JPH05241600A - Method and device for recording phoneme and reproducing speech - Google Patents

Abstract

PURPOSE: To provide a method and device for recording phoneme and reproducing speech by which data processing can be performed by reducing and compressing the quantity of data and, at the same time, holding phonemes and sound pressure thickness data. CONSTITUTION: The device for recording phoneme and reproducing speech is constituted by connecting a microphone 101 which converts a speech into a speech voltage, an amplifier 102 which amplifies the voltage from the microphone 101, a gain controller 102 which adjusts the magnitude of the amplified voltage outputted from the amplifier 102 to a fixed value, a microcomputer 104 which executes a program for extracting phoneme and sound pressure data at the controller 103, a data storing machine 105 which stores the phoneme and sound pressure data extracted by means of the microcomputer 104, a storing state indicator 106 which indicates the data storing state of the machine 105, and a computer 107 which confirms the data storing state of the machine 105 from the indicator 106 and, when stored data exist, causes the data to be stored on a diskette or its own memory to each other.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention extracts the time (phoneme) until the direction of the voltage is changed and the voltage (sound pressure) at the time when the voltage is changed, and records it. The present invention relates to a phoneme recording and voice reproduction method and device capable of generating and outputting a voice increase or decrease waveform in time (phoneme) and outputting a difference value obtained by obtaining a difference.

[0002]

2. Description of the Related Art First, a phoneme referred to in the present invention is a time (T) in which a voltage direction is kept constant, which is a time in which a voltage converted through a microphone into a minimum unit of sound increases or decreases. means.

In the present invention, the sound pressure
Is a vibrational pressure change when a sound wave moves in a medium (in air), and means a point (maximum value voltage) at which the direction of the voltage is changed.

Generally, most of the recording methods used at present use a voltage recording method for extracting and recording the sound pressure of voice, but the voltage recording method is largely pulse code modulation (Pulse Code). Modulation: PCM) and differential tracking pulse (adaptive differential) code modulation (AD)
PCM) is known.

That is, in the former voltage recording and reproducing method by the pulse code modulation method, the voltage of the sound converted by the microphone is sampled once at a fixed time to extract each peak value voltage, and each extracted peak value voltage is extracted. In this method, a voltage is converted into a digital numerical value and stored in a memory, and the numerical value stored in the memory is further converted into an analog voltage and reproduced as voice through a speaker. The latter differential tracking pulse code modulation method is considered to be a kind of method for compressing data recorded by the pulse code modulation (PCM) method.

That is, the differential tracking pulse code modulation method is a method of compressing the 8-bit voltage of pulse code modulation (PCM) data into 4 bits as shown in Table 1 below, and the first voltage of the 4 bits increases the next voltage. Is a decrease or decrease, and the remaining 3 bits are for displaying the increase / decrease.

[0007]

[Table 1]

However, in such a method, although the sound quality is good, the data capacity is consumed so much that it can be applied only to a partly short voice, and only the voltage value is used for the data. Therefore, the voice assist control is not possible at the time of synthesis, and it is the situation that it is used only for sentence-by-sentence synthesis.

[0009]

Therefore, in the present invention, in consideration of the above-mentioned drawbacks of the prior art, the time (phoneme portion) until the direction of the voltage is changed and the voltage is converted at the time of voice recording. It is divided into the sound pressure part of the highest value voltage at the time of being used, the two extracted data are utilized for the recording data, and the difference between the previous voltage and the current voltage is obtained at the time of voice reproduction output. To reduce the amount of data compared to general recording methods and at the same time retain the phoneme and sound pressure thickness data by generating a waveform that increases or decreases the voice at a certain time (phoneme) and can output it at a certain time (phoneme). It is an object of the present invention to provide a phoneme recording method and a voice reproducing method and a device thereof that can process data.

[0010]

The concrete means for achieving the above-mentioned object of the present invention will be described below with reference to the accompanying drawings.

FIG. 1 shows a block diagram of a phoneme recording apparatus according to the present invention. A microphone 101 for converting voice into a sound voltage, an amplifier 102 for amplifying the voltage converted by the microphone, and an output from this amplifier. A gain controller 103 for adjusting the magnitude of the amplified voltage to a constant value, a microcomputer 104 for executing a program for extracting phoneme and sound pressure data by the gain controller, and a phoneme and sound pressure data extracted by the microcomputer. A data storage device 105 to be stored, a storage state display device 106 for displaying the data storage state of the data storage device 105, and a diskette or its own memory if the stored state data of the storage state display device is confirmed. It is configured by interconnecting a computer 107 that is stored in the computer.

[0012]

Operation: That is, the voice voltage output from the microphone 101 is amplified, and after the gain control, the phoneme and sound pressure data are extracted by the microcomputer. The time until the direction of the voice voltage is changed (phoneme part ) And the sound pressure part of the highest value voltage at the time when the voltage is converted, the two extracted data are stored in the data storage device. Then, at the time of voice reproduction output, the difference between the previous voltage and the current voltage is obtained, and a voice increase or decrease waveform can be generated at this difference value and time (phoneme) and output at a fixed time (phoneme). The data can be processed by holding the phoneme and sound pressure thickness data at the same time as the reduction.

[0013]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First, the present invention will be summarized. The voltage (sound pressure) at the time when the direction of voltage is changed is extracted and recorded, and the difference between the extracted voltage and the current voltage is obtained. The present invention relates to a phoneme recording and voice reproduction method and device for generating a voice increase or decrease waveform with different difference values and outputting data that can be processed with less data amount than recorded data in a fixed time (phoneme). This is to solve the problems of the prior art that could not be put to practical use due to difficult problems such as data processing, noises, word feelings, pitch changes, etc., which occur when sounds are continuous.

Therefore, according to the present invention, at the time of voice recording, the phoneme portion of the time until the voltage direction is changed,
When the direction of the voltage is changed, it is divided into the sound pressure part of the highest value voltage, and the two extracted data are used for the recorded data, and the difference between the voltage of the recorded data and the current voltage at the time of voice reproduction output. By calculating the difference value and the phoneme to generate and output a voice increase or decrease waveform, the data amount is reduced compared to the recorded data, and
It is possible to solve the problem that the voltage difference of the voice connection part occurs, and the data in which the phoneme and the sound pressure are separated can be controlled by software such as sound compression, expansion, and sound volume.
There is an advantage that it can be applied to next-generation support computer development, such as enabling reduction and expansion of sound length (“car” of Hangul sound is “ka” and short time is “ka”).

The embodiment will be described in detail below. The microcomputer 104 shown in FIG. 1 is configured as follows. This microcomputer 104 is an A / D converter 110 that converts an input analog voltage into a digital voltage, and a noise that removes noise by monitoring the change and fluctuation rate of the value output via this A / D converter 110. A denominator 111, a voltage direction finder 112 for detecting that the voltage output from the noise eliminator 111 is converted into an increase-decrease or an increase-decrease, and an output of the noise eliminator 111. For a certain period of time or more, the silent direction discriminator 113 for inspecting the fluctuation state in the voltage direction and transmitting a silent message, the voltage direction finder 112 and the silent silence discriminator.
It is connected to the output side of 113, and the silent message and the inspected signal in the voltage direction are input, and the time (phoneme) in the same voltage direction is counted, and the time (phoneme) in the same voltage direction is counted. ) Is connected to the output terminal of the voltage direction finder 112 and the silence discriminator 113, and the voltage determined at each point where the voltage direction is changed by the voltage direction signal output from these. Is connected to the output terminals of the timer 114 and the voltage discriminator 115, and the phoneme and sound pressure data processed from these are stored depending on whether the external storage state indicator 106 is stored or not. The device 105 includes a data transmitter 116 that controls transmission determination or transmission suspension.

FIG. 2 is a block diagram showing the configuration of the audio output device according to the present invention. A computer 107 for transmitting data in its own memory or diskette according to a data storage display state, and a data stored in the computer 107 are stored. A data storage 205 and a storage status indicator 20 for displaying the data storage status of this data storage 205
6 and a microcomputer 200 for executing a series of audio reproduction processing programs for reading the data in the data storage 205 in which the data of the computer 107 is stored and dividing the phoneme and sound pressure data to obtain the difference in sound pressure. And a noise eliminator 201 that filters the high frequency of the staircase from the output voltage processed by this microcomputer 200 and finally converted to an analog voltage, and the output voltage from which the noise high frequency is removed through this noise eliminator 201 An amplifier 202 for amplifying a voice voltage that can be recognized visually and a speaker 203 for converting the voice voltage amplified by the amplifier 202 into a voice are connected to each other.

Further, inside the microcomputer 200 of the voice output device, a data separator 210 for separating the output data of the data storage 205 into phoneme and sound pressure data, and the sound pressure extracted and separated by this data separator. The voltage difference extractor 211 that calculates the difference between the data and the current sound pressure, and this voltage difference extractor 211
A triangular wave generator 212 for generating data of a triangular wave from the sound pressure position of the previous stage calculated in step 3 to the current sound pressure position;
High-frequency remover 213 for filtering out high-frequency components contained in the triangular wave generated from each wave generator 212
A voice waveform converter 214 for receiving the outputs of the triangular wave generator 212 and the high frequency remover 213 and shaping the waveform into a sine wave of a digital voice signal, and the voice waveform converter 214.
And a D / A converter 215 which is connected to the output terminal of and which converts the output data thereof into an analog signal.

FIG. 3 is a circuit diagram showing an embodiment in which the phoneme recording device shown in FIG. 1 and the voice output device shown in FIG. 2 are combined. A microcomputer 104 of the phoneme recording device and the voice output device is shown.
Are connected together to form a block. Such a microcomputer 104 has analog signal input terminals (A / D, D /
A), output ports (PA1 to PA7) for sending data to the computer 107, input ports (PB0 to PB7) for receiving data from the computer 107, a data reservoir 105 and a reservoir status indicator 106. It has control ports (PD0 to PD1) for controlling.

Between the computer 107 and the microcomputer 104, a plurality of analog switches 301 for grasping the presence / absence of storage in the shared data storage 105 in the computer 107.
And 302, and a gate array 300 for controlling the shared data storage 105 and the storage status indicator 106 in the computer 107.

The operation and effect of the device according to the present invention thus constructed will be described with reference to FIGS. 4 to 6. First, as shown in FIG. 1, when the voice is converted into a voltage in the microphone 101, the amplifier The amplifier 102 amplifies the voltage converted by the microphone 101, and is adjusted by the gain controller 103 so that the magnitude of the amplified voltage is constant. The output voltage from the gain controller 103 is the analog input terminal ( A / D).

Further, the microcomputer 104 converts the input voltage into a digital signal by the A / D converter 110 in the microcomputer 104 and outputs it, and the noise eliminator 111 changes and changes the output value of the A / D converter 110. The rate is monitored, and the noise caused by the change is removed. At the same time, the voltage direction finder 112 determines whether the output voltage obtained through the noise eliminator 111 is changed from increase to decrease or decrease to increase voltage. A signal is sent to the discriminator 115, and the silence discriminator 113 is a kind of timer 11
It is checked whether or not the voltage direction is changed for a certain period of time and the message is sent to the timer 114 and the voltage discriminator 115 as a silent message.

As a result, the timer 114 transmits the time during the same voltage direction, that is, the phoneme data, to the data transmitter 116.
At the same time, the voltage discriminator 115 sends the voltage at the point where the voltage direction changes, that is, the sound pressure data to the transmitter 116.

At this time, since the data transmitter 116 stores the phoneme and sound pressure data in the external data storage device 105, the contents of the storage device status display device 106 are confirmed and the data storage device 105 is confirmed.
If there is no stored data in (data stored by computer 107), each data input to data transmitter 116 is transmitted to data reservoir 105.

That is, as a method of confirming the contents of the reservoir status indicator 106 for transmitting the data of the microcomputer 104 to the computer 107, the reservoir status indicator 106 is first read as shown in FIGS. 3 and 4. Port for (PD0
), The output (Q) state of the reservoir state indicator 106 is discriminated.

In this judgment, if the output (Q) of the reservoir status indicator 106 is at "low" level, the data reservoir 10
It is determined that no data is stored in 5. At this time, the data of the microcomputer 104 is sent to the data storage 105 via its output ports (PA1 to PA7), and its control port (PC0) is set to the "high" level and then to the "low" level. After conversion, the data of the microcomputer 104 is stored in the data storage 105,
The output terminal (Q) of the reservoir status indicator 106 is kept in the "high" state and the time value of the internal timer 114 is set to 0 (ZE.
RO) to clear.

On the other hand, when the computer 107 designates an arbitrary address (A0 to An), the output (Q1) of the gate array 300 becomes "high" level and the analog switch 30.
With 1 open, check the output (Q) of the reservoir status indicator 106. At this time, if the output (Q) of the reservoir status indicator 106 is at "high" level, the computer 10
7 determines that the data is stored in the data storage device 105, and specifies another address to specify the gate array 300.
After making the output end (Q2) of the "high" level, "low"
Switch to a state. As a result, the stored data in the data storage 105 becomes readable by the computer 107, and the phoneme recording data by the microcomputer 104 is stored in the computer.
It can be stored on its own memory or on a separate diskette.

Next, a description will be given of a process of extracting data in the memory of the computer 107 and outputting it as a voice in a state where the phoneme and sound pressure data of the voice are extracted and stored in the memory of the computer 107 by the above process.

That is, a computer as shown in FIG.
In order to transfer the data stored in its own memory or diskette to the microcomputer 104, the computer 107
The process of inspecting the output terminals (Q) of the data storage unit 205 and the storage unit status display unit 206 and storing the stored data of the computer 107 in the data storage unit 205 according to the logic values thereof is performed in the same manner as above. To do.

That is, while the storage data of the computer 107 is stored in the data storage unit 205, the microcomputer 104
As shown in FIG. 6, the data in the data storage 205 is read, and the data separator 210 inside the data storage 210 separates the data into phoneme data and sound pressure data.

At this time, if the difference between the pre-stage sound pressure of the separated sound pressure data and the current sound pressure is calculated by the voltage difference extractor 211 and the difference is output, the triangular wave generator 212 is moved to the preceding stage. Triangular wave data up to the current sound pressure position is generated at the sound pressure position. The high frequency component generated during the generation of the triangular wave data, that is, the high frequency component whose direction is changed is removed by the high frequency remover 213. The triangular wave generator 212 and the high frequency remover
A voice waveform converter 214 connected to the output end of 213 removes the high frequency contained in the triangular wave and outputs a sine wave.

When this is further converted into an analog voltage by the D / A converter 215 and sent to the noise eliminator 201, the noise eliminator 201 outputs the analog voltage obtained by eliminating the high frequency of the step wave of the converted analog voltage. Will be the amplifier
Reference numeral 202 is a sound which is converted into an audible sound through the speaker 203 and is output.

The voice output process of the voice output device is as follows.
As shown in FIG. 5, the presence / absence of stored data in the data storage unit 205 is determined, and when the data is stored, the data is read, and the phoneme data and the sound pressure data are separated from the read data to separate the preceding sound pressure and the current sound. Calculate the difference from the pressure. After this is generated into a triangular wave, the process of further converting it into sine wave data and outputting it to an analog value is continuously repeated, and the data of the computer 107 is transmitted to the microcomputer 104, and the computer expression sound is actually generated. Can be output to the voice.

In the above description, the data reservoirs 105, 205 are latches as in FIG. 3, and the reservoir status indicators 10
6 and 206 are replaced by flip-flops, and computer 107 means a real computer slot.

[0034]

As described above, according to the present invention, since the phonemes and sound pressures of the voice can be extracted and recorded, there is an advantage that it can be applied to the development of the next-generation voice assist computer.

[Brief description of drawings]

FIG. 1 is a block diagram of a phoneme recording apparatus according to an embodiment of the present invention.

FIG. 2 is a block diagram of the audio output device.

FIG. 3 is a circuit diagram showing a specific embodiment of the present invention.

FIG. 4 is data (phoneme, sound pressure) of the method according to the embodiment of the present invention.
The flowchart explaining an extraction operation.

FIG. 5 is a flowchart similarly illustrating a voice reproduction output operation.

FIG. 6 is a waveform chart showing an example of a sound waveform change for explaining the phoneme extraction method according to the present invention.

[Explanation of symbols]

101 ... Microphone, 102, 202 ... Amplifier, 103 ... Gain controller, 104 ... Microcomputer, 105, 205 ... Data storage, 106
, 206 ... Reservoir status indicator, 107 ... Computer, 110
... A / D converter, 111, 201 ... Noise eliminator, 112 ... Voltage direction finder, 113 ... Silent discriminator, 114 ... Timer, 11
5 ... Voltage discriminator, 203 ... Speaker, 210 ... Data separator, 211 ... Voltage difference extractor, 212 ... Triangular wave generator, 213 ...
High frequency remover, 214 ... Voice waveform converter, 215 ... D / A converter, 300 ... Gate array, 301, 302 ... Analog switch.

Claims (5)

[Claims] 1. A microphone (101) for converting into a voice voltage, a gain controller (103) for adjusting the voltage converted by the microphone so that the magnitude of an amplified voltage is constant, and a gain controller for adjusting the gain. A microcomputer (104) which executes a program for converting the detected voice voltage into a digital voltage and extracting phoneme and sound pressure data from the voice voltage.
A data storage device (105) for storing phoneme and sound pressure data extracted from the microcomputer; a storage state display device (106) for displaying a data storage state of the data storage device (105); A phoneme recording device comprising a computer (107) for confirming the storage condition of a display unit and storing the stored data in a diskette or in its own memory, which is connected to each other. 2. An A / D for converting an input analog voltage into a digital voltage inside the microcomputer (104).
A converter (110) and a noise eliminator (111) for observing a change and a fluctuation rate of a value output through the A / D converter (110) to eliminate noise.
A voltage direction finder (112) for detecting that the voltage output from the noise eliminator (111) is converted from increase to decrease or decrease to increase, and A silence discriminator (113) which receives and outputs a silent message by inspecting the fluctuation state in the abnormal voltage direction for a certain period of time, and is connected to the output ends of the voltage direction finder and the silent discriminator, and a silent message and A timer (114) for inputting an inspected voltage direction signal, counting time (phonemes) in the same voltage direction, and outputting time (phoneme) in each period having the same voltage direction; It is connected to the output end of the detector and silence detector, and the voltage (sound pressure) at the point where the voltage direction is changed is determined by the voltage direction signal output from these, and the voltage direction is changed at each point where the voltage direction is changed. Voltage discriminator that sends out the discriminated voltage
(115) is connected to the output end of the timer and the voltage discriminator, and the phoneme and sound pressure data processed by these are stored in the external storage state indicator (106) according to whether the storage indicator is present or not.
The phoneme recording device according to claim 1, further comprising a data transmitter (116) for controlling transmission determination or transmission suspension in (105). 3. A computer (107) for transmitting data in its own memory or diskette according to a data storage state display, a data storage (205) for storing data transmitted from the computer (107), and the data. The storage device status indicator (206) for displaying the data storage condition of the storage device (205) and the phoneme and sound pressure of the data storage device (205) in which the data of the computer (107) is stored are read. A microcomputer (200) that executes a series of audio reproduction processing programs that divides the data and finds the difference in sound pressure, and filters the high frequency of the staircase wave from the output voltage processed by the microcomputer and finally converted to an analog voltage. A noise eliminator (201), and an amplifier (202) that amplifies the output voltage from which the noise high frequency is removed through the noise eliminator to a sound voltage that can be perceptually recognized, A voice reproduction output device comprising: a speaker (203) for converting the voice voltage amplified by the amplifier (202) into voice, and the speaker (203). 4. A data separator (210) for separating the output data of the data storage (205) into phoneme and sound pressure data inside the microcomputer (104), and the data separator extracted and separated by the data separator. A sound pressure difference extractor (211) for calculating the difference between the sound pressure data and the current sound pressure, and data from the sound pressure position of the previous stage calculated by the sound pressure difference extractor to the current sound pressure position is converted into a triangular wave. A triangle wave generator (212), a high frequency remover (213) for filtering out high frequency components contained in the triangle wave generated by the triangle wave generator, and the triangle wave generator and the high frequency remover. A voice waveform converter (214) for receiving the respective outputs of the above and shaping the waveform into a sine wave of a digital voice signal, and the voice waveform converter (214) is connected to the output end thereof and converts the output data thereof into an analog signal. / A converter (21
5. The phoneme reproduction output device according to claim 3, further comprising: 5. A time (phoneme) until a change direction of sound pressure (voltage) is changed in voice recording.
The phoneme recording method is characterized in that the voltage (sound pressure) at the point where the voltage is changed can be recorded as data.