JPH01219898A - Speech synthesizing device - Google Patents

Abstract

PURPOSE:To obtain a smooth voice by sectioning the generation time of respective phonemes into plural time zones by the phonemes which constitute a syllable, specifying phoneme parameters of the sectional area of acoustic tubes by the time zones, and interpolating the phoneme parameters and superposing the energy of a sound wave on a pattern prescribed by an exponential function. CONSTITUTION:The voice channel of a human is regarded as an acoustic tube group and made to correspond to a circuit element group of surge impedance components, and then a voice is simulated according to the current wave at the output terminal of the circuit element group. The propagation of the acoustic wave of an acoustic tube model is substituted with a flow of the current of the equivalent circuit and parameters of the pitch of the current source and the sectional area of the acoustic waves are prescribed by the phonemes and the parameters are interpolated as to joins of phonemes or joins of time zones sectioned in the phonemes. As for the energy, energy patterns corresponding to accent prescribed by an exponential function previously by words and the interpolated value of the energy is superposed on the energy pattern to obtain a voice closer to the human voice.

Description

DETAILED DESCRIPTION OF THE INVENTION A. Field of Industrial Application The present invention relates to a speech synthesis device using an acoustic tube model.

B3 Summary of the Invention The present invention regards the human vocal tract as a group of acoustic tubes, and by associating this with a group of circuit elements for surge impedance components, it is possible to simulate speech based on the current wave at the output end of the group of circuit elements. The generation device divides the generation time of each phoneme into multiple time periods for each phoneme that makes up a syllable, and specifies phoneme parameters such as the cross-sectional area of the acoustic tube and the energy of the sound source wave for each time period. In addition to interpolating the parameters, the group of interpolated values of the energy of the sound source wave is superimposed on a pattern defined by an exponential function according to the accent of the word, thereby creating a smooth sound that approximates human speech. This is what I did.

C0 Conventional technology Electronic devices that artificially synthesize and output so-called sounds, such as voice synthesis and music synthesizers (electronic musical instruments), have recently become LS for voice recognition and voice synthesis using one or several chips.
I can now be realized at low cost through audio information processing and semiconductor large-scale integrated circuit technology, and various systems have been proposed depending on the purpose of use and constraints. This speech synthesis method involves two methods: recording and editing raw human speech, which combines them appropriately and editing them into sentences; There is a parameter method that extracts only the parameters and controls them during the speech synthesis process to artificially create a speech signal.

In the parameter method, the audio waveform is sampled at certain intervals, the audio signal value at each sampling point is converted from analog to digital, and the values are displayed as codes of 0 and 1. In order to record faithfully, it is necessary to increase the number of bits, which requires a large memory capacity.

Therefore, various highly efficient encoding methods are being researched and developed in order to reduce the amount of information as much as possible.

One such method is a delta modulation method in which at least one bit corresponds to the information of one audio signal. In this method, one bit is used to determine whether the next audio signal value is higher or lower than the current value, and if it is higher, it is coded "1".
”, if it is low, the code “0” is given and the audio signal is encoded.In the actual system configuration, a constant amplitude step amount (delta) is determined, and the previous code is changed to prevent errors from accumulating. The residual signal between the audio value obtained by the encoding and the input audio signal is encoded.

This type of predictive coding is called a linear prediction method (predicting from several previous sample values) and the Percoll method (using a partial autocorrelation function called Percoll coefficient instead of the prediction coefficient of the linear prediction method). There is.

D1 Problems to be Solved by the Invention Of the conventional speech synthesis methods, the recording and editing method has a problem in that the types of word surprises and sentences that can be synthesized are limited.

In addition, in methods using predictive coding, it is difficult to create articulatory combinations that correspond to the joints between sounds, and a method for combining synthesis units has not been established. In the process of going from a steady state to a consonant through a transition, and then through a vowel transition to a steady vowel sound, the sound at the joint between the vowels becomes disconnected. Therefore, there is a problem that the sound lacks smoothness and does not give a natural feeling when heard by humans.

An object of the present invention is to provide a speech synthesis device that can synthesize sentences of any amount and provide a smooth sound that is close to actual human speech and can give a natural feeling to the listener. It's about doing.

Means and operations for solving the E0 problem (1) Basic concept In order to radiate sound outward from the mouth, a sound source is required, and this sound source is produced by the vocal cords. -The vocal cords have the function of intermittent exhalation by opening and closing two folds, and these intermittent intervals generate airflow called puffs, and when the vocal cords are tensed, tension is applied to these folds, causing the folds to open and close. The frequency increases, producing a high-frequency puff sound. When the exhalation flow is increased, the sound becomes louder.

When this sound source wave passes through a cylindrical acoustic tube like the vocal tract, a resonance phenomenon causes certain components of the sound waves from the open end to be emphasized and certain components to be attenuated, creating a complex vowel waveform.

Even if the sound source waves have the same waveform, the sound emitted from the mouth is affected by the shape of the vocal tract that the sound passes through before being emitted from the lips. That is, the sounds produced by humans are determined by the length and cross-sectional area of the vocal tract from the vocal cords to the lips, and the way the vocal cords vibrate.

The present invention has been made with this in mind, and considers the vocal tract as a group of acoustic tubes with variable cross-sectional areas.
Furthermore, the starting point is to realize the traveling wave phenomenon representing the transmission of sound waves in an acoustic tube using its equivalent circuit. If we consider the vocal tract as an acoustic tube, the propagation of sound waves in each acoustic tube can be thought of as a repetition of reflection and transmission phenomena at the boundary surfaces of each acoustic tube, dividing them into forward waves and backward waves. and transmission are quantitatively determined according to the degree of mismatch of acoustic characteristic impedances at the interface, that is, the ratio of the respective cross-sectional areas of adjacent acoustic tubes.

The reflection and transmission phenomena described above are the same as the transient phenomena that occur when impulse currents are passed through lines with different impedances in an electric circuit.

(2) Equivalent circuit Based on the above, a sound tube model consisting of n sound tubes S, ~Sn is shown in Figure 1 (A).
It can be expressed as an electric circuit in which a group of circuit elements (T, -T,) made up of lossless surge impedance components without resistance are connected in series as shown in FIG. A1~
A is the cross-sectional area of the acoustic tubes 81 to Sn, respectively. Here, in the present invention, basically, by applying the above-mentioned electric circuit and changing the impulse current supplied to it and the surge impedance of each circuit element T, ~Trl, the sound source wave of the acoustic tube model and each The sound obtained from the acoustic tube model is simulated by supplying the current output from the final stage circuit element Tn to a sounding section such as a speaker in response to changing the cross-sectional area of the acoustic tube. There is.

Specifically, we assume a circuit equivalent to the above electric circuit as shown in Figure 1 (2), change the current of the current source in this equivalent circuit with respect to time, and create the calculation formula as described later. Since the cross-sectional area ratio of the acoustic tube is introduced, each cross-sectional area A, -An is changed with respect to time, and the current value of each part is calculated based on this. In the same figure, P
is the current source, Zo is the impedance of the current source, Z1~Zn
are the surge impedance and Z of the circuit elements T and -Tn, respectively.
L is radiation impedance, 10^~f (n-+l^・
i+n〜i nB, l OA-a fn-11A+
a ts-a nB is the current in the current path corresponding to each symbol, W OA -Wi-1, ^, W, , ~WnB is the current source, I OA-1 + n-11A is the backward wave current, I + a ~
Ins indicates forward wave current. In this equivalent circuit,
For example, circuit element T1. Focusing on the connection part of Tt, the current source WIA corresponds to the current La flowing toward T through the circuit element T1, and the circuit element T.

Assuming a current source WIA corresponding to a current 11A flowing through it towards TI, currents ■ and B are circuit elements T□.

The reflected wave current 11B reflected to TI at the boundary of T and T
, the transmitted wave current alA transmitted to
This equivalently represents that 1A is divided into a reflected wave current 11A that is reflected to T at the boundary between circuit elements Tt and TI, and a transmitted wave current alB that is transmitted to T1. Further, FIG. 1 (1) is a schematic diagram schematically showing such a situation.

(3) Calculation First, the block including the first stage current source P shown in FIG. 1 can be thought of as a superposition of two circuits as shown in FIG. Therefore, if we set the voltage of current source P as ■, then
The currents al and a in the figure are expressed by equations (1) and (2), respectively, and as a result, the current a. A is expressed by equation (3).

al=v/zo+zt・(1)
at =Zo/Zo+Z+・lot −(2
) aO^=a, +a.

= l/Zo+Z+(V+Zo・l0A) −(3
) Now, if we start supplying current into the equivalent circuit for the first time, by setting IOA to zero, a. Find A.

Then, calculations are performed sequentially based on this value.

Taking as an example the calculation formulas for the current values of the first block and second block located at the left end of the figure, the following (4) to (
12) It is expressed as follows.

11oA': l/Zo+ Z+(V'+Zo・l0
A) −(4) ioA'=loA' IOA
”' (5) 10A'= l +a'
+ a +8' ・-(6) a+B
'=s+a(1+s+IIA)...
(7) i+a'-ala' l +e
”・(8)1 +a'= I OA'10a O
A'...(9) a+A'=
s+A(1+B+I IA) ・=
(10) 11A'-a IA' I +8
”' (11)! +A'-12B'
+ a tB' ・-(12
) As such calculations proceed, the arithmetic expressions for the final stage block are expressed as equations (13) to (15).

a nB': zL, / Zn + "LL-1nB
+++ (13)j nB"" a nB' l
n8I ns"" i (n-11A+ a 1n-1
1A H... (14) In this way, the final stage acoustic tube Sn
A current Ins corresponding to the sound wave emitted from the waveform is determined. However, SIB+SIA are coefficients expressed by the cross-sectional area ratio of adjacent acoustic tubes, respectively (15) and (
16) Formula % Formula % A series of calculations of the current values of the blocks from the first stage to the final stage are executed instantaneously, and these calculations are performed one after another at a predetermined timing. Here, the above (4) to (1)
In formula 4), the value with a dash is the calculated value at time t, and the value without a dash is the calculated value obtained by the calculation one time before the calculation at time t. The digital value lnB obtained in this way is converted from digital to analog to create an analog current, and this current is supplied to a speaker or the like to obtain sound. The timing of the calculation is determined in consideration of the speed of sound. For example, by setting the propagation time of one acoustic tube as the calculation time interval, the backward wave current (2) is determined. A~1 (n-11A and forward wave current IIB-1nB each circuit element T at the same speed as the speed of sound,
~1. A state equivalent to the state flowing through A is created, thereby matching the acoustic tube model and the electric circuit model.

The present invention is based on the realization of the equivalent model and calculations described above, and specifically, the utterance time of each phoneme is divided into one or more time periods for each phoneme that makes up a syllable, and each phoneme is divided into one or more time periods. For each time period, each initial value of the pitch, which is the repetition frequency of the sound source wave, the energy of this sound source wave, and the cross-sectional area of the acoustic tube, and each initial value X of the next time period is determined from the above-mentioned initial value Xo of the relevant time period.
a phoneme parameter storage unit that stores constants and sound source wave patterns that define how to change to , , and a parameter interpolation processing unit that performs interpolation processing of the pitch, energy, and cross-sectional area corresponding to the input phoneme data. Then, the energy patterns of the sound source waves corresponding to the accent of each word are classified into multiple types, and the energy pattern for each type is defined by an exponential function, and the pattern corresponding to the input word is determined from among these patterns. and a pattern processing section that selects a current value outputted from the output end of the circuit element group based on the parameters interpolated by the parameter interpolation processing section, and performs interpolation processing on the calculated value of energy. a calculation unit that uses a value obtained by superimposing an exponential function corresponding to the pattern selected by the pattern processing unit on the interpolated value group, and a voice generation unit that generates sound based on the calculation result of the calculation unit, The parameter interpolation processing unit performs interpolation calculations many times during each of the time periods using the initial value X0, the X1 corresponding to the target value, and the constant, and for the energy, an exponential function for interpolation is used. It is characterized by being executed based on.

F. Embodiment FIG. 1 is a diagram showing a block configuration of an embodiment of the present invention. 1 is a Japanese language processing unit, which converts the input Japanese text into pronunciations by referring to clause breaks and a dictionary. Reference numeral 2 denotes a sentence processing section which performs processing for adding intonation to sentences. 3 is a syllable processing unit, which adds accents to the syllables that make up a sentence according to the intonation. For example, the sentence "Sakura ga Saita" is broken down into syllables such as rsAj, rKUJ, rRAJ, etc., and an accent is added to each syllable. The intonation of a sound is determined by the repetition frequency of the sound source wave, which will be explained later.
Since it is determined by the energy and time, adding an accent means determining coefficients for these parameters. In particular, the determination of coefficients for energy is executed in the pattern processing unit 31 within the syllable processing unit 3.

Regarding the pattern processing unit 31, energy patterns of sound source waves corresponding to the accent of each word are classified and stored into three types, for example, a head height pattern, a tail height pattern, and a middle height pattern, and these three types of energy patterns are stored. It has a function of selecting a pattern corresponding to a word to be uttered from patterns. Figure 4 (A) to (T) are respectively “
This is a diagram showing the results of analyzing the speech actually uttered by humans for the words "It worked,""Abnormal", and "The circuit breaker", where the solid line shows the change in the energy of each syllable, and the dotted line shows the energy of each syllable. The energy patterns corresponding to the accents of the words obtained by connecting the peak values of are shown. In this example, the patterns shown by the dotted lines in FIG. Figure 5 (A) - (T)
The pattern processing unit 3 is defined by an exponential function as shown in FIG.
1 in advance.

In this case, each stored pattern has an exponential function defined by a different formula, for example, assigned to an ascending portion and a descending portion. To extract this energy pattern, for example, a code indicating the type of energy pattern is attached to the word registered in the dictionary in advance, and the pattern processing section 31 selects an energy pattern corresponding to this code.

4 is a phoneme processing unit, 4□ is a syllable parameter storage unit,
The phoneme processing unit 4 includes a syllable parameter storage unit 4. which defines the correspondence between syllables, vowels, and consonants, which are units of phonemes, for input syllable data such as rsAJ. For example, the phoneme "S" for the syllable rsAJ.

Take out rAJ.

5 is a parameter interpolation processing unit; 5 is a phoneme parameter storage unit; 5. is a sound source parameter storage section.

As shown in FIG. 6, the phoneme parameter storage unit 5□ divides the utterance time of each phoneme into a plurality of time periods, for example, three time periods OI to 03, and stores the duration, pitch, which is the repetition frequency of the sound source wave, and this value for each time period. Stores each initial value of the energy of the sound source wave and the cross-sectional area of the sound tube, a time constant that specifies the manner of change from each of the above-mentioned initial values in the relevant time period to each initial value in the next time period, and a sound source wave pattern. ing. In this example, the human vocal tract (17 cN for a male) is connected to an acoustic tube of length lcx.
The model is made up of individually connected objects, and therefore, 17 cross-sectional area values (A, to A17) are determined for each time period. Further, in the sound source parameter storage unit 5°, waveform components of three types of sound source wave patterns 61 to G3 are stored as 50 sample data, as shown in FIG. 7, for example. The parameter interpolation processing unit 5 processes each time period (0, ~
This is the part that performs interpolation processing of pitch, energy, and cross-sectional area in 03), and this processing is performed based on the pitch, energy, and cross-sectional area of the relevant time period.
Let the initial value of each parameter of energy and cross-sectional area be Xo, the initial value of the next time period be Xr, the nth interpolated value be X(n), and the time constant corresponding to each parameter be D. This is a process in which calculations are performed n times during the time period according to the recurrence formula shown in equation (17). However, the initial value X (0)
is the above Xo.

X(n)=D (X, -X(n-1)) +X(n-1
)...(17) For example, regarding the energy interpolation process in the time period O1, since X oh<E 1 % X r corresponds to E2, it is calculated according to equation (18).

X(n)-DE+(Ex X(n l))+X(n
1') = (18) Here, the above equation (17) is transformed into the following (1
This is a recurrence formula of equation 9).

X=X, -e-'' -(19) That is, when equation (19) is differentiated, equation (20) is established, and therefore (21) is established.

d x / d t = De −” − (20) Δ
X=X(n+1)-X(n)=Δt, De-D t
I n 1−Δt −D(X, −X(n)) −
(21) Therefore, the formula (22) is obtained.

X(n+1)=Δt-D(X,-X(n))×X(n
)...(22) Here, since the time interval of the interpolation calculation is constant, Δt-D can be collectively replaced with a time constant, which is expressed as equation (17).

Reference numeral 6 denotes a calculation unit, which calculates the digital value of the current 1na shown in FIG. 1 at the same timing as the interpolation calculation, for example, at a time interval of lOOμS, based on the parameters calculated by the parameter interpolation processing unit 5. 7 is a digital/analog (D/A) converter, which generates a current wave (analog current) based on the digital value obtained by the calculation section 6. Reference numeral 8 denotes a voice generating section such as a speaker, which generates voice based on analog current.

Next, the operation of the above embodiment will be described.

A Japanese sentence inputted by a word processor or the like passes through a Japanese language processing section 11, a sentence processing section 2, and a syllable processing section 3, where intonation etc. are added and it is divided into syllable units.Furthermore, each syllable is divided into phonemes by a phoneme processing section 4. Decomposed. Next, the parameter interpolation processing unit calculates the pitch of each phoneme,
The energy and cross-sectional area are stored in the phoneme parameter storage section 5. , and interpolation processing is performed on these parameters for each time period (0° to 0.). This interpolation process is (1
For example, energy in time period 01 is performed according to equation (18). 8th
The figure shows this situation, and each interpolated value of energy E (1), E (2) obtained by interpolation calculation is shown in the figure.
-E (n) are arranged along the curve expressed by the following equation (23).

E = E t e-Dt- (23) Also, each time period 0
Sample data of the sound source wave patterns defined for each of 1 to 03 is taken out from the sound source parameter storage section 5, and this sample data and interpolated values such as pitch are given to the calculation section 6. , the calculation detailed in section (3) "Operation" is executed. In this calculation, the syllable processing unit 3
In , the coefficient or function corresponding to the accent added to each syllable unit is multiplied by each parameter obtained by the parameter interpolation processing unit 5, and the intonation of the sentence is calculated. In particular, regarding energy, the pattern processing unit 31 selects an energy pattern (see Figure 5) corresponding to the accent of the word based on the code attached to the word extracted from the dictionary, and In order to draw the energy pattern, read out the exponential function value that defines the pattern (the value on the vertical axis in Figure 5), multiply the read value by the interpolated value of energy, and then calculate the combined value. is used as an element of the calculation.

In this way, the digital value of the current wave corresponding to the sound wave emitted from the final stage sound tube is obtained, and based on this value, the D/
A current wave is created by the A converter 7, and a corresponding sound is emitted from the vocalization sound 8.

G0 Effects of the Invention According to the present invention, the propagation of sound waves in a sound tube model is replaced by the flow of current in an equivalent circuit, parameters such as the pitch of the current source and the cross-sectional area of the sound tube are defined for each phoneme, and the Since parameter interpolation processing is performed on the seams between 2 and 2 or between divided time periods within a phoneme, smooth speech can be obtained, giving a natural feel to the listener.

As for the energy, interpolation processing is performed based on an exponential function, so the arrangement of interpolated values is similar to that of actual speech, and an energy pattern corresponding to an accent defined in advance by an exponential function is assigned to each word, and the energy Since the interpolated energy value is superimposed on the pattern, it is possible to obtain a voice that is even more human-like. Furthermore, instead of storing all parameter values in the area corresponding to the joints between phonemes in the memory, it is sufficient to store data in units of phonemes or units of time, so the memory capacity can be reduced.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram showing an equivalent model of an acoustic tube, FIG. 2 is an equivalent circuit diagram showing a block including a current source, FIG. 3 is a block diagram showing an embodiment of the present invention, and FIGS. 4 and 5 are explanatory diagrams each showing an energy pattern. FIG. 6 is a data diagram of phoneme parameters, FIG. 7 is an explanatory diagram showing a sound source wave pattern, and FIG. 8 is an explanatory diagram showing the state of parameter interpolation processing. 31... Pattern processing unit, 4... Phoneme processing unit, 41
. . . syllable parameter storage section, 5. . . parameter interpolation processing section, 5. . . . phoneme parameter storage unit, 5. . . . Sound source wave pattern storage section, 6. Arithmetic section, 7. Digital/analog conversion section, 8. Generation section. Fig. 2 1 Door” Prefecture and Including Professional 1, 16 Supposedly I Registerable Title Fig. 3 V Station Sorting etc. A- Times m Cause Fig. 4 Construction Rule R- Pattern C Explanation Field: 1) HA DA N KI WA 5th art sequence 1000 yen ba tsun oLdA fig. initial hi tsu tsun Odaka bahoon 6th figuration parameter data l Fig. 7 Mr/1Jljo+

Claims (1)

[Claims] (1) By regarding the human vocal tract as a plurality of acoustic tubes connected in tandem, and by associating the acoustic tube group with the circuit element group of the surge impedance component and making the sound source correspond with the current source, the acoustic tube In a speech synthesizer that simulates a speech wave emitted from an output end of a group of circuit elements based on a current wave of an output end of a group of circuit elements, the utterance time of each phoneme is divided into one or more time periods for each phoneme constituting a syllable. For each time period, the initial values of the pitch, which is the repetition frequency of the sound source wave, the energy of this sound source wave, and the cross-sectional area of the acoustic tube, and the initial values a phoneme parameter storage unit that stores constants and sound source wave patterns that define how to change to each initial value X_r; and the pitch corresponding to the input phoneme data;
A parameter interpolation processing unit performs energy and cross-sectional area interpolation processing, classifies the energy pattern of the sound source wave corresponding to the accent of each word into multiple types, and defines the energy pattern for each type using an exponential function. , a pattern processing section that selects a pattern corresponding to the input word from among these patterns, and a current value output from the output end of the circuit element group based on the parameters interpolated by the parameter interpolation processing section. A calculation unit that uses a value obtained by superimposing an exponential function corresponding to the pattern selected by the pattern processing unit on the interpolation-processed interpolation value group, and a calculation result of this calculation unit. and a voice generation unit that generates a sound based on the parameter interpolation processing unit, wherein the parameter interpolation processing unit performs interpolation calculation multiple times during each time period using the initial value X_o, the X_r corresponding to the target value, and the constant. A speech synthesis device characterized in that the energy is determined based on an exponential function for interpolation.