; in this embodiment, when a plurality of vowels are successively pronounced one by one, a new In order to smoothly switch between the vowel to be pronounced and the vowel to be extinguished, the switching signal that specifies the pronunciation of the new vowel and the one that specifies the pronunciation of the vowel to be extinguished are simultaneously generated.
In the switching circuit 3 described later, the pronunciation of two types of vowels is smoothly switched based on desired crossfade information XF.

As an example, the switching signal generating circuit 12 generates the switching signal E (E1 to ElO) based on the count information of a counter that counts low frequency pulses at a predetermined period from the start of key depression, as shown in FIG. A circuit is used in which the pronunciation of successive vowels is switched at a predetermined period. The switching signal generating circuit 12 shown in FIG. 4 includes a variable frequency type low frequency pulse oscillator 12a that outputs a low frequency pulse of a predetermined period, and a low frequency pulse oscillator 12a that outputs a low frequency pulse of a predetermined period. A 5-ary counter (corresponding to 5 types of vowels) 12C that counts input low-frequency pulses, and this 5-ary counter 12C
a decoder 12d that decodes the count value of and outputs the decoded output as switching signals E1 to ElO, a switching vowel number setting switch 12e that sets how many types of vowels are to be switched at a predetermined period, and a quinary counter 12C.
When the count value matches the value set by the vowel number setting switch 12e, the AND gate 12b is deactivated,
An inverter 12f for stopping the counting operation of the quinary counter 12C and a key-on pulse KO at the start of key depression
Crossfade information generating circuit 1 of the switching circuit 13 whose OR signal of the output of the NP and the AND gate 12b will be described later.
OR gate 12 outputs as reset signal RS at 30
It is equipped with g. In this case, the switching signals E1 to ElO output from the decoder 12d are ``based on the same decode output'' for E1 and ElO, E2 and E6) E3 and E7) E4 and E8) E5 and E9, but new In order to distinguish between the vowel to be pronounced and the previous vowel to be extinguished, different signal names are used. for example,
When vowels (a) to (0) are to be pronounced in sequence, after vowel a has been pronounced by switching signal E1, when switching signals E2 and E6 become a logic n'' signal, the vowel is pronounced as a new vowel to be pronounced. The vowel (1) is specified, and the vowel (a) is also specified as the vowel to be extinguished.In this case, the number of vowels to be pronounced is determined by switching the vowel number setting switch 12.
It becomes as shown in FIG. 5 according to the set value by e. In other words, when the vowel number setting switch 12e is set to the setting value "1", only the vowel (a) is pronounced as shown in FIG. a)
is produced for a predetermined period of time (corresponding to the low frequency pulse period of the low frequency pulse oscillator 12a), then the vowel (1) is produced. 13 switches the amplitude information FnR(a) to FnR(0) output in parallel from the amplitude information generation circuit 11 based on the switching signal E (E1 to ElO) to generate the harmonics output from the harmonic component generation circuit 8. A switching circuit that outputs an amplitude coefficient CnR (V) for each component, the second
The detailed circuit diagram is shown in the figure.

In the same figure, 13a is a selection circuit that selects in what combination five types of vowels are to be sequentially pronounced, and is a selection circuit that selects amplitude information FnR that is output in parallel from the amplitude information generation circuit 11.
Five select switches 130a, 131a, 13 input each of (a) to FnR(0) in parallel, select one of them, and output it as selected vowel amplitude information D1 to D5.
2a, 133a, and 134a. 13b~1
3f is an AND gate that passes the selected vowel amplitude information D1 to D5 only when the corresponding switching signals E1 to E5 are input, and outputs the output as amplitude information DI of the vowel to be newly pronounced via the OR gate 13g. , 13h
.about.13m is an AND gate that allows selected vowel amplitude information D1 to D5 to pass only when the corresponding switching signals E6 to ElO are input, and outputs the output as amplitude information DD of the vowel to be extinguished via an OR gate 13n. These AND gates 13h to 13m are inhibited by the output of the flip-flop 13P when the first vowel is generated in response to a key depression.

130 smooths amplitude information DI of a vowel to be newly pronounced and amplitude information DD of a vowel to be extinguished, which are output from the OR gates 13g and ]3n, respectively, based on desired cross-fade information XF (fade-in information XFi, fade-out information XFO). A fade-out information memory 13, which is a cross-fade circuit for switching to the first address, stores fade-out information XFO whose value becomes smaller as it approaches the terminal address as shown in FIG. 6a.
10, a fade-in information memory 1320 storing fade-in information XFi whose value increases sequentially as it approaches the terminal address as shown in FIG. After being reset by the low frequency pulse oscillator 1330 and the set signal RS output from the switching signal generation circuit 12, the low frequency pulses inputted via the AND gate 1340 are counted and the count output is obtained. The fade out
I blue report memo l river 310 and fade-in information memory 1
Counter 1350 outputs as address signal to 320
, a multiplier 1360 that multiplies the fade-out information XFO and the amplitude information DD of the vowel to be extinguished, a multiplier 1370 that multiplies the fade-in information XFi and the amplitude information DI of the vowel to be newly pronounced; 13
At step 70, the amplitude information DI. DD and outputs the summed output as an amplitude coefficient CnR (V) for each harmonic component output from the harmonic component generation circuit 8, the fade-out information memo 1 river 3
When the counter 350 outputs a maximum value output M (logic n'') indicating the maximum value of the count value, this maximum value output M is inverted by the inverter 1390. , thereby disabling AND gate 1340.

As a result, the counter 1350 stops counting at the maximum count value, and starts counting again when the set signal RS is input again from the switching signal generating circuit 12. In this case, since the switching signal generation circuit 12 generates the set signal RS the number of times corresponding to the set number of pronounced vowels, the fade-out information 1310 and the fade-in information memorandum 320 generate the set signal RS the number of times corresponding to the set number of pronounced vowels. The fade-out information XFO and fade-in information XFi are read out, and the amplitude information DD. Crossfade information XF will be added to DI. 13P
is a flip-flop that is set by the key-on pulse KONP and outputs its inverted output Q to the AND gates 13h to 13m as an inhibit signal. This prohibits the vowel amplitude information DD from being output.

When the first vowel is generated, the flip-flop 13P registers the counter 13 of the crossfade circuit 130.
50 outputs the maximum direct output M, the one shot circuit 13 is triggered and outputs a one shot pulse of a predetermined width.
It is configured to be reset depending on the output of q. Next, returning to FIG. 1, 14 is an envelope waveform that outputs an envelope waveform signal ENV as shown in FIG. 7b when a key-on signal KON as shown in FIG. 7a is inputted with a key press operation. A generator 15 multiplies the amplitude coefficient CnR(V) outputted from the switching circuit 13 by the envelope waveform signal EN to obtain the amplitude coefficient CnR(V) as shown in FIG. 7b.
A multiplier that gives an amplitude envelope as shown in ]
6 is the amplitude coefficient C″NR(
) and each sup. output from the harmonic component generation circuit 8.
Sine amplitude value Sinvvn for each harmonic at the π point
qR to obtain the amplitude value of each harmonic Fn=C″NR(V
)・． A harmonic amplitude multiplier 17 outputs πSinwnqR, and a harmonic amplitude multiplier 17 calculates the amplitude value Fn of each harmonic output from the harmonic amplitude multiplier 16 for each sample point (calculation interval timing???・Willi: Kousei circuit, 18 is musical tone signal generation circuit 1
This is a sound system that generates musical tones based on musical tone signals output from 7.

(2) Description of operation of this embodiment In the electronic musical instrument configured as described above, the switching circuit 13 selects the vowels to be produced as combinations of (a), (1), (u), (e), and (0). The selector switches 130a to 134a in the selection circuit 13a (FIG. 2) are set, and the switching vowel number setting switch 12e (FIG. 4) in the switching signal generation circuit 12 is set to the setting value "5".
The operation when pronouncing the vowels (0) one by one will be explained.

First, when a key on the keyboard section is pressed under the above setting conditions, the key switch circuit 1 outputs the key information KD of logic "1" to the output line corresponding to the pressed key, and also outputs the key information KD of logic "1" to the output line corresponding to the pressed key. A key-on signal KON indicating that a key has been pressed is output.

When the key switch circuit outputs key information KD corresponding to the pressed key, the frequency number memory 3 is addressed by the key information KD, thereby reading out the frequency number R corresponding to the pitch of the pressed key. When the key switch circuit 1 outputs the key-on signal KON, the one-shot circuit 4 is triggered by the rising edge of the key-on signal KON and outputs a narrow one-shot pulse as the key-on pulse KONP. The frequency number R read out from the frequency number memory 3 is input to the harmonic component generation circuit 8,
As a result, the harmonic component generation circuit 8 generates each high frequency signal.
A sine amplitude value SinwnqR of the π harmonic (including the fundamental wave) is output.

On the other hand, the counter 7 of the frequency signal generating circuit 10 counts the clock pano response Tc and supplies the count output to the multiplier 9 as the order number n.

The multiplier 9 multiplies the frequency number R output from the frequency number memory 3 by the order number n output from the counter 7, and uses the multiplied value NR to calculate the frequency of each harmonic component corresponding to the pitch of the pressed key. The frequency shown is output as signal NR. This frequency signal NR is supplied as an address signal to the amplitude information memorandums 1a to 11e (FIG. 2) of the amplitude information generating circuit 11.
The amplitude level of each formant frequency of vowels (a) to (0) stored at the address corresponding to the frequency signal NR from each of 1e is the amplitude information FnR(a) to FnR(
0) in parallel. Amplitude information FnR(a) read out in parallel from this amplitude information generating circuit 11
~FnR(0) is input to the switching circuit 13, and the selection circuit 13a of the switching circuit 13 selects the select switch 130 in accordance with the combination of pronunciation vowels set in advance.
Selected vowel amplitude information D1 to D5 from a to 134a, respectively
is output as On the other hand, the switching signal generation circuit 12 (fourth
In the figure), after the quinary counter 12C is reset by the key-on pulse KONP output from the one shot circuit 4, the low frequency pulses from the low frequency pulse oscillator 12a are counted, and the count output is decoded by the decoder 12d. Switching signals E (E1 to ElO) are sequentially output.

In this case, the frequency signal N is
Since R changes at high speed, while the switching signal E is E1, the amplitude information FnR(a) of each formant frequency constituting the vowel a is obtained from the amplitude information memory 11a as FlR(a), F2
R(a), F3R(a)...FwR(a
) are read out repeatedly in sequence. Amplitude information FnR(a) (FlR(
a), F2R(a)...FwR(a))
is the AND gate 13b while the switching signal E indicates E1.
The signal passes through the OR gate 13g, becomes amplitude information DI, and is input to the crossfade circuit 130. Then, in the crossfade circuit 130, the amplitude information DI is given crossfade information XF as time passes, and is output from the adder 1380 as an amplitude coefficient CnR(V), that is, the harmonic component generation circuit 8. The sine amplitude value SinvvlqR, . π
．． πSinw:2qRc・・・・・・・・・
Amplitude coefficient ClR(V) for SinvvwqR, C2
R(V)......Output as CwR(V).

Therefore, when the sine amplitude value SinπWnqR of each harmonic is multiplied by the amplitude coefficient CnR (V) of the corresponding frequency in the harmonic amplitude multiplier 16, the amplitude of each formant frequency component constituting the vowel (a) is This means that the value Fn is set.

Therefore, the amplitude value Fn of each formant frequency component of the vowel (a) obtained in this way is accumulated in sequence in the musical sound signal generation circuit 17, and the accumulated value 1Y1Fn is converted into an analog signal and then supplied to the sound system 18. Then, the sound system]8 produces a musical sound of the vowel a. The sound system 18 continues to pronounce the vowel (a) while the switching signal E indicates E1. And the switching signal E is from E1! , a new vowel (
1) Amplitude information FnR of each formant frequency constituting
(1) (FlR(1), F2R(1)...
...FwR(1)) passes through the AND gate 13C, passes through the OR gate 13g, becomes amplitude information DI, and is input to the crossfade circuit 130. At the same time, the vowel (a
) is the amplitude information FnR(
A) becomes the amplitude information DD of the vowel to be eliminated and is output from the OR gate 13n. The amplitude information DI corresponding to vowel (1) and the amplitude information DD corresponding to vowel a are smoothly switched in the cross-fade circuit 130 under the control of fade-in information XFi and fade-out information XFO, respectively. As a result, the crossfade circuit 130 outputs vowel (1) and vowel (a).
Amplitude information FnR(1
), FnR(a) is smoothly switched and the amplitude coefficient C
It is output as nR(V). Therefore, during the time period when the amplitude coefficient CnR(V) is smoothly switched, the amplitude levels of each formant frequency constituting the two vowels, vowel (1) and vowel (a), are output in duplicate. Therefore, after giving the amplitude envelope by the envelope waveform signal ENV to the amplitude coefficient CnR (V) in this time period, the amplitude is calculated by multiplying by the sine amplitude value Sin QR of each harmonic corresponding to π. When the value Fn is set and a musical tone signal is generated based on this amplitude value Fn, the generated musical tone is a diphthong of vowels (a) and (1).

This kind of action is the vowel (1) and (u), (u) and (e
), (e) and (0), and the sound system 18 outputs 5 in the manner shown in FIG. 5V.
The types of vowels (a) to (0) are sequentially switched one by one, and each switch is smoothly performed and pronounced. In this case, the pitches of vowels (a) to (0) correspond to the pitches of the same key pressed, and therefore, if keys of different pitches are pressed consecutively, the pitches of the vowels (a) to (0) correspond to the pitches of the same key pressed. Vowels (a) to (0) of different pitches are successively pronounced. Therefore, the select switches 130a to 130a in the switching circuit 13
] By properly selecting the vowels using the vowel 34a and appropriately setting the number of vowels to be switched in the switching signal generating circuit 12, humming and skitting performances using diphthongs can be performed. Effects of the Invention As explained above, the electronic musical instrument according to the present invention generates phoneme instruction data that sequentially instructs the pronunciation of each of a plurality of phonemes,
The phoneme of the human voice to be produced is switched according to this data.

Second, it is possible to automatically switch the phoneme of a human voice sound without a troublesome switch operation, and a wide variety of human voice sounds can be produced, which has the effect of improving playability.
In addition, in this invention, when switching the phoneme of a human voice,
Since the amplitude level of the phoneme to be extinguished is gradually decreased and the amplitude level of the phoneme to be newly pronounced is gradually increased, the phoneme can be switched smoothly.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing one embodiment of an electronic musical instrument according to the present invention, FIG. 2 is a circuit diagram showing an example of the amplitude information generating circuit and switching circuit of FIG. 1, and FIG. 3 is a block diagram showing an example of the amplitude information generating circuit and switching circuit of FIG. A diagram showing an example of formant characteristics of a vowel stored in the amplitude information memory of the generation circuit, FIG. 4 is a circuit diagram showing an example of the switching signal generation circuit of FIG. 1, and FIG. Figure 6 is a diagram for explaining the operation of the signal generation circuit.
FIG. 7 is a waveform diagram showing an example of the envelope signal EN. 1... Key switch circuit, 3... Frequency number memory, 8... Harmonic component generation circuit,
10... Frequency signal generation circuit, 11...
- Amplitude information generation circuit, 12... Switching signal generation circuit, 13... Switching circuit, 16... Harmonic amplitude multiplier, 17... Musical tone signal generation circuit.

Claims (1)

[Scope of Claims] 1. In an electronic musical instrument configured to form and produce musical tones with fixed formant characteristics corresponding to the human voice of a desired phoneme, each of a plurality of phonemes is generated according to the elapse of time from the time when the musical tones are produced. A phoneme instruction data generating means for sequentially generating phoneme instruction data to instruct, and controlling the phoneme of the musical tone in accordance with the content of the phoneme instruction data generated from the phoneme instruction data generation means, and When the instruction data changes from the first data to the second data, the amplitude level of the musical tone of the phoneme corresponding to the content of the first data is gradually decreased, and the amplitude level of the musical tone corresponding to the content of the second data is gradually decreased. An electronic musical instrument is provided with musical tone control means for controlling the amplitude level of musical tones of phonemes to be gradually increased, and the phonemes of human voice sounds to be produced are automatically and sequentially switched over time. 2. The electronic musical instrument according to claim 1, wherein the phoneme instruction data generating means includes means for selecting and setting the number of phonemes specified by the phoneme instruction data. 3. The electronic musical instrument according to claim 1, wherein the musical tone control means includes means for setting a pronunciation order of a plurality of phonemes.