JPS58184196A - Electronic musical instrument - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] This invention generates human voice sounds! 1 Electronic Toki KI41, % generates a human voice, that is, a human voice sound, in the 6° enemy near electronic Qin system related to the electronic Qin # that was generated in % and changed the phoneme of the 6 human voice sounds to a kosei measure! Some attempts have been made to use it as a pack chorus for a song. This electronic Qin generates human voice sounds by forming a defined phono-lumant filter using the noodle method and breaking it down. It is set by selecting it as you like.

By the way, when you listen to a human voice with a tone of ``Tou'' in I/IaO, it feels like a single sound, and the sound tends to become clearer as you listen to it.Therefore, it is recommended to switch the phonology in the middle of the song. Therefore, K has to operate the phoneme selection switch each time, and the operation becomes extremely troublesome, which is not surprising when playing electronic appliances.

This invention is based on the above-mentioned point K, and is designed so that human voice sounds can be played independently for each predetermined maki ball.

This light ##rr theory will be explained below based on the drawings.

FIG. 1 is a basic block diagram showing a single hook of an electronic musical instrument according to the present invention. Jll board & l is analogy #IA
Board UK, below il! It has a Nishiki'LK, pedal keyboard PK, and 6 key operations on each keyboard UK, LK, and PK are detected by key press detection Ml! +2 detected. The key press detection unit 2 is &! Detects the key being pressed in the [board section] and outputs the key information (hereinafter, represented by the key code IC) representing the pressed key.・Pronunciation assignment? ! 1hm is for assigning each pressed key, which is added from the hS detection section 2 and suppressed in the wrinkled section 1 based on the key code KCK, to one of a plurality of musical tone dulling channels f, and -g to each channel. A key code KC indicating the hit key and a key-on signal KON indicating whether the suppression of the WX key continues or has been released are output in a temporal manner in synchronization with the channel timing of each channel.

Here, the enemy code KC belongs to the pressed key -r61
! 2-pit hardboard code IIL indicating the board (LK or PK on υ), K1 and octa-7i of the pressure key! A 3-bit octave code indicating the I range Bm+B! ,B,
and a 4-bit note code indicating the note name N4 s Nm
This is 9-bit data consisting of I Nm # NB. Further, the key-on signal KON is a signal 6141'l' when the key press tp continues, and becomes sOa when the press is released tt6 (stored).

The key code kc and key-on ＠KON outputted from the sound generation assignment hall 3 are supplied to the -11mine sound signal form 叡酡4 and the make sound detection unit. 3 to 6 key codes KC and key-on bias @KO
In response to the NK response, musical tone signals corresponding to *'* are applied to each channel and outputted to the downstream & section 8, which mixes the formed musical tone signals of each channel as appropriate. Also, vocal sound detection & SS is ゛, keyboard tI
It detects the human voice sound (Baumol sound) generated from the key being pressed by Al, and is currently being suppressed based on the key code KC and key-on signal KONK given temporally by the pronunciation assignment section 3. A predetermined 1 from the hardness of
or digit number plow is detected, and a vocal sound key code VKC and a vocal sound key-on signal VKON are outputted in accordance with the detected key and supplied to the vocal Toji sound signal Jijf forming section 6. This vocal sound detection s5 is configured as shown in FIG. 2, for example.

In the implementation of the vocal sound detection station 5 shown in FIG.
Among the keys currently being suppressed on a predetermined keyboard (UK, LK, or PK), a single key corresponding to the highest i-pressed key is detected, and the vocal sound key hood VKC is detected in response to the highest pressing pressure #IKtI#. and a vocal sound key-on signal VKON is output.

211th! In J, the keyboard code Kglg of the key code KC4a of each channel and the key-on signal WON given time-wise from the sound generation-applying section 3 is input to the decoder 51, and the octave code Bi%B
! , note code N4-N.

The 2F key-on signal KON is manually input to the gate 52. The decoder 51 decodes the keyboard codes ic, 1, and 1, and outputs the upper folded keyboard number υ, ``'F keyboard signal VL, and pedal key number P, which indicates the keyboard to which the key assigned to each channel belongs. Output these fB@u, L, P
are respectively supplied to one input terminal of the AND gate A++4*Am. Ant game) As w Muyuki,
The other input terminal of system 1 is the upper keyboard★selection switch BUs, the lower keyboard selection switch SL%, and the header keyboard backward selection switch 8P.
Each output signal of is added respectively.゛In this case,
= Selection switch 8u * 8L t' 8p is a switch that selects from among the pressed keys of the keyboard when detecting the highest pressed key in this vocal sound detection 115. Switch 8t+ is turned on, and the output signal of switch 8U becomes elm, which enables AND gate A1 to operate, and outputs 6 upper keyboard signal U from decoder U, enabling gate 52 via OR gate G1. In addition to terminal E, gate 52 is rendered conductive. Therefore, in this case, when the lower keyboard IJK riddle is assigned and the key code KC and key-on signal @KON of 6 channels are given, the decoder 51 sends the upper and 11 keys at the channel timing of this channel.
11 signal U ('1' signal) is output,
The gate 52 becomes conductive, and the octave code Bs""B1, note code N, and 8N of the channel.

and the key-on signal KONt- is passed. That is,
Octave code 83 for upper keyboard UK channel
~B1, No) Code N4~N Otori and key-on signal K
Only ON is selected by gate 52 and the highest tone is detected - path 53
added to. Also, when the selection switch SL or 8P is turned on, the keys of the lower keyboard LK or pedal @@PK are assigned in the same way as above, and the 6-channel octave chord B, ~B, % note is assigned. Code N4
~N1 and the key-on signal KON are selected by the game F52 and added to the highest sound test path 53, and the selection switch 8u e IL * @p selects IIl! When the switch is turned on to m-, each am channel corresponding to the turned on switch is kilc chest 6 otatarp code 1. A-B,, note code N4-Nl
, key-on KON or all selection 2 and the highest sound detection-path saK can be added.

Forged treble detection @Route 53, octave chord Bm ``''' Ill, note chord) added from gate 58! 4~NS and Imaichi ON signal KO) i (ill selection switch 8u e 8L ## board #CsI corresponding to the switch that is turned on) #L with the leg under pregnancy suppression Octave chords F Bi to Hatake 8 and note codes N4 to N1 corresponding to the highest pressed key are detected. This detection 6 added octave code BB~ml
%Nonito Code N4~Teeth [The channel timing of each channel is performed by performing a 6R fine touch ratio once, and the octave code B of the highest pressed key detected as a result of the comparison, ~1. The % note codes N4 to N1 are stored at the next channel timing 1 cycle and output as the vocal amount middle code VKC. In addition, in the highest pitch detection-m53, the highest pitch #sho-detection KM mentioned above
Continuously, a vocal sound key-on signal VKON is formed, and this signal vxoh becomes 1 bit indicating a key depression when the highest pitch suppression is detected and a vocal sound key code vKC having some essence other than rOJ is output. , the highest pressed key is no longer detected and the content of the wolf key code VKC is 1.
When it becomes 0'', it becomes 10' which indicates -art.

Note that when the highest note pressed key detected by this vocal sound key-on signal VKON U is changed (that is, when the content of the key code VKC is changed), K is set to 8 during the predetermined study period.
10# and 6゜This is the key-on signal VKON t-chi time U #O-.The vocal Hata tone signal generated in the 6-vocal Hata tone signal image weight 6, which will be described later from K, is pressed by the old highest note. This is in order to make the transition clear when switching from the one that corresponds to the key pressed to the one that corresponds to the new highest-pressing key. The code VKC and the E-KAL sound source-on-Yu code VXO14 are supplied to the vocal Qin sound signal engraving IE section 6.

On the other hand, the −1 notification occurrence noodle 7 is 1 in the preset measure 19.
It outputs the control signal P. 1. In the implementation shown in FIG. 3, whose results are shown in FIG.
For example, Suzuran 111111 corresponds to the length of an s half note.
) pulse train counter? 2 makes it the first point, the second point) - 1 6 m corresponding to the time length of t J
When the count reaches 72, the bar pulse L is outputted from the count 72. This bar pulse L is counted by the carantuff 3.Meanwhile, the music department arbitrarily sets the bar 1lJWk to be changed in advance (for example, every 2 bars). I want to switch the sound I#I
(2 measures are set for each session).

At #75, the actual small tt1 sheet output from the counter 13 and the set small l1lb are in the ratio IIR.
and when it becomes Mo-B, W from that EQI11i1 child
# Koji signal is output. In addition, in this example,
From Tenboku Tsuta &! l#711! II board s1 suppression #1
It may be started after one cycle of operation.

Vocal music message @Kage wo part 6 is vocal sound key code V
Based on KC and the vocal sound key-on signal VKON K, a musical tone signal corresponding to a predetermined human voice tone is formed and output, and the six musical tone signals to be formed are switched by the 1^ signal PK given from the klIiI tone generation signal. . In Gikal Sound Collection @ Form 6, l1il-signal generation body 7
Kara-餉 - 4Mt of corresponding musical tones in a predetermined order or f! :, occurs in the order of meaning. Therefore, as the tl back progresses on the keyboard 1, the phonology of the i number formed in this Hata Onga 8 formation lic changes every predetermined number of measures. Signal forming section 6
The tone ^ (pitch) of the musical tone signal formed by Imaichi code V
The highest pitch suppression shown by KC! It corresponds to IKm.

The human voice musical tone signal outputted from the vocal musical tone signal type 11 is supplied with the musical tone signal corresponding to each restraint mK outputted from the general musical tone signal generator 4, and then sent to an amplifier, a speaker, etc. These sounds are manually produced by the volume Rand System 9, which consists of the following, and are pronounced as musical tones.

In this way, the performance on the keyboard s1 progresses.K is a predetermined small l.
Ela in which the sound of the human voice changes automatically every 1 WI Next, we will explain the specific preparation and operation of the vocal Qin Yin Yugo W& part 6.

114- is Vukal Qin Shiyu number form ``Weibe 6-tube Dijigaby''
<ka&mon) Tll1iL□. 1゜□ cut! 11111
M+-Road 61 is the 1-signal Pfl from Tokyo 7 where the signal is generated.
The address signal generator 62 and the selector 64 are supplied with the six children's selection code 8 which appropriately translates the sound of the human voice. This sound is off! -The control circuit 61 is shown in Fig. 5&)K
As shown, an M-adic counter 6 that converts the control signal P
11, a decoder 61b for decoding the Callan F value of this counter 61a, and K, or a decoder 61b for decoding the Callan F value of this counter 61a,
Random number generation @61c that generates the numbers "1" to rMJ in any order as shown in , and random number data generated from random number generation @61c when receiving the 1lII-signal P [latch rough 'tb path 61d] , and a decoder 61e that decodes the random number data latched and output by the latch 1 circuit 61d. According to the phoneme switching koji-making route 61 having the configuration shown in Figure 1-1), the output value of the counter 61m increases every time the performance progresses by a predetermined number of measures, and the output value corresponds to the output value. The sounds 111i111 selection signal 8 are output in a predetermined order (for example, the touch of A, Tsu, O, U, Ru, Wa), and when the sound 1lIII11 selection signal 8 changes by the predetermined number of sounds -, the first phoneme is output again. Returns to appropriate translation signal 8 and repeats the same pattern from then on ・For this, figure N (b) K
In the phoneme switching of structure 1 shown - Koji Ichiro 61, the predetermined small IIb
For each number, the random number generator 61c outputs a child's choice @8 corresponding to the randomly generated random number Vt%i. Therefore, in this case, the working order of the material is completely chaotic, and the 6° address signal generating section 62 outputs this sound from the sound V8 and the vocal sound detection unit, and generates the 6-cal sound key-on signal. Based on VION, @ number memory 5spc is stored for each sound! Atto 1 for reading 1111 number data
Sugagomunit output 16゜This address signal generation capital 62
This is shown in Figure 61, and its work will be explained with reference to Figure #47 and Figure 8 KIh.

In Kame@-, phoneme-kiri! 1IllI Goro 61 to 7 dress Gage - Mushibe 62 receives a human power 1 from 1IllI Goro 61, and an attack pulse is generated from 1 & il bribe 1G, an attack pulse AP with a human voice rising shape 11t*-1 or 1 also decays. Pulse JjIhI! The sound of human voices coming from E-Road 11
The decay pulse DP that regulates the lower 6-state 1 specifies the address from which the coefficient take for the continuous part of the human voice is read from the coefficient memory 6s from the rough + stain address memory 12. are generated respectively. In this case, the attack pulse period P and the decay pulse period DPfJ are set according to the phoneme indicated by the phoneme selection signal 8, and the attack pulse generation Ig circuit 1G and the decay pulse generation Ig1wI
11, for example, the number of oscillations is fIll by Minetaka @S.
It is composed of six hired members, Midori. Further, the sustain address memory 12 stores the above-mentioned sustain address memory 8 at each address corresponding to each note.
The remaining address A1 corresponding to the phoneme indicated by the signal S is read out from the input K as an address signal.

Sate, W-Ak child detection f14htPuvlS7WAff
) K-like gong-cal sound key-on 4II issue VKO
)! When address g1@ is input to the generating section 621C, its rising edge is differentiated by the differential circuit 13 and output as the cow-on pulse KNP. This key-on pulse KNP resets the (1 day) advance counter 11 and KR8
7 rip 70 g 15 is set. As a result, the AND condition of Antogame 4 is established, and the attack pulse AP or Antogame 4, ORGATE G! , the counter 17 starts counting the attack pulses P. At this time, since the ant game clock 16 is set as described below, the AND condition is not satisfied and the decay pulse Dr is not supplied to the counter 1.

FIG. 8 shows the temporal keying of the count value of the counter 1, where the counter 17 counts attack pulses P and sends the count value to the coefficient menu 63 as an address signal DIE. When the count value of the output 6° counter 17 becomes equal to the sustain address memo 9'' 12, the signal IQb1 is output from the comparison 7ts, and the R8 preamp 7 4-tube 15 is reset. As a result, the AND condition of the antgame 4 no longer stands at 1, the completed pulse sum P is not supplied to the counter 17, and the count value of the counter 17 remains stopped at the constant pulse sum 1. This count value (M1) continues to be output as the address signal MDR. This state continues until the vocal sound key-on signal VKON falls. ゛When the vocal sound key-on signal VKON falls, the differential i path 14 Detecting the falling edge of the key-off pulse KFPtjlJT6゜The result R
Ant game because 87 Ritsugu Hoop 1 @ is set.
Tom S's AND condition is set to 1, and the 6 day pulse DI' generated from the 6 day pulse generation path 11 becomes the counter 1'.
tK is input. From the stiffness K, the counter 17 performs a counting operation of the decay pulse Dr from the above-mentioned count value (11). The count value of this chest counter 17 is also sent to the coefficient memory 63 as the address code DR.60 When the count value of the chest counter 17 reaches the maximum value N (all bits are 111), the The condition is met, and the RB7 lip 7th knob 16 is set to 6°.As a result, the counter 17 of the Decay Pulse Dr stops, and the count value of the counter 17 becomes the maximum value N.
The count value N continues to be sent to the coefficient memory 8 as the address signal ADR. When you return to the parent's office, the λμ sound & is synthesized with the unvoiced sound and the voiceless sound, and the unvoiced sound is synthesized. Since the filter is made from scratch, 41kWk memory 63 is required for that purpose.
v- is generated as a human voice sound and one sound - (analog number 1
"a", "tsu"..."ru", "wa", etc.) musical tones [Na coefficients of voiced and voiceless sounds necessary to make shadow 1,
Ol! 6 Parcor (Parcor)
Coefficients and notes for each note 96 l a m 115
m b...K is stored in advance.・Then, from the address signal generation capital 62, kT&! When the address signal MDR as shown in 3N) is output, the memory corresponding to each phoneme of coefficient letters 9 and 3 is stored.
・Ato & ＾Read out sequentially according to DR, selector @4
The person and the fish are each fed to the sweetfish. selector 64
Then, the gklk message 9 also selects the phonetic sound from the sound switching Ib#- path 61 among the silent vibration intermission, voice vibration S coefficient, and Percoll coefficient that are output corresponding to each sound 1 from 3. Selects and outputs only the modulus 1 [(voiceless nawa gIkWk1 voiced furiwa *&, percoll coefficient) corresponding to the predetermined phoneme based on the t8. If it is a voiced sound such as the sound -, the unvoiced vibration coefficient is always set to zero, and if it is a phoneme such as "ba" or "yu" in which Km% sound is mixed at the beginning of the sound, The number of vocal oscillations increases slowly with time as shown in Figure 7 (C), and the number of silent oscillations increases slowly with time as shown in Figure 7 (C), while the number of silent oscillations increases during the first rise as shown in Figure 7 (C). K112 is set so that the beater sounds sharply by the leap and then becomes zero. Note that the Percoll coefficient is determined by KW & so that it changes over time in the rising and falling parts of the sound as shown in xh(e)K1.

Among the 6'4i voices and unvoiced sounds that form the human voice, the unvoiced signal is 10 blue.The noise signal output from the noise source 65 is multiplied by the unvoiced ring image number output from the selector 64. Multiply by 6Sg to create 1. On the other hand, the voice signal is created as follows. First, an example in which the wave number F (constant &) corresponding to the 6 keys of 1 is stored in each address in advance is made up of 11 ROMKs. From the frequency information memory 67, read out the 6M1 wave number Fill, which is written in the address marked &* from the l-cal sound key code VKCK.
The j# wave number r continuously retrieved from the frequency information memory 67 is sequentially accumulated at fixed time intervals in the achievator 6I, and the accumulated value is outputted as an address signal of the sound-wrinkle shadow memory 69 □・□・. The waveform memory 69 contains 61rR@ values at each point of the religion, which is obtained by dividing the six fundamental wave shadows (for example, the first period of an asymmetrical triangular wave) that influence the Southern tone into one point of religion. is stored in each address as a digital data, and this amplitude value is stored in the accumulator 6.
Based on the address signal output from I, the next address is output. The rain sound source signal extracted from the sound source waveform memory 69 is output from the selector 64 at *m*go, and is multiplied by the ps vibration coefficient to become a sound signal. is added to the unvoiced sound signal outputted from the multiplier, . Digital filter 82
For example, as shown in Special Issue No. 56-1257941, the phonological 7-or! The point is multiplied by several wrinkles, addition is carried out, and the filter is filtered from a IIIK filter made up of memory and K, and the filter characteristics are determined by &lii from the percoll ill from the selector 6 to produce a predetermined sound of 11sK.
b0 that forms a corresponding human voice sound signal (vocal musical sound signal)
The digital human voice signal outputted from the digital filter 82 is converted into an analog mechanical musical sound signal by a D7M transformation 83.

In this way, the musical tone signal generated from the optical signal 6 is
-The Cal Qin sound signal is mixed with the sound collection signal generated from the general Qin beginning unit @formation hall 4 in the mixing section 8, as shown in 1-, and is amplified and output from the sound system 9! Figure 1゜tM9 is a trout that represents the main melody of a song. Generally, there are many 2 measures as 7 rays or movements in one song, so
- #Gase Kabuto 7's small 1 test - 2 bars at ro 14! -If you set it, every two measures from ****飜藶＠
P is output and 6° vocal tone signal shadow 1 part 60 section switching iiIm-path 61 receives this 1-signal P[, as explained in tt'g about the continuation that causes neck envy. As you progress to the 2nd grade, the sound of people changes in the 9th figure.
In a fruitful example of this invention shown in sl-, 1I11i
Using the output of the i41 number generation counter 72 (see Figure 3), a 1-year-old rhythm pattern OJ rhythm five-note m4 # is formed, and the rhythm sound source signal of the rhythm pattern whose performance idea has been Kanazawa is sent to the mixer sK and It is also possible to mix it with a general musical tone signal and a human voice signal and use it as a rhythm accompaniment tone. That is, the tenbotasic oscillator 71 in FIG. cow〉ri7
2 may be shared with the automatic rhythm playing device.

FIGS. 12 and 13 show other embodiments of the primary musical tone signal formation Wb7.

Figure 10 shows an example of the implementation of the analog filter using an analog filter. l+I corresponding to each sound 911
Human voice sounds are processed using IJkl. This *** example generates human voice sounds through analog processing, and for this purpose it is very difficult to use the digital filter shown in Figure 4. Voltage l1 instead of 6 multiplication 166.80 in
III WiJ Variable Gain Increaser (VCA) Theft 1 and 2
2, the analog filter 24 has a center frequency fl
e ft + fs-characteristics are fll"' + fl
Bandpass 74 with Fi4 wave number characteristics as shown in Fig. 11 of l *' fall k # jlPF+ * BP
The coefficient memory 6s also stores 7 ormant characteristics corresponding to each sound ("a", [tsu J... "ru", "wa"). Bandpass filter IPFI e required for Kaiisu 6
BPFI v Center frequency fl of each bPFg #
fje ,, f crow and resonance characteristics fil +1mm +
flB, the 1st voice coefficient and the %less vibration image are stored as analog values, and the sound output from the child-switching-ma path 61 is stored as an analog value. - Appropriate translation signal vs. vocal sound output from corridor 5 and 6 vocal sound key-on V
When the address signal DR is output from the address signal generator s62 for KON and KM, the center-wave number fl m b + ts for each tone is output from the coefficient memory 63, both of which are 1M.
Characteristics fll sfo.

Heavy..., voice vibration coefficient, and silent vibration coefficient analysis
1"・1: ν date evening or company 2 selector 64 human power terminal 1-M
K is input. The selector 64 properly translates the above-mentioned analog four-digital sound from the sound-switching-II1gl path 61゛, which is a predetermined sound based on the appropriate translation sou 88, and outputs the output t6゜ik noodle color code generation Tsumugiji 20 is 餑. It consists of a frequency information memory 67, an accumulator 68, and a sound source waveform memory 69'' that operate according to the implementation shown in Figure 4.
The six waveforms recorded in the source waveform memory 69 are processed as analog values and stored at six points. Therefore, the sound source signal generator 1 20 outputs a five-point analog source signal to the focal sound key hood VKC from the horn sound output s5. This voiced sound source signal is turned into a sound signal by the VCA 21 which is turned on based on the voice signal from the selector 64. On the other hand, the unvoiced sound source signal generated by the analog-configured noise sound -65 is also changed by -1!1!41 by the VCC Muk 2, which is gain-controlled based on the unvoiced modulation fiM coefficient from the selector 64, and becomes an unvoiced sound signal. Become. The % sound 4m and unvoiced automatic signals obtained in this way are added by addition -23, and are outputted from the filter characteristic number coselector 64 of the bandpass filter BPFm, which is supplied to the analog filter 24 and is supplied to the analog filter 24. The filter characteristics are changed by an analog component that adds a self-centered density f1 and a resonance characteristic tms, and similarly a bandpass filter PF.

And the filter characteristics of BPFi are the center frequencies f, , f, and m vibration characteristics from the selector 64 '11*tll
111 m2 by setting the Ajilogeta, and these three band pass filters BPFs e Bpy',
l BPFI provides ti7x luminant characteristics as shown in FIG. The center of t output from the selector 64 - t
Since the analog data of the ll number and resonance characteristics change in relation to #l, the 7-ormant characteristics of the analog filter 24 also change over time, resulting in a bird-quality vocal Hata tone signal.

&JIll signal generation signal may be m-4B per predetermined small detail
When @P is output, the selector 64 selects the sound based on the %P -1 selection signal 8 or the sound indicating the -16 center mm.
By overselecting the coefficient data of number, resonance characteristics, sound, and h%1&- coefficient, and controlling the gain of the vCM 21, 22 and the filter characteristics of the analog filter 24, ”-Kal Toon signal is formed.

The analog filter 24, which is similar to the above-mentioned fruit and hook, is composed of three bandpass filters TIPF@ + BPFI + BP.
It is a hook in which Fl is connected to the ik column, but even if they are connected in parallel to class L, and the filter characteristics of each bandpass filter and the level of their output are controlled for each stoichiometry, the vocal musical sound signal cannot be processed. can be formed (for example, special-defect 5b-t7t**% public participation) 0 Plow 12 is a vocal sound collection (
s! In this embodiment, a formation hall 6 is constructed, and this implementation uses a high l11iIE oscillator generation IP+ path to imitate a human voice sound 6, and adjusts &-° of tltm to r in response to a predetermined sound. #j
For example, special #JB5
No. 5-21063.

The chest tIkk information memory 30 has a key 11 and a key 1, similar to the Ml & number memory 67 in the fruit hook shown in No. 4 (2).
The frequency number F corresponding to the note #F is stored, and when addressed by the key code VKCK from the vocal sound detection section, the frequency number F corresponding to the highest note #F is displayed. O^-wave component generation path 31 where F is read out
is the 6M1 wave number output from the wave number f#-memory 30? Based on the sinusoidal vibration of each harmonic - aqF (m
m1.2=3-v%q-xw2sm) is generated in hours, minutes-, and multiplied by 6 degrees. 1ikfI
The I report memory 30 is also read out and 6 chest wave number F is read out.
#! The multiplication value -F is output as a frequency Wk signal signal r indicating the pigeon wave of each harmonic component.

&111tm&occurrence-m 33! J multiplication −32 to I
B force; are 61' Kojiba & scratch QaFKm and the predetermined sound - (praise h\shita - "a", "tsu", ... "ru"
, "wa") corresponding to each of the 7 ormant frequencies 1.
'3 a 133 b *... output in parallel from 6o, that is, each of these memories 38as33b@...
are stored in the vibration level or digital value of each wave number in the 7 ormant characteristic corresponding to a predetermined society, and from the multiplication -32, When 1 address is received, Sorikikienaiharu is read out.On the other hand, Onshokiri Ik
The aIil1gi path 34 outputs the sound-1 selection signal 8 when the signal P is output from the 1-偲-go raw noodles 7 in the same way as the sound-kiri-dan-1lIIl-path 61 in the embodiment shown in Figure 4. do. Selector 35 is island 4#! The same function as the selector 64 of J is carried out, and each sound outputted in parallel from the vibration information transmitter Tsumugiji 33 -Kllsu6&-of the information -switch&
Based on the sound from the -1 path 34, a predetermined transmission signal is selected based on the appropriate translation signal 8, and is outputted from the high-wave component generation path 31 as the vibration coefficient for one wave and one minute. Output. - etc.
Is the child & proper translation signal S fake 1 degree bellow 1 waveform? It is also supplied to route 36, and this envelope 1 wave carving occurs.
At step 36, a vocal sound key-on signal VKON is manually generated from the vocal amount detection s5, and an envelope M shadow signal corresponding to the child indicated by the i*st Henzawa Kayo S signal is generated. The envelope waveform signal includes the rise, sustain level, and fall of the waveform for each note. The vibration coefficient output from the selector 35 and the envelope waveform signal generated from the envelope waveform generation island 36 are multiplied by Higashi N & 37, and the envelope characteristic is given to the valence number.
In this way, Furi-Envelope-7, was given 2 Furi S! ill
Number and 1IA-Wrinkle 1 minute dulling-Roron・The sine vibration value of each ^ frequency component outputted from th; nqF is High-Hanawa multiplication 9
38 and output as a signal for each high II wave. Musical tone signal form - path 39 accumulates the vibrations of each harmonic wave output from the high frequency amplitude multiplier @Sa for each point, and converts the accumulated value to the corresponding ryo naag@. Then, this anada signal is output as an l-cal musical tone signal.
The structure and its operation are discussed in %m1fi5ft-210
No. 113 public INK is shown, so I omit the sparrow here.The 13th is an example in which the Vocal Eion signal carving 1 study was made by the crane 1 using the Pengha taihen chest method, and this is the fruit.

For example, m- is 41111i111i65-18623
As shown in the issue of CS, the 17&- wave hook method of 1211i11 is shown. VKC or human power is read b, and the number FtIK is read out and 6o,
In addition to the above-mentioned IHk hook, the m-signal PK from the IIItJP signal slows down, and the phoneme switching signal 8 is output from the phoneme switching path 44. 7 Ormant information memory 46 stores each sound rl1M(r
A J, r') J... "Wa"etc.)'s 7 Orman) characteristic as K11ll data as the center of 74 Le Manto 11 wave number KtJI 46 Theta Rc, 7 Cheetham to the level of Ormant (RI and 7 or!nt's wide bl (@) KM, Richi I IO) These data are read from the address signal DRK, and the selector 4 is inputted to the selector 4, 0, and 1 to MK. Pieces (071 to -q>) Shadow!-Road PCB
, FCt −・= FC.

Yanagi 1 is lower than K - respectively the above-mentioned 7 ormant-shaped intermittent data RCs (1),! (1) and frequency k +
The formant component signal is 1e1 and output by performing a constant calculation. Note that the interval between %7 Olman) formation-step νC1~FC is 16 to be realized.
It is only necessary to provide 7 ormans of phonemes; for example, if you are trying to synthesize human voice sounds, you only need three at most.

In addition, each 7 ormanshi ha kil k FC1~F
The variant structure 1 of C11 and its operation are described in feature 185 above.
Since it is shown in the details of Toyumi Publication No. 8623 K11P, the parent here will be omitted.

In this way, 7 ors each! 11g1l road FC, ~FC,
The forma signal 1 minute signal formed by 1 is multiplied by 1l by addition -42 and becomes a digital signal. The sound collection 414@ is wrinkled and a vocal roar signal of the corresponding phoneme is generated.
The waveform of the human voice sound of each of the above 1i11I methods can be stored in the tii#* type memory and used as wthtb-i. In this case,
The method for storing the waveform of a human voice in the waveform memo 9 is P.
CM (pulse code changing), APCM (a response pulse x code changing), DPCM (differential pulse code changing), mu DP
There are various methods such as CM (adaptive differential pulse code tm), DM (delta encyclopedia), and LPCM (linear pulse code production), any of which may be used. Roughly speaking, in such a storage method, each '11! In particular, by providing a waveform memory that stores the human voice sound waveforms described above, natural human voice sounds can be obtained. Although I have only set the ``l'' corresponding to the ``sound control'' key, instead of the ``I'' and ``I'' keys, instead of the ``I'' and ``I'' keys, you can also set the key to produce a human voice sound corresponding to the ``1'' key. (or more than one 1i6)K
Correspondingly, the human voice may be pronounced as 16, or the range of the human voice to be pronounced may be different from the range of the pressed key (for example, the human voice may be one octave above the range of the pressed key). or in a lower register).

SaraK also detects the eel rice (temporary key depression) that is pressed on the noodle board N, 1, and detects the phoneme of human voice sounds and the sound 1tll.
You may also use K to do lll[Il. Zara K also, top l
In eitkm, the general musical tone signal forming section 4tjI may be omitted and only human voice sounds are produced.
The seller timing of 7P) may be within K11l plus l for every δr fixed work hours. -rM11
i4I- and 1-signal P [good enough to be generated.

It's Sumire on the 8th floor, this is from the #S board in the msi gallery or the first tier.
Even in 6-electron Qin & r7D6
゜As mentioned above, in this case, the sound of the 6 human voices produced in 11 was automatically changed to 1b every time (predetermined small noodles change), so the troublesome switch I11 was used. [You can automatically wrinkle the human voice sounds and phonology,
6゜4・wim's simple theory of 6゜4・wim* Work 1 is a summary of Denshi East i based on this work. - 1 Fushi block - 1, Rattan 2 - This is the number 1 in the 6 electronic telephone number according to this call - One part of the power sound detection part - One of the hooks A block diagram showing #b Kirinari, 餉3- is the block machine 1 of the system + 5 trillion yen part, and the 4th factor is the power used in the electronic assembly according to this invention.
One implementation of the Cal Qin it signal forming unit - [Block diagram shown]
Figure 5 is on-off I! 62 fruits that are members of 1Ml1-
The block h-1 No. 6- is the address signal generation center tIII used for forming the vocal sound collection shown in No. 41.
Block h-1-7- showing AeIil and FIG.
1 is a diagram showing the quantification theory of 6 human tones produced by 6 electronic Qin instruments in -41 with Rakumasu, Kame 1O (2)
is a block imw showing the other real 11 IAH of the vocal electric tone signal form 1 of Kono-yo 6m child music, and Figure 11 shows the pigeon wave intermittent characteristic of the 10th do pass filter.
2 and 913 are further embodiments of the main part of the vocal tone signal forming section for electronic music according to the present invention.

1... S* section, 2... Key insertion detection section, 3... Sound generation -
Corridor, 4...General musical sound signal shadow & part, 5...-One-cal sound imitation part, 6...Vocal Hata sound signal formation part, 7...
・Control signal slowdown capital, 71...11- from Tempofretsuta
172.73...Counter, 74...-measure number setting, 75...Ratio & profit.

4I snoring - person day * Takashi - management company stock company representative coffee - person Bencher - Ki Hiroo man,,': ・ Dejinkafu ■ Happy 8th figure Rin) 4ff Do 1 person) 119 figure Heaaa Wah 111 Heoh 7-Ra 1111

Claims (3)

[Claims] (1) A keyboard section, a human voice collection form one hand wave that forms a musical sound signal of a human voice corresponding to the key pressed on the keyboard section, and a tempo clock that indicates the tempo of the performance. 1IiII- signal generating means for outputting a 1-signal at #&l1h- at a predetermined time, and a phoneme switching means for switching the phonology of the human vocal sound formed by the 1. human vocal sound forming means based on the ° control and Sakai code. Electronic East featuring Nobue Mokoto. (2) The electronic musical instrument according to claim 1, wherein the generation timing of the koji signal outputted from the koji signal generating means is arbitrarily determined. (3) Claim 1, wherein the order of the sounds - the sounds produced by the switch IIi is predetermined.
Electronic drop according to paragraph 2134 - (4) The human voice electric sound forming means selects a predetermined trowel of -L11m from among the legs suppressed by the keyboard section, and produces this - Any one of the 1st to 3rd terms of the range of III Xu Kojiku which forms an electric tone signal corresponding to the translated key.