JPH05197396A - Voice synthesizer - Google Patents

Abstract

PURPOSE:To obtain a voice synthesizer in which a musical instrument tone and a DTMF signal can be also generated and for which miniaturization can be attained by outputting one or more kinds of vocal cord tone waveforms generated at a voice waveform generating part and one or more kinds of sine waveforms generated at a sine waveform generating part comprehensively from a waveform unifying part. CONSTITUTION:A voice synthesizer 30 is provided with the sine waveform generating parts 32, 33 and the waveform unifying part 34 in addition to the voice waveform generating part 31, and outputs plural waveforms obtained from those generating parts by unifying in one waveform by the waveform unifying part 34 comprehensively. Thereby, synthetic voice can be outputted and also, a sine wave can be accurately outputted, and a chord and the DTMF signal can be generated. The number of components in the waveform unifying part can be reduced by remarking the unification of plural kinds of voice when the vocal chord is generated in conventional technique, and sharing the sine waveform generating part with the waveform unifying part by incorporating in the synthesizer.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice synthesizer for electronically synthesizing a human voice and outputting the same, for synthesizing a buzzer sound in addition to a human voice, and particularly for controlling a telephone line. To a voice synthesizer capable of synthesizing the DTMF signal.

[0002]

2. Description of the Related Art A speech synthesizer according to the prior art aims to synthesize a human voice. For example, it is realized by a sound source and a vocal tract transfer function simulating a human voice generation mechanism. Therefore, a spectrum and a sound source intensity are generally used as the plurality of characteristic parameters (characteristic parameter series) used in the speech synthesizer, and they are arranged in time series in the order of utterance of speech.

Here, the spectrum is the result of Fourier transform of the speech waveform, and the dominant portion characterizing the vowel therein is called a formant, and its frequency (formant frequency) is one of the characteristic parameters. Used.

The portion where the waveform of the spectrum forms a pole and a zero is called a pole-zero pair, and the frequency (band) of the pole-zero pair.
Is used as one of the feature parameters. The characteristic parameter is composed of a set of various parameters such as the formant frequency and the pole-zero pair frequency.

FIG. 4 shows an example of a concrete circuit configuration of such a speech synthesizer (Japanese Patent Laid-Open No. 3-63699).
In this circuit, radiation characteristics are given to each sound source signal generator.

In FIG. 4, reference numeral 101 is a polynomial waveform generator, which in this example generates a signal of a vocal cord sound source waveform in which high-order emphasis of a fourth-order polynomial waveform is performed. It is also possible to use a waveform obtained by another calculation formula or a waveform extracted by analysis from natural speech.

Reference numeral 102 is a random number generator which generates a random waveform using random numbers and outputs a signal of a frictional sound source waveform with high-frequency emphasis.

Reference numeral 103 is an impulse waveform generator for generating a signal of a burst sound source waveform in which high-frequency emphasis is applied to a step waveform, that is, an impulse waveform.

The buzzbar / nasal sound branch 11, the aspirated sound / voiced sound branch 12, the fricative sound branch 13 and the burst part branch 14 respectively are a resonator and a predetermined frequency part for emphasizing a predetermined frequency part of the signal. The anti-resonator that attenuates is cascaded. In particular, the aspirator / voiced sound branching unit 12 has a resonator 1 for creating a waveform indicating a pole-zero pair of vowels.
26, 128 and anti-resonators 127, 129 are provided.

In such a circuit configuration, each waveform generation component 10 is generated by an operation signal according to the voice to be synthesized.
1 to 103 are activated, and the intensity of the sound source signal is adjusted by the corresponding intensity modulator. After that, a specific frequency band is emphasized and attenuated by the resonator and the anti-resonator of each branching portion, and the adder 15 outputs an audio signal that has been subjected to the articulation processing and the radiation characteristic addition processing.

[0011]

The conventional speech synthesizer of this type is dedicated to the synthesis of human voice and is not suitable for synthesizing other sounds. From a practical point of view, it is desirable to synthesize a buzzer sound other than the human voice. If the sound is a periodic sound consisting of a single fundamental frequency such as a buzzer sound and its harmonics, the polynomial waveform generator 101 generates a sound of that fundamental frequency, and the resonators / anti-resonators 121 to 129 are used. By controlling the power of the required harmonic component, it is possible to create a desired sound.

However, due to the structure of the synthesizer, the periodic sound source is limited to the polynomial waveform generator 101 corresponding to the vocal cord sound source, so that a sound in which two or more periodic sounds are mixed cannot be synthesized. Furthermore, in the case of synthetic speech, when the fundamental frequency is relatively low and precise precision is not required, the fundamental period is limited to an integral multiple of the sampling frequency in order to easily realize a polynomial waveform generator. There are many. In this case, the fundamental frequencies that can be set are discrete values such as 200 Hz and 204.08 Hz for voice, and 1428.57 and 16 for buzzer and the like.
It becomes a discrete value such as 66.67 Hz (when the sampling frequency is 10000 Hz).

Therefore, the following sounds could not be successfully synthesized.

Sound of chime When the marimba is hit in order of "do", "mi", "so" and "do", the next "mi" sounds before the first "do" is completely attenuated. You can play beautiful chords by mixing the sounds of. To create this chord, two or more sound sources with different fundamental frequencies are required.

DTMF (Dual Tone Mul)
ti Frequency) A DTMF signal (dial tone signal), which is widely used for selecting a telephone line such as a push button type telephone, is composed of two sine waves having no harmonic relationship. Either 697, 770, 852, or 941 Hz is used as the lower frequency, and 1209, 1 as the higher frequency.
Any of 336, 1477, 1633 Hz is used.
Therefore, in order to create this signal, it is necessary to accurately create a plurality of sine waveforms (within ± 1.5% of the predetermined frequency).

Therefore, as long as the conventional voice synthesizer is used, a dedicated circuit is required in addition to the voice synthesizer in order to generate these sounds. FIG. 5 shows a circuit configuration when a musical instrument sound can be generated in addition to the synthetic voice and is connected to a telephone line.

In FIG. 5, reference numeral 201 is a voice synthesizer. In this configuration example, the voice synthesizer 202 of FIG.
DSP (Digital Signal Procedures)
sor: digital signal processor). The characteristic parameters to be given to the speech synthesizer 202 are read from the ROM (Read Only Memory) 204 to C
A PU (Central Processing Unit) 203 reads the data, edits and processes it as needed, and supplies it to the speech synthesizer 202. An embodiment is also possible in which the ROM 204 stores characteristic parameters of the entire sentence to be speech-synthesized, and the characteristic parameters are given to the speech synthesizer as they are according to a speech synthesis instruction.

Further, the ROM 204 stores characteristic parameters in units of syllables and phonemes, and the characteristic parameters are connected according to the order of syllables and phonemes appearing in the sentence to be speech-synthesized and then given to the speech synthesizer. It is also possible to perform voice synthesis using a so-called voice rule synthesis method. A RAM (Random Access Memory) 205 is used as a work area for processing performed by the CPU 203. The voice waveform calculated by the voice synthesizer 202 realized by the DSP 202 is represented by a digital code. A DA (digital-analog) converter 206 converts the voice waveform into an analog signal. Reference numeral 207 is an LPF (low-pass filter) for anti-aliasing, but it may be unnecessary depending on the characteristics of the analog circuit connected to the subsequent stage.

Reference numeral 210 is a musical instrument sound generating circuit, for example,
Create a chime sound by tapping the marimba in order of "do", "mi", "so", and "do".

Reference numeral 211 is a circuit for adding analog signals. As a result, the signal line from the voice synthesizer 201 and the signal line from the musical instrument sound generation circuit 210 can be integrated.

220 is an NCU (Network Co)
control unit (network control unit), and is used for selecting a party to talk on a telephone line.
The selection of the call destination is generally performed by the connection / disconnection of the telephone line 224 (dial pulse) or the DTMF signal. It is provided. The line transformer 223 is a telephone line 224,
This is for isolating the signal generation circuit in the device from direct current, and has the function of transmitting only AC signals. 224 is a circuit for adding analog signals. As a result, the signal line from the circuit 211 and the signal line from the DTMF signal generation circuit 221 are combined into a single line transformer 224.
Can be supplied to.

As described above, as long as the conventional voice synthesizer 202 is used, the instrument sound generating circuit 210 and the DTM are generated.
The F signal generation circuit 221 is separately required and the analog addition circuits 211 and 224 are required, which hinders downsizing and cost reduction of the entire apparatus.

Therefore, an object of the present invention is to provide a voice synthesizer capable of generating a musical instrument sound or a DTMF signal even in the voice synthesizer and reducing the size of the device.

[0024]

In order to achieve such an object, the present invention provides a speech waveform generator capable of generating at least one or more first waveforms of vocal cord sounds, and at least a sine waveform shape. The present invention is characterized by comprising a sine waveform generation unit capable of generating one or more types of second waveforms, and a waveform batch unit that collectively outputs the generated first and second waveforms.

[0025]

In the present invention, focusing on the fact that in the configuration of the prior art, a plurality of types of voice waveforms are put together when a vocal cord sound is generated, a sine waveform creating section is built in the device to share the waveform batch section. By doing so, it is possible to reduce the number of parts in the waveform batch section.

[0026]

Embodiments of the present invention will now be described in detail with reference to the drawings.

The basic configuration of the speech synthesizer according to the present invention is shown in FIG.
Shown in.

The voice synthesizer 30 of the present embodiment has, in addition to the voice waveform generator 31, sine wave generators 32 and 33 and a waveform collective unit 3.
4 is included in the voice synthesizer, and a plurality of waveforms obtained from these generation units are collectively output as one waveform by the waveform batch unit 34.

In the speech synthesizer 30, in addition to the synthesized speech from the speech waveform generator 31, sine waveform generators 32, 3
The sine waveform created by 3 can be output as a collective waveform in the collective waveform unit 34. This allows
In addition to outputting synthetic speech, it is possible to output an accurate sine waveform, and by outputting by combining multiple sine waveforms,
Create chords and DTMF signals.

The waveform batch unit 34 includes a voice waveform generator 31,
Collective processing is performed by adding and subtracting the waveforms created by the sine wave generators 32 and 33. As another method, the sine waveforms generated by the sine wave generation units 32 and 33 may be added or subtracted from each other, and the collective processing may be performed by selecting only one of the voice waveform and the sine waveform.

When the waveform lumping process adds or subtracts the waveforms created by the voice waveform generating unit 31 and the sine wave generating units 32 and 33, it is possible to simultaneously output a sound using a sine wave together with the synthesized voice. it can.

FIG. 2 shows a concrete configuration example of the circuit of FIG. In this example, as the voice waveform generator 31, the same one as the conventional circuit of FIG. 2 is used, and three sine waveform generators are added thereto to form a voice synthesizer.

In the figure, each of the elements 5 to 15 and 101 to 146 has the same function and configuration as the parts denoted by the same reference numerals in FIG. 21 to 23 are sine waveform generators, which generate a sine waveform having a predetermined frequency. Reference numerals 24-26 are amplitude control units that adjust the amplitude of the sine wave. That is, it is possible to adjust to a loud sound or a quiet sound. It is necessary to overlap two or more waveforms for a chord, but in this embodiment, since three sine waveforms can be overlapped, a more beautiful chord can be created.

The sound of the marimba can be simulated by a sine waveform whose amplitude decays exponentially. When producing sounds in order, it is sufficient to use a plurality of sine wave generators in order. Therefore, when you hit "do", "re", "mi", and "fa" in that order, you can simulate as follows.

The first "do" is, while the sine wave generator 21 creates a sine wave of that frequency, the amplitude controller 24
The sound of "do" is simulated by gradually decreasing the amplitude.

While creating the sound of “do” while attenuating the sound of “do”, the amplitude of the next “re” is gradually decreased by the amplitude control unit 25 while creating the sine wave of that frequency by the sine waveform generation unit 22. Simulate by doing.

Similarly, while the sound of "do" and "re" is being created while being attenuated, the next "mi" is the sine waveform generating section 23.
While creating a sine wave at that frequency, the amplitude control unit 26
Then, it is simulated by gradually reducing the amplitude.

In order to produce the final "Fa", the sine wave generator 21 creates a sine wave of the frequency of "Fa" while attenuating the sounds of "Le" and "Mi" while controlling the amplitude. The part 24 simulates by gradually reducing the amplitude.

Further, since two sine waves may be overlapped in the DTMF signal, two of the three sine waveform generators are used to create a sine wave having a predetermined frequency and amplitude, and the adder 1
It is sufficient to add 5 and output.

When the adder 15 is used as the waveform collective section 34, a sine wave sound can be simultaneously output together with the synthesized voice. For example, in music education, while producing a single sine wave or multiple sine waves (chords),
A comment such as "This is a sound of 440 Hz" can be given by voice.

As another embodiment of the waveform collective section 34,
An embodiment is possible in which the sine waves are added, but the voice and the sine wave are switched, and only one of them is output. In this case, it is not possible to superimpose the sine wave and the sound and output them. However, in the case of implementing the present invention by programming a general-purpose DSP, due to restrictions on processing time, working memory area, etc.
In some cases, sine waves cannot be generated at the same time. In such cases, a switchable embodiment is suitable.

It is also possible to perform the above circuits by a digital processing circuit. FIG. 3 shows an application example using a digital processing voice synthesizer. In the figure, the same parts as those of the circuit components of FIG. 5 are designated by the same reference numerals.

FIG. 3 shows an application example in which the voice synthesizer 202 'has a DTMF signal generating function among the functions constituting the NCU for selecting a telephone line. 202 'is
The voice synthesizer according to the present invention can output a DTMF signal, a music, a buzzer sound, etc., which can be produced by combining a sine wave, in addition to the synthesized voice, under the control of the CPU 203. Therefore, music or the like can be output without using the musical instrument sound generating circuit 210 of FIG. 5, which has been conventionally required. Similarly, when connecting to a telephone line, DT
The MF signal generation circuit 221 and the analog addition circuit 224 are not required and can be directly connected to the line transformer 223.

[0044]

As described above, according to the present invention, in addition to synthetic speech, it is possible to output sounds that can be expressed by superposition of sine waves. Therefore, an electronic device using the speech synthesizer of the present invention can be used. There are remarkable effects in improving the expression means, simplifying the circuit configuration, and reducing the cost.

[Brief description of drawings]

FIG. 1 is a block diagram showing a basic configuration of an embodiment of the present invention.

FIG. 2 is a block diagram showing a specific circuit configuration of an embodiment of the present invention.

FIG. 3 is a block diagram showing a circuit configuration of another embodiment of the present invention.

FIG. 4 is a block diagram showing a circuit configuration of a conventional speech synthesizer.

FIG. 5 is a block diagram showing a circuit configuration of another conventional speech synthesizer.

[Explanation of symbols]

 15 Adder 31 Voice Waveform Generating Unit 32 Sine Wave (Shape) Generating Unit 34 Waveform Collecting Unit 202 DSP

Claims (1)

[Claims] 1. A voice waveform generation unit capable of generating at least one type of first waveform for vocal cord sounds, a sine waveform generation unit capable of generating at least one type of second waveform having a sine waveform shape, A voice synthesizing apparatus comprising: a waveform batch unit that collectively outputs the generated first waveform and second waveform.