JP3266761B2 - Speech rate converter and method of using speech rate converter - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech speed conversion device which is a digital voice signal processing device for converting the speed without changing the pitch of voice, and more particularly, for use in connection with a recording / reproducing device such as a tape recorder. According to the method.

[0002]

2. Description of the Related Art In recent years, attempts have been made to perform hearing aid over the entire auditory system including a decrease in language processing speed by changing only the speed without changing the pitch of speech by digital signal processing. Such speech rate conversion technology is described, for example, in the Proceedings of the Acoustical Society of Japan (March 1994, 2-
4-7) It is described in detail in "Evaluation of speech rate converted speech by hearing-impaired persons" and "The Real-Time Speech Rate Conversion Type Listening System" of the Acoustical Society of Japan (1994, Vol. 50, No. 7, P509-P520). Particularly recently, a portable speech speed conversion device has been developed in which the speech speed conversion function is housed in a portable size device. This device is described in "A portable digital speech-rate converte
r and its evaluation by hearing-impairedlistener
s ", in Proc. of Int. Conf. on Spoken Language Pro
cessing (ICSLP94,) pp.2055-2058, Yokohama, Sep.
1994.

FIG. 8 shows the operation of the conventional speech speed converter. In the conventional speech speed conversion device, first, an input voice signal is converted into a digital value by A / D conversion means 10. The digitized voice is alternately input to two input buffers 30 having a fixed time length (hereinafter, referred to as a frame). While inputting to one input frame buffer, a speech speed conversion process 40 for expanding and contracting the time axis of the audio without changing its pitch is performed on the audio data of the input buffer already input. Apply. Then, the processing result is recorded in the output ring buffer 50.
The output ring buffer 50 is a memory in which addresses are used in a ring so that the next address after the last address is the first address. Independently from the writing of the speech speed conversion processing result to the output ring buffer 50, the output ring buffer 50
By taking out the processed data one by one and simultaneously outputting the audio signal through the D / A conversion means 60,
The speech speed conversion process and the output of the conversion result proceed simultaneously, and real-time speech speed conversion is realized. Therefore, the output sound is output with a delay time with respect to the input sound, and the delay time spreads with time. That is, the listener can hear the sound slowly but slowly from reality. This speech speed conversion operation is controlled by the user through a slow button 160 provided on the device, and the speech speed conversion is performed while the slow button is pressed, and the voice speed conversion is performed while the slow button is not pressed. In this case, the input voice is output as it is. In FIG. 8, this operation is represented by the switch means 20 controlled by the slow button. The above operation is realized by the device configuration shown in FIG. The speech speed conversion process shown in FIG. 8 is realized by software operating on the microprocessor 110. The input buffer 30 and the output ring buffer 50 shown in FIG. 8 are secured in the external RAM 120 connected to the processor, and the configuration in FIG. 8 is realized by address control in a program. The control information from the slow button 160 is provided through the control circuit 130 and the I / O circuit 112 on the microprocessor, and the above-described switch operation is realized by software. In FIG. 9, the A / D converter 111 and the D / A converter 114 are also provided inside the microprocessor, and the control of the operation is represented by a program. And the software is a microprocessor
It is stored as a program in ROM 115 inside 110.

[0004]

SUMMARY OF THE INVENTION
Without changing the pitch of the voice, that is, the pitch of the voice,
It is effective for learning and training a foreign language because the voice can be made slow only at the speed. In actual language learning, audio recorded on a tape or the like is often used as a teaching material.

[0005] However, in the above-mentioned conventional apparatus, a speech speed conversion device is connected to a recording / reproducing device such as a tape recorder, and a means necessary for speech speed conversion of a long-time voice without interruption is considered. Had not been. That is, in the above-mentioned conventional speech speed conversion device, by outputting the speech after the speech speed conversion while recording it in the memory, the speech that is slowly converted while the time delay from the input speech is widened is output. For this reason, voice information having a length corresponding to the time difference between input and output is always recorded in the memory, and if this time difference exceeds the memory capacity, there is a problem that the speech speed cannot be further continuously converted. .

An object of the present invention is to connect a speech speed converter to a recording / reproducing device such as a tape recorder so that long-time recorded / reproduced audio can be continuously converted in speech speed without interruption due to the limitation of memory capacity. It is to provide a means of.

[0007]

SUMMARY OF THE INVENTION It is an object of the present invention to provide a speech rate converter which takes in an input speech from a recording / reproducing apparatus, extends a time axis without changing the pitch of the input speech, and outputs a speech rate converted speech. The apparatus is achieved by a speech speed conversion device having control signal generating means for controlling a reproduction operation of the recording and reproduction device.

[0008] The control signal generating means is configured to control the reproduction and the pause operation of the recording / reproducing device in accordance with the volume of the voice data stored in the memory inside the speech speed converting device.

In the present invention, the control signal for controlling the reproducing operation of the recording / reproducing device is output in accordance with the amount of audio data stored in the memory means inside the voice speed converting device along with the voice speed converting process. Is done.

On the other hand, the speech speed conversion device of the present invention comprises an A / D conversion means, an input buffer constituted by a plurality of frame buffers, and a speech signal which expands and contracts the time length of the audio signal in frame units without changing the pitch. Speed conversion processing means, an output buffer having a ring buffer configuration in which addresses are used in a ring,
A monitoring means comprising D / A conversion means, and monitoring the amount of audio data stored in the output ring buffer;
Control signal generating means for generating a control signal for controlling a reproduction operation of the recording / reproducing means, wherein the monitoring means performs the A / D conversion in accordance with the amount of audio data stored in the output buffer. The operation of the control means and the operation of the control signal generating means are controlled.

Further, the speech speed conversion device of the present invention comprises an A / D conversion means, an input buffer having a ring buffer configuration in which addresses are used in a circular fashion, and a reproduction address control means for controlling an address read from the input buffer. A speech speed conversion processing unit that expands and contracts the time length of the audio signal in frame units without changing the pitch, an output buffer composed of a plurality of frame buffers, and a D / A conversion unit, and the input ring Monitoring means for monitoring the amount of audio data stored in the buffer, and control signal generating means for generating a control signal for controlling the reproduction operation of the recording / reproducing device, wherein the monitoring means The operation of the A / D converter and the operation of the control signal generator are controlled according to the amount of audio data to be stored.

In the control by the monitoring means, when the amount of audio data stored in the input or output ring buffer approaches the maximum amount that can be held in the input or output ring buffer, the reproduction of the recording / reproducing device and the A / D When the operation of the conversion means is temporarily stopped, the speech speed conversion process is continued, and the stored voice data is output. When the amount of stored voice data in the input or output buffer approaches 0, the reproduction of the recording / reproducing apparatus is performed again. It operates to start the operation and the operation of the A / D conversion means.

At this time, the speech speed conversion device has a repeat function for repeatedly reproducing the past voice, and when the repeat function is executed, the monitoring means is configured to control the recording / reproducing device and the A / D conversion means. Is temporarily stopped to prevent new input audio data from entering the input ring buffer during the repeat operation.

In the present invention, the control signal for controlling the operation of the recording / reproducing apparatus is a wireless remote control signal.

Further, the speech speed converter of the present invention comprises a low-pass filter for anti-aliasing at the input and output of an analog audio signal, an A / D converter,
A / A converter, a microprocessor, an external memory, a control button, and a control circuit for inputting the state of the control button to the microprocessor, and a control for controlling the recording / reproducing device. The above-described series of operations is realized by a software program on the microprocessor having means for generating a signal.

Further, according to the utilization method of the present invention, a sound is reproduced by a recording / reproducing apparatus having two output channels, the same sound signal is output from both of the two output channels, and two separately prepared input signals are provided. In a system for duplicating and recording voice by a recording / reproducing apparatus having channels, a voice speed for executing a voice speed conversion process for expanding and contracting the time length between signal paths of one channel without changing the pitch of the voice signal. There is also a method of using a speech speed conversion device in which a conversion device is provided, and only one channel voice is subjected to speech speed conversion processing and recorded, and the other channel is recorded with a normal speed voice.

[0017]

[0018]

According to the present invention, the speech speed is converted by the speech speed converter connected to the recording / reproducing device. Inside the speech speed conversion device, while recording the result of performing speech speed conversion processing on the input audio signal in the buffer memory,
Output at the same time at the sampling interval. As a result, voice data having a length corresponding to the time difference between the input and output voices is stored in the memory inside the speech speed conversion device. The control signal generating means acts to temporarily stop the reproducing operation of the recording / reproducing apparatus when the accumulated audio data capacity becomes substantially equal to the memory capacity. The sound stored in the memory is output at any time, and when a sufficient free space is generated in the memory, the sound recording / reproducing device operates again to perform the reproducing operation. As a result, when the speech speed conversion device is connected to the recording / reproducing device and used, the speech speed conversion function can be continuously used for a long time regardless of the memory capacity.

[0019]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below in detail with reference to embodiments.

FIG. 1 is a block diagram of the first embodiment of the present invention. In this embodiment, the function of reproducing the sound of a recording medium such as a CD or a cassette tape recorder is built in the recording / reproducing apparatus 200. Furthermore, the above-mentioned recording / reproducing device 2
00 is provided with a function for controlling its operation by a remote control signal such as infrared rays given from the outside.

In the speech speed converter 100 of this embodiment, an external input terminal is provided in addition to the microphone 101, and the audio output of the recording / reproducing device is input through the external input terminal. I have. The input audio signal is supplied to the analog amplifier 102 and the low-pass filter 1
03, to the microprocessor 110. In the microprocessor 110, the input audio signal is digitized by the A / D converter 111, and the CPU 1 is operated in accordance with a program stored in the ROM 115 in the same manner as a conventional speech speed conversion device.
13 executes the speech speed conversion process. The digital voice signal after the speech speed conversion is converted into an analog signal by a D / A converter 114 in the processor 110, and output through a low-pass filter 105 and an analog amplifier 104. The output sound is provided to headphones 300 or speakers 400. RAM120
Are allocated to various buffers necessary for the speech speed conversion processing. The control circuit 130 sets the pressed state of the slow button 160 to I /
The information is transmitted to the microprocessor 110 through the O circuit 112.
The control signal generating circuit 140 generates remote control information for controlling the operation of the recording / reproducing device 200 according to an instruction from the microprocessor 110. Remote control signal generation means 15
0 emits an infrared signal corresponding to the remote control information generated by the control signal generation circuit 140 to the recording / reproducing apparatus 200.

FIG. 2 schematically shows the operation of the speech speed converter executed inside the microprocessor. In the present embodiment, a monitoring means 70 is provided in addition to each means necessary for the operation of the conventional speech speed conversion device shown in FIG.
The monitoring means 70 monitors the state of the output ring buffer 50 and gives an instruction to the control signal generation circuit 140 and the A / D conversion means 10 according to the amount of audio data stored in the output ring buffer 50. The control signal generation circuit, as described in FIG.
A remote control signal such as an infrared ray is generated through 150. The remote control signal generating means 150 generates a pause signal for temporarily stopping the reproduction operation of the recording / reproducing device connected to the speech speed conversion device, and a pause release signal for restarting the reproduction. Further, the A / D conversion means 10 stops or restarts the A / D conversion operation in response to a command from the monitoring means.

FIG. 3 shows the operation timing of each means controlled by the monitoring means 70. First, when reproduction of the recording / reproduction device starts, the speech speed conversion device continues the speech speed conversion process of the input voice, and outputs the voice that has been slowly converted.
Then, an amount of audio data corresponding to a time delay between input and output is accumulated in the output ring buffer. Next, when the time delay reaches the maximum value that can be held in the output ring buffer,
The operation of the A / D conversion means 10 is stopped by the monitoring means 70, a pause signal is generated through the control signal generation circuit 140, and the recording / reproduction device stops reproduction of the sound. Since the input audio is stopped and the A / D conversion operation is also stopped, there is no data input to the input buffer. Since there is no more data to be processed, the speech speed conversion process 40 also suspends its operation. However, the output from the output ring buffer operates so as to be continued. As a result, there is no new input in the output ring buffer, and the stored data continues to be output, so that the amount of stored audio data decreases. When the stored audio data runs out, the monitoring means again gives a command to the control signal generation circuit and the A / D conversion means, generates a remote control signal for releasing the pause of the recording / reproduction device, and restarts the A / D conversion operation. To work. By repeating this series of operations, it becomes possible for the sound reproduced by the recording / reproducing device to continue the speech speed conversion for a long time, exceeding the limit imposed by the size of the output ring buffer.

In the above description, for the sake of convenience, the state change occurs when the amount of stored audio data reaches the maximum value that can be held in the output ring buffer and when the output ring buffer becomes empty. did. However, in practice, the recording / reproducing device may not respond instantaneously to the remote control signal, and it may take time to pause or resume reproduction. If there is a time delay when resuming playback, the speech speed converter
Despite resuming the D conversion and speech speed conversion processing, there is no audio input from the recording and playback device, and as a result,
The voice after speech speed conversion becomes discontinuous. In order to avoid such a problem, it is desirable to allow time for the above-mentioned state change to occur. That is, although the amount of accumulated audio data is not the maximum value that can be held in the output ring buffer, the above-mentioned value is approached when the audio data is sufficiently close to the maximum value and before all the audio data accumulated in the output ring buffer is output. It is desirable to operate the monitoring means so as to cause a state change.

FIG. 4 shows a second embodiment of the present invention.
In this embodiment, the input buffer 30 is a ring buffer, and the output buffer 50 is composed of two frame buffers. The input audio signal is recorded in the input ring buffer 30 immediately after the A / D conversion. At the time of the through operation in which the speech speed conversion process is not performed, the audio data once recorded in the input buffer 30 is not changed.
It is provided to the D / A conversion means 60. Also, at the time of speech speed conversion, it is sent to the speech speed conversion processing 40 in frame units. Speech speed conversion processing
The audio data of the frame length subjected to 40 is recorded in one output buffer, and during that time, the audio data for which the speech speed conversion processing has already been completed is output from the other output buffer.
It is output through the D / A conversion means 60. By inputting voice to the input ring buffer 30 and pipeline processing using two output buffers, a real-time speech speed conversion operation can be realized as in the first embodiment. In this embodiment, a reproduction address control means 80 for controlling a data output address from the input ring buffer and a repeat button 161 connected thereto are provided behind the input ring buffer. The reproduction address control means 80 controls a reading position of data transferred from the input ring buffer 30. That is, in the case of outputting voice through without performing the speech speed conversion process, the reproduction address control means 80 controls the read address so as to immediately output the data input immediately before. However, the write address operates so as to advance one by one. As a result, in the memory, the past sound for a certain period of time corresponding to one round of the ring buffer is always held. Further, when the slow button 160 is pressed to enter a state in which a speech speed conversion process is performed, the reproduction address control means 80 operates to move the output address in frame units. Further, the repeat button
When the button 161 is pressed, the operation is performed so that the output address goes back in the past by a certain number of frames. This allows
In the present embodiment, in addition to the speech speed conversion operation, a repeat operation of repeatedly reproducing the past voice can be realized. Further, when the slow button 160 is pressed during the repeat operation, the repeat reproduction sound can be slowly heard. This function is effective for language learning. Also, in the present embodiment, as in the first embodiment, a monitoring unit 70 is provided, monitors the amount of audio data stored in the input ring buffer 30, and performs A / D conversion in accordance with the amount. 10 and an operation for controlling an external recording / reproducing device. This operation is the same as that of the first embodiment, and operates at the timing shown in FIG. These operations are realized by a program on a microprocessor, as in the first embodiment.

FIG. 5 shows that during the above-mentioned repeat reproduction,
It is a figure which showed the operation timing of each means which the said monitoring means controls. In this embodiment, the playback operation of the connected recording / playback device is also temporarily stopped during the repeat playback. This is because if the repeat is repeated without stopping the input, the input ring buffer eventually becomes full, and the subsequent speech speed conversion operation cannot be continued. First, when the repeat button is pressed while listening to the voice through, the repeat reproduction is started from the address of the input ring buffer that goes back in the past. At this time, the monitoring means 70 stops the operation of the A / D conversion means 10 and stops the input to the input ring buffer 30, and at the same time, emits a pause signal of the remote control through the control signal generation means 140, and Stop the playback operation of. Next, when the repeat reproduction is completed, a pause release signal of the remote controller is generated, and the reproduction of the recording / reproduction device is restarted. With this series of operations, new audio data is not input to the input ring buffer during the repeat period, so that even if the repeat operation is repeated many times, the input ring buffer does not overflow.

Now, in the speech speed conversion device of the present invention,
As shown in FIG. 1, a standard usage method is for the user to directly hear the output through headphones or speakers. However, there is also a method in which the voice slowly converted through the speech speed converter of the present invention is copied and recorded on another recording medium and used later.

FIG. 6 shows an example of connection between devices in such a use method. This usage example uses a playback device that has a stereo playback audio output function of two channels on the left and right,
This is a method in which the same sound is output to both channels, and then only the sound of one channel is subjected to speech speed conversion and recorded on another medium. According to this method, the sound of one channel is recorded at a normal speed, and the converted sound is recorded in the other channel slowly. The user of the duplicate recording can listen to the same audio at two different speeds by switching between the left and right audio channels according to the purpose.

FIG. 7 shows an example in which the same method is applied to video dubbing. In this case, a time delay occurs between the video and the voice whose speech speed has been converted. However, in the speech speed conversion device of the present invention, the playback device side temporarily stops at the timing shown in FIG. 2, and this time delay is eliminated. That is, in a video created by this method, sound of a normal speed synchronized with an image is recorded on one channel, and slowly converted sound is recorded on the other sound channel. And while normal speed audio and images sometimes stop,
The slow sound of the other channel is recorded continuously. The stop state of the image and the normal speed sound operates so that the slowly converted sound catches up with the image and is released.

In the above embodiment, the description has been made using wireless control using a remote control signal such as infrared rays. However, a control line is provided between the speech speed converter and the recording / reproducing device, and these are directly connected. Even if the same control is performed, exactly the same effect can be obtained.

[0031]

According to the present invention, it is possible to continuously listen to the recorded voice for a long time while performing the speech speed conversion process. Further, the repeat operation or the repeat operation can be used in combination with the speech speed conversion process, and the shortage of the buffer memory for holding the input voice during the repeat operation can be avoided.

[0032]

[Brief description of the drawings]

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

FIG. 2 is an operation explanatory diagram of the first embodiment of the present invention.

FIG. 3 is an explanatory diagram of operation timing in the embodiment of the present invention.

FIG. 4 is an explanatory diagram of the operation of the second embodiment of the present invention.

FIG. 5 is an explanatory diagram of operation timing in a second embodiment of the present invention.

FIG. 6 is a diagram of an example of use of the present invention.

FIG. 7 is a diagram of a usage example using a video of the present invention.

FIG. 8 is a diagram illustrating the operation of a conventional speech speed conversion device.

FIG. 9 is a configuration diagram of a conventional speech speed conversion device.

[Explanation of symbols]

10: A / D conversion means, 20: switch means, 30: input buffer, 40: speech speed conversion processing means, 50: output buffer, 60: D /
A conversion means, 70 monitoring means, 80 address control means, 100
... Speech speed converter, 101 ... Microphone, 102,104 ... Analog amplifier, 103,105 ... Low pass filter, 110 ... Microprocessor, 111 ... A / D converter, 112 ... I / O circuit, 113 ... CP
U, 114 D / A converter, 115 ROM, 120 external RAM, 130 control circuit, 140 control signal generation circuit, 150 remote control signal generation means, 160 slow button, 161 repeat button, 20
0,201… Recording / reproducing device, 300… Headphone, 400… Speaker, 500,501… VCR

──────────────────────────────────────────────────続 き Continuation of front page (56) References JP-A-62-271254 (JP, A) JP-A-55-156999 (JP, A) JP-A-55-48812 (JP, A) (58) Field (Int.Cl. ⁷ , DB name) G10L 21/04

Claims (8)

(57) [Claims] 1. A captures input speech from the recording and reproducing means, a speech speed converting device that outputs voice converted speech speed to extend the time axis without changing the pitch of the input speech, the speech speed
Stored in the memory means inside the speech speed converter with the conversion process
Control signals output according to the amount of audio data
And the reproduction operation of the recording / reproducing unit is performed by the control signal.
A speech speed conversion device comprising control means for controlling . 2. The speech speed conversion device according to claim 1, wherein said speech speed conversion device includes an A / D converter, an input buffer including a plurality of frame buffers, and a speech speed conversion device that expands and contracts the time length of the audio signal in frame units without changing the pitch. Processing means, an output buffer having a ring buffer configuration in which addresses are used in a circular fashion, and D / A conversion means, and monitoring means for monitoring the amount of audio data stored in the output ring buffer; Control signal generation means for generating a control signal for controlling the reproduction operation of the reproduction means, wherein the monitoring means corresponds to the audio data amount stored in the output buffer,
The speech speed conversion device according to claim 1, wherein the operation of the D conversion means and the operation of the control signal generation means are controlled. 3. The speech speed conversion device according to claim 1 , wherein the A / D conversion means, an input buffer having a ring buffer configuration in which addresses are used in a ring form, a reproduction address control means for controlling an address read from the input buffer, A speech speed conversion processing unit that expands and contracts the time length of the unit audio signal without changing the pitch, an output buffer configured by a plurality of frame buffers, and a D / A conversion unit, and the input ring buffer Monitoring means for monitoring the amount of audio data to be stored, and control signal generating means for generating a control signal for controlling the reproduction operation of the recording / reproducing unit, wherein the monitoring means is stored in the input buffer. And controlling the operation of the A / D conversion unit and the operation of the control signal generation unit in accordance with the amount of audio data to be transmitted.
The speech speed conversion device according to 1. 4. The control by the monitoring means, wherein the amount of audio data stored in the input or output ring buffer is:
When approaching the maximum amount that can be held in the input or output ring buffer, the playback of the recording / playback unit and the operation of the A / D conversion unit are temporarily stopped, and then the speech speed conversion process is continued and the accumulated voice data is When the amount of audio data output and stored in the input or output buffer approaches 0, the operation of the recording / reproducing unit and the operation of the A / D conversion means are started again. The speech speed conversion device according to claim 2 or 3. 5. The speech speed conversion device according to claim 1 , further comprising a repeat function for repeatedly reproducing the past voice, and when the repeat function is executed, the monitoring means includes the recording / reproducing unit and the A / D conversion means. 4. The speech speed conversion device according to claim 3, wherein the operation is temporarily stopped to prevent new input voice data from entering the input ring buffer during the repeat operation. 6. The control signal for controlling the operation of the recording / reproducing unit is a wireless remote control signal.
The speech speed conversion device according to claim 1. 7. A speech speed conversion device comprising: a low-pass filter for anti-aliasing at an input and an output of an analog audio signal; an A / D converter; a D / A converter;
A means having a microprocessor, an external memory, a control button, and a control circuit for inputting a state of the control button to the microprocessor, and generating a control signal for controlling the recording / playback unit; Having, and the above-described series of operations is realized by a software program on the microprocessor,
The speech speed conversion device according to any one of claims 1 to 6. 8. A sound reproducing unit that reproduces sound from a recording / reproducing unit having two output channels, outputs the same sound signal from both of the two output channels, and outputs two separately prepared input channels. In a system for duplicating and recording voice by a recording / reproducing unit having a voice speed conversion device for performing a voice speed conversion process for expanding or contracting the time length between signal paths of one channel without changing the pitch of the voice signal the provided recording and speech speed conversion processing only the audio piece channels, to the other channel normal speed of sound was FEATURES <br/> to be recorded, utilizing the speech speed converting device Method.