JPS6180300A - Singing sound generator - 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 Field of the Invention The present invention relates to a singing voice generating device, and more particularly to a singing voice generating device that synthesizes formant frequencies and generates and outputs them as pre-singing voices.

Conventional sound synthesis systems for playing technical music include those for synthesizing musical instrument sounds and those for synthesizing singing voices. Among these, singing voice synthesis systems include, for example, [Singing voice synthesis system for personal computers using formant type speech synthesis LSI 1 (Ishihara, Fushikida, Mitome, Izuchi: 1981 National Conference of the Institute of Electronics and Communication Engineers, 6l.
-198) are known. This method decomposes the input audio character string and musical score character string into parameters whose basic units are consonants and vowels, and also uses the "C pitch parameter" that follows the scale data in the musical score character string. is read from the table, and then, according to the table created by decomposing the consonant and vowel parameters placed in -, the formant parameter and amplitude retrieved from the data ROM are combined with the pitch parameter of the above-mentioned scale to generate synthetic parameters. A time length table is created according to the note length data in the character string.Then, the above synthesis parameters are sequentially transferred to a speech synthesis LSI (Large Scale Integrated Circuit) at predetermined frame periods according to the time length table, and then transferred here. A composite waveform is generated by 10 kl-I Z +J thin rings.

Problems to be Solved by the Invention However, the above-mentioned conventional singing voice systems can only produce singing voice sounds, but they cannot produce musical changes such as vibrato (they are rotational and lack musical expression). There were problems such as.

Also, a device that generates human voice sound at a specified pitch.
There is also a device disclosed in Japanese Patent Application Laid-Open No. 55-77799, but since this device approximates consonants with noise, it is not possible to accurately reproduce rapid changes in formant frequency, and therefore There was a problem that human voice sounds were unclear and unnatural. Furthermore, in the past, only the singing sound was generated, and there was a problem in that it was not possible to cover up the musical instrument sounds that would serve as accompaniment sounds.

Therefore, the present invention particularly focuses on MID I (Musica
lI nstrument D 1g1tal r
2 using the interface of the interrace) standard.
An object of the present invention is to provide a singing voice generating device that can be connected to two synthesizers and solves the above-mentioned problems.

Means for Solving the Problems The singing voice generating device according to the present invention includes a table creation means, a pitch parameter conversion means 9, a whitening parameter creation means, first and second data conversion means, a data transfer means, a first and a second synthesizer and an interface means. The table creation means creates a first table by decomposing the input lyrics data into parameters with consonants and vowels as units, and the pitch parameter conversion means converts the pitch parameters in accordance with the scale data in the input musical score data, for example, in advance. Read from the second table corresponding to the prepared scale. The synthesis parameter creation means generates an edited and interpolated synthesis parameter by combining the pitch parameter with the formant parameter read from the memory according to the first table, and also creates a time length table according to the note value data in the musical score data. Create and temporarily store them. Further, the first data conversion means converts the synthesis parameters into data of a specific standard,
Further, the second data conversion means converts the input musical score data for musical instrument sounds into data of a specific standard.

The data transfer means transfers data related to a plurality of pomanto frequencies among the output data of the first data conversion means, and generates pitch frequency data related to pitch frequencies based on the time length table only during the vowel utterance period. At the same time, the output data of the second data conversion means is sequentially output. Further, the interface means separately supplies data regarding a plurality of formant frequencies and pitch frequency data among the data transferred by the data transfer means as control signals for the plurality of variable frequency oscillators in the first synthesizer, The output data of the data conversion means is supplied as the m+even signal of the plurality of variable frequency oscillators in the second synthesizer.

The plurality of variable frequency oscillators in the first synthesizer +J-Eve are controlled to oscillate and output signals of a plurality of Bormann 6 tot frequencies that constitute consonants and vowels, and another variable frequency oscillator generates and outputs a signal with a pitch frequency that determines the scale according to the musical score data during the vowel part voicing period, so that a singing voice sound is produced from the speaker of the first synthesizer. Meanwhile, in synchronization with this, the plurality of variable frequency oscillators of the second synthesizer are controlled based on the musical score data for musical instrument sounds, so that musical instrument sounds are extracted from the second synthesizer. DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in more detail below with reference to the drawings and embodiments.

Embodiment FIG. 1 shows a block system diagram of an embodiment of the apparatus of the present invention. In the figure, the lyric keyboard 1 has a structure in which, as shown in FIG. 2, for example, a plurality of alphabet keys 17 are arranged in three rows on the keyboard body 16, and alphabet keys representing vowels are arranged in the bottom row. A plurality of keys 17 are selectively pressed according to the lyrics (audio character string), which is a collection of letters written in Roman letters. For example, the character "bi" is input by pressing two keys, ``dou'' and ``da''. As shown in Fig. 2, the lyrics keyboard 1 further has an accent addition key 18 and a voice type indication key 1 at the bottom position.
91 to 194 are arranged respectively3.As is well known, the voice types are classified into J according to vocal range and timbre, and into 6 scales, but here, as an example, -1-19+, 192 ．．
Reference numbers 193 and 194 are used to indicate four types of voices: bass, tenor, aldo, and soprano, respectively. The lyrics data (letter sound data, voice type data, and accent data) input using this lyrics keyboard 1 is l1
0 (input/output) interface 2 to a central processing unit (CPU) 3.

Also, as a keyboard of Haku, note value keyboard 4] Citrol keyboard 5. Furthermore, there are respective keyboards in first and second synthesizers 10 and 11, which will be described later. As shown in FIG. 3, for example, the note value keyboard 4 has a structure in which a plurality of key groups 21 indicating note values (current prices) are arranged in a matrix. On the other hand, ]n 1 ~ roll keyboard 5
is a keyboard for specifying the pitch of a note within the vocal range of the numeral kutabe, for example, by a number within the clause range from ``1'' to ``88'', or for specifying the strength of the blindness. Note value keyboard 4
The output data of both the control keyboard 50 and the control keyboard 50 are supplied to the CPtJ3 via the T10 interface 2 as musical score data for singing voices (score character strings) or musical score data for musical instrument sounds.

CPU3 has random access memory (RAM)6.
A read-only memory (ROM) 7 and a memory (eg, magnetic disk) 8 are each connected via a bidirectional bus. The RAM 6 is a memory circuit for data storage and work of the CPU 3, and also stores pitch parameters and the like corresponding to musical scales to be described later. Also ROM7
A control program for the CPU 3 and a formant parameter table regarding formant frequencies are stored in advance. Furthermore, the memory 8 is a memory circuit for storing input or processed data. Further, the T10 interface 2 is connected to a MIDI interface 9, which will be described later, via a bidirectional bus, and the MTr) I interface 9 is connected to a MIDI interface 9, which will be described later, via a bidirectional bus. -C is connected via the bus.

c PLJ The lyrics data and musical score data input to 3 (musical instrument sound score pulley (yo video))
Video random access via La 12 (
V, RAM) 13, and the video A=1in I~1-
1-La 12 (This J, Liberia S No. 1: After this change 1 is dropped, it is converted into a three primary color signal by the three color signal separation circuit 1/l and C
It is supplied to an RT (cathode ray tube) 15. Also, CRT15
is supplied with three primary colors based on the data generated by the CP (+3). This allows the operator to display the note values displayed on the CRT 15, for example.

While looking at the four tables showing scales, lyrics, and other necessary items, you can operate the keyboards 1, 71, and 5 to confirm and correct the input of desired data, and also confirm the input of musical score data for musical instruments. Corrections can be made in the same way.

Further, CI) U3 has a configuration as shown in FIG. 4, and performs operations according to the flowchart shown in FIG. After the CPU 3 is first initialized (shown in step 35 in FIG. 5), the data receiving section 25 in FIG. 4 reads the keyboard 1. /l and 5 sequentially receive singing voice data consisting of lyrics data and musical score data inputted via the I10 interface 2, and musical instrument sound musical score data inputted from the keyboard 4.5, respectively;
(Steps 36 and 37 in Figure 5). The table creation means 26 shown in FIG. 4 in CPtJa decomposes the above-mentioned imported lyric data into first parameters in which consonants and vowels are units, thereby creating a first daple, which is stored in the RAM 6. (Step 38 in FIG. 5). In other words, it is known that the majority of Japanese syllables consist of combinations of consonants and vowels in a broad sense, and it is thought that the blue expansion of a word is expressed by a sequence of consonants and ffi sounds as units. , the first parameter can indicate the individual syllables representing the lyrics.

In addition, the pitch parameter converting means 27 reads a pitch parameter indicating the fundamental frequency (pitch frequency) FO of the vowel that defines the scale from a table stored in advance in the RAM 6, according to the scale data in the imported musical score data. i (i.e., converts it into a pitch parameter,
This is shown in FIG. 5 as step 39). Further, the synthesis parameter creation means 28 in the CPU 5 shown in FIG. The pitch parameters converted by the pitch parameter conversion means 27 are combined, and the parameters are edited and interpolated (such as adding breaks to the sounds so that the sounds change smoothly) to create synthesis parameters. A time length table is created according to the syllable data in the musical score data, and each is temporarily stored in the RAM 6 (step 40 in FIG. 5).

Here, the singing voice consists of individual syllables indicating the lyrics and the "n" generated together with the syllables, the former being indicated by the consonant and vowel in the first parameter, and the pomant frequency as shown in Figure 6. It is known to exhibit characteristics over time. In other words, it is considered that the identification of blind voices is performed based on the fundamental frequency (pitch frequency) and a plurality of polmantos that make up the speech waveform. 1st. The second and third formants (hereinafter also referred to as formant frequencies) are F+,
In the case of F2 and F3, the consonant part of the - syllable is first pronounced with a gradual change in formant frequency, and then the vowel part is pronounced with a substantially constant formant frequency. Note that during the vowel part utterance period, a pitch frequency is also generated together with the formant frequencies F1 to F3.

Changes in the formant frequencies F1 to F3 in the consonant part differ depending on the consonant. For example, the formant frequencies of the consonants 1, d, and b are shown in Figure 7 (A), (B), and (C).
It is generally known that the change occurs as shown in the figure below. Here, the formant parameters in the synthesis parameters described above are parameters for determining the three formant frequencies F+ to F3 listed in -V. The pitch parameter also defines the pitch frequency.

On the other hand, the 8th verse (lyrics) and the notes generated by it are determined by the sound, size, and length of the note, and this is input into the CPU 3 as musical score data as described above, but it is not converted into a musical scale. Accordingly, the formant frequency vs. time characteristic shown in FIG. 6 becomes a characteristic as if it were moved in parallel in the vertical axis direction, and the pitch parameter (pitch frequency) determines the moving part (pitch of the sound). Also, the quality of the sound can be increased by 1q depending on the duration of the vowel part. Furthermore, the size of the blindness is input via the control keyboard 5, and is included in the synthesis parameters as well as the volume parameter.

This synthesis parameter is converted into data in a format that conforms to the known MTDT standard by the first MIDI data conversion means 29 (this is then taken by the MIDI data).
(The process of step 41 in FIG. 5 is performed.)

). Here, the MIDI standard converts the various operations performed when playing a musical instrument (pressing a key, turning a polycomb, etc.) into several pieces of data, and can be used to control multiple instruments or control them. - This is a standard for sending and receiving data between devices that use the Internet. MIDI data is transmitted as 2 to 3 bytes of serial data, the first byte of which is a status, and the subsequent data bytes are messages that indicate what function they have.

The data length including the status is undefined, 1 to 3 bytes, and the first bit of each byte is “1” for the status and data.
'' is recognized as status, and ``0'' is recognized as data. The above messages can be broadly divided into channel messages for sending individual operation data and system messages for controlling the network.

On the other hand, the musical instrument sound score data received by the data receiving section 25 is supplied to the second MIDI data converting means 30, where it is converted into second MIDI data (
Step 42 in FIG. 5).

This second MIDI data is supplied to the data transfer means 31 together with the first MIDI data taken out from the first MIDI data conversion means 29, where it undergoes data processing in steps 43 to 56 shown in FIG. Sunawa 15,
In FIG. 5, the data transfer means 31 sets the variable J for counting the total number of notes to [1] (step 43), and sets the other counting variables 11 and I2 to [0], respectively.
(step 45), from the first synthesizer 10 M T I) I inter 7-[-s 9 and I
10 It is determined whether the synthesizer 1f10 is capable of receiving data based on the data coming in through the interface 2,
If reception is not possible, proceed to step 53, which will be described later.
Come on, second Shin 1'! The synthesizer 11 determines whether or not it is receivable based on the data that comes in from the If 11 via the MIDI interface 9 and the I10 interface 2, and if it is in a receivable state, it performs data processing in step 54, which will be described later. If the reception is not possible, the process returns to step 45 and the first synthesizer live 10
Check to see if it can be received.

1st Synth 4J-Iza 10 can receive (enable)
If this occurs, the formant frequencies F1 to F3 of the indicated consonant to be reinserted in the first MIDI data.
After transferring the first data relating to the following data to the synthesizer 10 through the 10 interface 2 and the Mll) I interface 9, the value of the variable 11 is added by [1] (step 46.47). The processing operations of steps 46 and 47 are repeated until the value of variable 11 after sieving becomes equal to or greater than N (step 4B).

Here, the value of N indicates the number of frequency changes to be made for each of the first to third formant frequencies F1 to F3 at regular intervals (for example, 0.33 m5) during the voicing time period of the consonant part. . As a result, in steps 46 to 4B described above, the frequency change of the formant frequencies F1 to F3 in the consonant part voicing time period as shown in FIGS. 6 and 7 is performed in a linear approximation. The first data of is transferred, and the formant frequency data is rewritten every approximately 311 IS (3×3×320 μs).

When the value of variable 11 becomes N or more, then the first MID
Formant frequencies F1 to F3 of vowels in the I data. The second data regarding pitch frequency FO, volume, etc. is the first data.
The signal is transferred to the synthesizer 10, where pitch frequency Fo is set, etc. (step /19). Then, based on the time length table, a time calculation is performed to check the length of one interval of the vowel part to determine the time length according to the note value, and when the predetermined time length is reached, the second data is The transmission of is determined to be A-f (step 50). The data transfer means 31 then clears the first data (step 51
), the value of the variable is increased by 11'' (state 52).

After that, it is determined whether the synthesizer 11 can receive data based on the data from the second synthesizer 11,
When the data is ready for reception, the data transfer means 31 transfers the second MIDI data to the synthesizer 11 via the I10 interface 2 and the MIDI interface 9 (step 53.5/l). Incidentally, it is possible to determine which of the synths (Vibes 10 and 11) the data is to be transferred from the MIDI interface T9 to by a channel message of the MIDI data stator 18.

The data transfer means 31 next increases the value of the variable 12 by [11 (step 55), and then checks whether the value of the variable J is greater than the total number of notes m1 of the musical score for singing voice, and whether the value of the variable I
It is determined whether the value of 2 is greater than the total number of notes m2 of the musical score for musical instruments.

The data processing operations of steps 45 to 55 described in - are repeated until the value of I2 exceeds "+ * ' +112,
The operation ends when both ml and I2 are exceeded.

The first and second MIDI data extracted in this way are supplied to the Ml and DI interface 9 through the I10 interface 2, and further distributed and supplied from there to the first synthesizer 10 or the second synthesizer 11.

FIG. 8 shows a circuit diagram of an example of the MIDI interface 9. The MIDI data (first and second MIDI data) from the I10 interface 2 is sent to the signal processing device 58.
The data is supplied to the input/output terminals o to D7, where it is parallel-to-serial converted and sent to inverters 59, 60, . Resistance 61. The signal is transferred to the synthesizer 10 or 11 via the connector 62 or the like.

Also, data from the synthesizer 10 or 11 is sent to the connector 63. Signal processing device 5 via photocoupler 6/l, respectively.
8 and output in parallel from its data input/output terminal Do-r)7. Furthermore, the signal station]]! The device 58 is supplied with a clock of, for example, 500 kHz from the clock module IC65.

Furthermore, for the network when connecting synthesizers 10 and 11, the inputs are directly connected to inverters 66 and 67.
, resistor 68 and connector TI-IRU terminal) 60.

FIG. 9 shows a block diagram of an embodiment of the first synthesizer 10. MID [1st from interface 9
MIDI data is sent to the input terminal 701 via an input/output interface (not shown) in the synthesizer 10.
06 are input sequentially.

Input data at input terminals 701-704 are supplied to voltage controlled oscillators (VCO) 71-74 as control signals. Here, VCs 071, 72 and 73 are the first VCs, respectively. Generates and outputs signals of the second and third formant frequencies, and
O74 oscillates and outputs a signal of pitch frequency FO. Furthermore, the data input to the input terminal 705 is transmitted to a low frequency oscillator (l).
FO) 75 and controls its oscillation operation. The output signal of this L F075 is supplied to VC071 to VC074 to obtain a vibrato effect by variable control of the output oscillation frequency, and the output signal of L F075 is supplied to voltage control filters (V, CF) 78 and Voltage controlled amplifier (V
It is possible to emphasize only specific harmonics or obtain a tremolo effect by supplying each signal to CA>79.

Each output signal of VC071 to 74 is sent to the ring modulator 7.
6, where it is converted into a signal with a sum or difference frequency of each frequency, and then supplied to a mixer 77, where V
It is mixed with each output signal of C071 to C074. The ring modifier 6 is used to obtain a specific tone, and its operation is controlled by data from a terminal 706. The output mixed signal of mixer 77 is VCF7B, VCA7
9 and an amplifier 80, a signal having a waveform as shown in FIG. 10 is outputted to a speaker 81.

In FIG. 10, the king indicates the pitch period.

Here, predetermined data is applied only to the terminals 701 to 703 during the first consonant production period, so the first . second and third formant frequencies F+,
F2 and [3 are output simultaneously, and their frequency values are divided into the N stages and sequentially changed, but V
No signal is output from the CO74. On the other hand, in the vowel voicing time period, the pitch frequency Fo that determines the musical scale is generated and output from the VCO 74 for the first time based on a part of the second data, and the VCs 071 to 73 continue to produce the first and third formants. Signals of frequencies F1 to F3 are generated at constant values, respectively. As a result, the lyrics inputted by the lyrics keyboard 1 are pronounced by the speaker 81 in a human voice according to the musical score and reinsertion inputted by the note value keyboard 4 and the control 4--board 5. Furthermore, musical changes such as vibrato, tremolo, etc. can be applied to the human voice sound, that is, the vocal sound. As a result, the singing voice becomes more nuanced and more desirable.

Note that the synthesizer 10 shown in FIG. 9 can also obtain only musical instrument sounds by using the keyboard 82 as in the conventional case.

On the other hand, the second MIDI data is applied to the VCO, LFO, etc. in the second synthesizer 11 having the same configuration as in FIG. pronounced. Therefore, the singing voice from the first synthesizer 10 is emitted in synchronization with the musical instrument sound from the second synthesizer 11 until t before the singing voice from the first synthesizer 10.

In the present invention, the pitch parameter converting means is not limited to reading from a table, but can also calculate the value. Further, the number of formants may be a plurality other than three, and the waveform of the consonant synthesized speech signal supplied to the speaker 81 is as shown in FIG. It may also be a waveform synthesized with.

As described in ``Effects of the Invention'', according to the present invention, the pre-singing voice and the instrumental sound can be generated separately or simultaneously using two synthesizers, so that the pre-singing voice with the instrumental sound as the accompaniment sound can be generated separately or simultaneously. Because it synthesizes formant frequencies, etc., it can generate natural and clear singing voices compared to conventional devices that use why-noise as consonants, and it also uses a synthesizer. Therefore, it is possible to create a musical change before the singing voice, and compared to conventional devices, it has many features such as being able to produce a much more attractive and expressive voice before the singing voice. It is something.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the device of the present invention;
2 and 3 are diagrams showing the schematic configuration of each keyboard in the block system shown in FIG. Flowchart jl-1-1 for explaining the operation of the main parts of the inventive device Figure 6 is a diagram showing the relationship between the consonant part, ffl sound part and formant frequency of synthesized speech, and Figure 7 is a diagram showing the relationship between the formant frequency of each consonant and time. Figure 8 shows the relationship between M in the block system shown in Figure 1.
A circuit diagram of an example of an IDI interface, FIG.
FIG. 10 and FIG. 11 are block system diagrams showing one embodiment of the synthesizer in the block system shown in FIG. . DESCRIPTION OF SYMBOLS 1... Lyrics keyboard, 2... I10 interface, 3... Central processing unit (CPLJ), 4... Syllable keyboard, 5... Control keyboard, 6...
...Random access memory (RAM), 7...
Read-only memory (ROM>, 9...MIDI
Interface, 10.11...Synthesizer, 1
2... Video controller, 13... Video random access memory (V, RAM), 25... Data receiving unit, 26... Table creation means, 27...
pitch parameter conversion means, 28... synthesis parameter creation means, 29... first Mll) I data conversion means, 30... 20th M1r) I data conversion means, 31.
...Data transfer means, 70+"706...Input terminal,
71・~7/I...Electric 1-f-controlled oscillator (VCO)
, 75...Low frequency oscillator (1-[)), 77...Miki 1no. Patent applicant: Victor Japan Co., Ltd. = 26-

Claims (1)

[Claims] table creation means for creating a first table by decomposing input lyrics data into parameters having consonants and vowels as units; pitch parameter conversion means for converting into pitch parameters according to scale data in input musical score data; The formant parameters read from the memory according to the first table are combined with the pitch parameters to generate edited and interpolated synthesis parameters, and a time length table is created according to the note value data in the musical score data. a first data conversion means that converts the synthesis parameters into data of a specific standard; and a second data converter that converts the musical score data for the input instrument sound into data of the specific standard. a converting means, first and second synthesizers that mix and output output signals of a plurality of variable frequency oscillators, and transfer data regarding a plurality of formant frequencies among the output data of the first data converting means; and generates and outputs pitch frequency data regarding the pitch parameter based on the time length table only during the vowel part utterance period,
data transfer means for sequentially outputting the output data of the second data conversion means; and a data transfer means for sequentially outputting the output data of the second data conversion means; Interface means for separately supplying the plurality of variable frequency oscillators as control signals, and supplying output data of the second data conversion means as the control signals for the plurality of variable frequency oscillators in the second synthesizer; A singing voice generating device characterized by: