JPH07129190A - Talk speed change method and device and electronic device - Google Patents

Abstract

PURPOSE:To provide possibility of hearing a talk with a retarded talk speed with no change in the feature of the hearer's voice by changing the talk speed without changing the pitch of the input voice only during the specified time when the hearer requires it. CONSTITUTION:A voice is fed to a microphone 321, emitted as a voice signal, digital converted at certain time intervals by an A/D converter 5 upon passing through an amplifier 10 and low-pass filter 7, and subjected to a talk speed converting process to be made by a software 11 on a digital signal processor (DSP) 1. A PTL, switch 4 is connected with a terminal 13 for external interrupt flag contained in the DSP 1, and the software 11 on the DSP 1 judges whether talk speed conversion should take place depending upon the numerical value of a flag register 14. When the hearer needs talk speed conversion, it is done without changing the pitch of the input voice only during the specified time. Accordingly this talk speed converting device can be used even in the situation such as in talking, and the hearer can select the voice to which a talk speed conversion should be applied.

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 method, a speech speed conversion device and an electronic device for modulating the speed of voice,
In particular, the present invention relates to a technique effectively applied to a control technique for using the device for dialogues and the like.

[0002]

2. Description of the Related Art As a means for assisting the hearing of a hearing-impaired person, an analog type hearing aid which processes an amplitude and a frequency characteristic of a voice by using an analog circuit has been mainly used. On the other hand, in recent years, research and development for applying digital signal processing to compensation for hearing impairment have been actively conducted. For this research and development trend, see, for example, the Acoustical Society of Japan (1
991, Vol. 47, No. 10, P760-P765) "Application of Digital Technology to Deafness Compensation" and J. Acoust. So
c. Am. (90 (2), Pt.1, Aug. 1991) "Speech-p
erception aids for hearing-impaied people: Curre
nt status and needed research ”etc.

Generally, in order to compensate for the loss of hearing, the sound pressure level is amplified and the dynamic range is compressed for each frequency according to the hearing characteristics of the user. In a conventional analog hearing aid, such processing is realized by an analog circuit. Further, in the recently developed digital hearing aid, this processing is realized by software such as a digital filter, so that the user's hearing characteristics can be more precisely adapted.

Under these circumstances, in recent years, hearing aids for the entire auditory system including a decline in higher-order language processing speed are obtained by changing only the speed without changing the pitch of voice by digital signal processing. An attempt is made to go. For such a speech rate conversion technology, for example, Technical Report of IEICE (Vol.92 No.207 SP92-54 "Development of portable DSP system for voice processing for the elderly") and the same (SP92-55). It is described in detail in "High-quality real-time speech speed conversion system".

[0005]

In the above-mentioned conventional technique, the speech for converting the speech speed is intended for broadcasting speech of television / radio or speech recorded on a tape recorder or the like. That is, only the voice unilaterally given to the listener was the target of speech speed conversion.

However, considering that the conventional hearing aid can be used regardless of the type of input voice, it is desirable that the speech speed conversion apparatus can also handle voices other than the above. In particular,
If you can listen slowly to the voice of the other party in a dialogue, it can be used not only to assist the hearing of the elderly and the hearing impaired, but also to assist in hearing in foreign language conversations that normal hearing people are not familiar with. It will be possible.

An object of the present invention is to provide a technique capable of converting the speech speed as needed and reproducing it.

Another object of the present invention is to provide a technique capable of storing the original voice data and constantly converting the speech speed based on the original voice data.

Another object of the present invention is to provide a control means of the device for utilizing the speech speed conversion device for dialogue and the like.

Another object of the present invention is to provide a technique capable of expanding the application range of the speech speed conversion device.

Another object of the present invention is to provide a technique capable of effectively utilizing the storage device (memory) of the speech speed conversion device.

Another object of the present invention is to provide a technique capable of returning a read pointer to a storage device (memory) of a speech speed conversion device.

Another object of the present invention is to provide a technology capable of controlling connected AV equipment in a speech speed conversion apparatus.

Another object of the present invention is to provide a technique capable of continuously converting a voice speed in a voice speed conversion device.

Another object of the present invention is to provide a technique capable of improving the operability of the speech speed conversion device.

Another object of the present invention is to provide a technique capable of reducing the power consumption of the speech speed conversion device.

The above and other objects and novel features of the present invention will be apparent from the description of this specification and the accompanying drawings.

[0018]

The outline of the representative ones of the inventions disclosed by the present application will be briefly described as follows.

(1) A voice speed conversion method in which a voice is input and only the speed of the voice is changed without changing the pitch of the input voice, which is a time specified when the listener needs the voice speed conversion. During this period, the speech speed conversion processing of the input voice is performed, and during the other time, the speech speed conversion is not performed.

(2) A voice speed having means for inputting voice, voice speed conversion processing means for changing the speed of the input voice, and means for outputting the output of the voice speed conversion processing means to the listener's ear as voice. A conversion device, wherein the speech speed conversion device is provided with a speech speed conversion processing switch, and the speech speed of the input voice is changed and output only while the speech speed conversion processing switch is ON. , The speech speed conversion processing switch is off (OFF)
A means for outputting the input voice without changing the speech speed is provided during the operation.

(3) A speech speed conversion method in which original speech is encoded and accumulated, the accumulated encoded speech is read out, and only the speed of the speech is changed without changing the pitch of the original speech. The speech speed conversion processing of the input voice is performed only during the designated time when the speech speed conversion is required, and the speech speed conversion is not performed during the other time.

(4) Means for inputting the original voice, storage means for encoding and accumulating the input voice, and voice speed conversion processing means for reading the accumulated encoded voice and changing the speed of the input voice. And a speech speed conversion device having means for outputting the output of the speech speed conversion processing means to the listener's ear as voice,
The speech speed conversion device is provided with a speech speed conversion processing switch, and the speech speed of the input voice is changed and output only while the speech speed conversion processing switch is on (ON). A means for outputting the input voice without changing the speech speed is provided while the power switch is turned off.

(5) The storage means has means for storing in frame units.

(6) The determination of the waveform expansion / contraction processing in the speech speed conversion processing is made by comparing the power of the frame with a threshold value, and the threshold value is made variable.

(7) The speech speed conversion device is provided with a speech speed selection switch for selecting a speech speed, and means for changing the speech speed selected by the speech speed selection switch.

(8) The speech speed converter is provided with means (AV control) for controlling audio / video equipment.

(9) The speech speed conversion apparatus is provided with a repeat switch, and means for repeating reproduced voice while the repeat switch is on.

(10) The repeat means is a means for backing up for several seconds each time it is pressed, a means for occasionally generating an intermittent sound while returning, a means for preventing further returning after reaching the end of the ring buffer, a repeat At least one of the means for selecting the speech speed at time is included.

(11) The means for selecting the speech speed at the time of repeat is a default value of repeat, slow repeat,
It has at least two or more of the repeat and the repeat being gradually quickened by fast listening.

(12) In the speech speed conversion device, when a delay from real time occurs due to the speech speed conversion or the repeat operation, the delay amount is adjusted while the stored information is reproduced. It is provided with a chase means for carrying out.

(13) The chasing means is means for starting the chasing when the mode for slowly reproducing is finished,
Means to start chasing when playing to the point where the repeat starts after repeat, means to select the speech speed at the time of chasing, means to automatically shift to the through mode to output the input sound as it is when catching up, and notification signal sound (message) when catching up At least one of the means for generating

(14) The means for selecting the speech speed at the time of chasing has at least one of a means for skipping to reality at once, a means for chasing reality by listening fast, and a means for moving in parallel with delay. is there.

(15) At least one of the voice speed conversion processing switch, the voice speed selection switch, the repeat switch, and the reset switch is provided on one easily accessible peripheral portion on one side surface of the voice speed conversion device. It is a thing.

(16) The reset switch has means for stopping the operation when the switch is turned on during the repeat operation or the chase operation, skipping the operation, and then shifting to the through mode.

(17) The speech speed conversion processing means is provided as software executed by a digital signal processor having a terminal for inputting an interrupt request signal from the outside, and the speech speed conversion processing switch talks. The control of the speed conversion processing or the switching of the speech speed conversion speed is given to the digital signal processor through a terminal for inputting the interrupt request signal.

(18) Means for listening to the output voice of the speech speed conversion device through the binaural headphones.

(19) A microphone that converts an acoustic signal into an electric signal, an analog amplifier that amplifies the microphone output, a low-pass filter that removes high-frequency components of the output of the analog amplifier, and an analog signal that is output from the low-pass filter is digitalized. A / D converter for converting into a signal, a digital signal processor for executing processing for changing the speed of voice by digital signal processing, storage means for storing input voice data and data of signal processing result, and the digital signal processor Means for controlling the processing for changing the speed of the sound, the means for changing the processing parameters, the D / A converter for converting the digital sound data into an analog value, and the high frequency of the output of the D / A converter. A second low-pass filter for removing the component and an output of the second low-pass filter. A second analog amplifier for amplifying an a speech speed converting device and a headphone for providing an output of the analog amplifier in the second both ears is converted to an acoustic signal.

(20) A microphone that converts an acoustic signal into an electric signal, an analog amplifier that amplifies the microphone output, a low-pass filter that removes high-frequency components of the output of the analog amplifier, and an analog signal that is output from the low-pass filter is digitalized. A / D converter for converting into a signal, storage means for storing input voice data and data of signal processing result, and digital for executing processing for reading the stored information and changing the speed of voice by digital signal processing A signal processor, a means for controlling the processing for modulating the speed of the sound performed by the digital signal processor, a means for changing the processing parameters, a D / A converter for converting the digital sound data into an analog value,
A second low-pass filter for removing high-frequency components of the output of the D / A converter, a second analog amplifier for amplifying the output of the second low-pass filter, and an output of the second analog amplifier as an acoustic signal. It is a speech speed conversion device having headphones that are converted and given to both ears.

(21) The speech speed conversion processing means is carried out by frame-by-frame pipeline processing using a plurality of input frame buffers, and for each frame of data, pitch extraction processing is first performed for the beginning of the frame. Then, the pitch of that portion is detected, the detected data of one pitch length is transferred to the output buffer, and the window function that changes from 0 to 1 and the data of 1 pitch to 0 for the data of two pitch lengths. Multiply the changing window function, add the data of the results of applying each window function to create a composite waveform with a time length of 2 pitches, insert it after the 1 pitch of data that was transferred earlier,
The pitch detection process is performed again with the position separated by two pitches from the position on the data that has been subjected to the pitch extraction process first, as the beginning.
Pitch detection is performed at that position, and a series of procedures for transferring data for n (n is an integer) pitch to the output buffer in units of the pitch length obtained by the last pitch detection is repeatedly performed over the entire frame. It is a speech speed converter.

(22) The speech speed conversion processing means calculates the average power of the data in the input frame, and is executed only when the average power is larger than a preset threshold value. , The data included in the frame is directly transferred to the output buffer.

(23) In the threshold processing for the average power of the data in the input frame, a second threshold is provided, and a frame having an average power smaller than the second threshold is preset. When the data continues for a time longer than the time threshold, it is prohibited to transfer the data of the frames, which are continuous beyond the time threshold and have a smaller average power than the second threshold, to the output buffer. It was done.

(24) Each of the switches is a switch having a soft contact feeling so that the microphone does not pick up the click sound of the switch.

(25) Each of the switches has a surface form having a different tactile sensation so that which switch can be seen without looking.

(26) A cloth rubbing noise preventing means is provided for changing the distance between the microphone and the main body of the apparatus to prevent the microphone and clothing from directly contacting each other when the apparatus is put in a chest pocket and used. ing.

(27) A display means for visually observing a time delay amount from the present is provided at a predetermined position of the speech speed conversion device.

(28) A ring buffer is used as the storage means, and means for managing the delay time is provided by a counter indicating the time delay on the ring buffer.

(29) In addition to the through mode, a standby mode for lowering the clock cycle of the processor and performing the same processing as the slow mode is provided.

(30) The power switch is set to three stages of ON (ON), OFF (OFF), and ON and OFF intermediate, and when the switch is adjusted to the intermediate position, the analog input / output systems in the audio signal processing system are mutually connected. It operates in the analog through mode in which a short circuit occurs and power supply to the digital processing system between the analog input / output systems is stopped.

(31) A telephone set in which any one of the voice speed conversion means (2), (4) to (30) is provided between the handset of the telephone set and the main body of the telephone set.

(32) A telephone exchange in which the speech speed converting means according to any one of claims 2, 4 to 30 is provided in the telephone exchange.

[0051]

According to (1) and (2) of the above-mentioned means,
When inputting a voice and changing only the voice speed without changing the pitch of the input voice, the voice speed conversion process of the input voice is performed only during the time specified when the voice speed conversion is required. , Since the speech speed conversion is not performed during the other time, it is possible to use the speech speed conversion device not only in the one-sided voice given to the listener like radio voice but also in the situation such as dialogue. Therefore, the listener can select the voice to be subjected to the speech speed conversion without disturbing the listener's own speech.

Further, in a hearing aid, a foreign language learning device, a telephone, etc., it is possible to listen at a slow speech speed without changing the characteristics of the speaker's voice.

According to (3) and (4) of the above-mentioned means, the original voice is encoded and accumulated, the accumulated encoded voice is read out, and the pitch of the original voice is not changed, When changing only the speed of voice, etc., the voice speed conversion process of the input voice is performed only during the time specified when the voice speed conversion is required, and the voice speed conversion is performed during other times. In addition to the effects of (1) and (2) above, effective use of memory, original voice repeat function, voice memory function, repeat voice speech speed conversion function, and fast playback function can be added. .

According to item (5) of the above-mentioned means, since data is stored in frame units, the efficiency of writing and reading can be improved.

According to item (6) of the above-mentioned means, the waveform expansion processing, the shortening processing, and the silent section deletion processing in the speech speed conversion processing are determined by comparing the power of the frame with the threshold value, and Since the threshold value is changed according to the volume of the input voice, the speech speed conversion process can be performed according to the usage environment condition.

According to (7) of the above-mentioned means, the speech speed conversion device is provided with a speech speed selection switch for selecting the speech speed, and the speech speed selected by the speech speed selection switch is changed. Since the means is provided, the listener can select the speech speed of the voice heard.

According to (8) of the above-mentioned means, means for controlling the audio / video equipment by the voice speed conversion device (AV
By providing the control), when the memory capacity is insufficient, regardless of the speech speed conversion expansion / contraction rate,
When a signal to pause the playback operation of an external device is issued, voice input to the speech speed converter is temporarily stopped, and when there is free space in the memory, the pause signal output is stopped and the external device is restarted. Since the operation to start voice input from the device is repeated,
It is possible to continuously use the speech speed conversion for a long time.

According to (9) to (11) of the above-mentioned means, the speech speed conversion device is provided with a switch for repeating the speech speed, and the reproduced voice is reproduced while the repeat switch is on (ON). Is provided, it is possible to convert the speech speed of the repeat voice.

According to (12) to (14) of the above-mentioned means, since the chase means for chasing the information stored in the speech speed conversion device to the desired point is provided, the application range of the speech speed conversion device. Can be expanded, the operation time can be shortened, and the usability can be improved.

According to (15) and (16) of the above-mentioned means, the voice speed conversion processing switch, the voice speed selection switch, and the repeat switch are provided in one easily accessible peripheral portion on one side of the voice speed conversion device. Since at least one of the operation switch and the reset switch is provided, the application range of the speech speed conversion device can be expanded, the operation time can be shortened, and the usability can be improved.

According to (17) to (23) of the above-mentioned means, the speech speed conversion processing means is software executed by a digital signal processor having a terminal for inputting an interrupt request signal from the outside. Provided as
The control of the voice speed conversion processing by the voice speed conversion processing switch or the switching of the voice speed conversion speed is given to the digital signal processor through a terminal for inputting the interrupt request signal.

Further, the speech speed conversion processing means is carried out by frame-by-frame pipeline processing using a plurality of input frame buffers. For each frame of data, first, pitch extraction processing is performed on the beginning of the frame. Then, the pitch of that portion is detected, the detected data for one pitch length is transferred to the output buffer, and the window function that changes from 0 to 1 and the change from 1 to 0 for the data for two pitch lengths. Multiply the window function, and add the resulting data of each window function to create a composite waveform with a time length of 2 pitches, insert it after the 1 pitch of data that was previously transferred, and The pitch detection process is performed again starting from the position separated by 2 pitches from the position on the data subjected to the extraction process, the pitch is detected at that position, and the pitch length obtained by the final pitch detection is calculated. n (n is an integer) to perform repeated over a series of procedures for transferring to the output buffer data of pitch in the entire frame.

Further, the speech speed conversion processing means calculates the average power of the data in the input frame, is executed only when the average power is larger than a preset threshold value, and when it is smaller, The data contained in the frame is transferred to the output buffer as it is.

In the threshold processing for the average power of the data in the input frame, a second threshold is provided, and a frame having an average power smaller than the second threshold has a preset time. When the data continues for a time longer than the threshold value, the data of the frames having the average power smaller than the second threshold value and continuously exceeding the time threshold value are transferred to the output buffer.

By configuring the speech speed conversion processing means in this way, it is possible to improve the efficiency of the speech speed conversion processing and prevent the quality deterioration of the reproduced voice.

According to item (24) of the above-mentioned means, since the microphone does not pick up the click sound of the switch, it is possible to prevent a large noise when the switch is operated.

According to item (25) of the above-mentioned means, since the surface has a different tactile sensation so that the user can recognize which switch without looking at it, the operability can be improved.

According to (26) in the above-mentioned means, since the cloth rubbing noise preventing means for the microphone is provided, it is possible to reduce the intrusion of noise.

According to (27) of the above-mentioned means, since the display means for visually observing the time delay amount from the present is provided at the predetermined position of the speech speed conversion device, the operation time is shortened and the usability is improved. It can be improved.

According to (28) of the above-mentioned means, since the ring buffer is used as the storage means and the means for managing the delay time by the counter showing the time delay on the ring buffer is provided, the repeat processing, A chase process or the like can be easily performed.

According to (29) of the above-mentioned means, since the standby mode is provided in addition to the through mode, the power consumption can be reduced.

According to (30) of the above-mentioned means, the power switch is turned on (ON), turned off (OFF), and has three stages of on and off, and the analog slow mode is provided. You can measure.

According to (31) of the above-mentioned means, since the speech speed converting means is provided between the handset of the telephone and the main body of the device, the listener's own utterance is not disturbed,
The listener can select the voice for which the speed conversion is applied.

Further, it is possible to listen at a slow speech speed without changing the characteristics of the speaker's voice on the telephone.

According to (32) of the above-mentioned means, since the speech speed converting means is provided in the telephone exchange, the listener can hear the voice to be subjected to the speech speed conversion without disturbing the listener's own speech. You can choose.

[0076]

Embodiments of the present invention will now be described in detail with reference to the drawings. In all the drawings for explaining the embodiments, parts having the same function are designated by the same reference numerals, and repeated description thereof will be omitted.

(Embodiment 1) FIG. 1 is a block diagram showing a schematic configuration of an internal circuit of Embodiment 1 according to the present invention. Reference numeral 1 is a DSP (digital signal processor), and 11 is software for performing speech speed conversion processing. , 12 is a serial port,
Reference numeral 13 is an external interrupt flag terminal, 14 is a flag register, 2 is a voice memory (output buffer), 3 is a selector switch, 4 is a PTL (Push-To-Listen) switch,
5 is an A / D converter, 6 is a D / A converter, 7 is a low-pass filter, 8 is a low-pass filter, 9 is an analog amplifier,
10 is an analog amplifier, 321 is a microphone, 325
Are headphones for both ears (earphones).

In the speech speed converting apparatus of the first embodiment, as shown in FIG.
As shown in, the voice is input to the microphone 321 and output as a voice signal (electrical signal). The audio signal is input to the A / D converter 5 through the amplifier 10 and the low-pass filter 7, and is converted from an analog value to a digital value at a preset time interval in the A / D converter 5.

The audio signal converted into the digital value is input to the DSP 1. Then, the voice speed conversion processing of the audio signal is realized by the software 11 on the DSP 1. The PTL switch 4 is connected to the external interrupt flag terminal 13 of the DSP 1,
The state of is expressed as a numerical value of the flag register 14 inside the DSP 1 corresponding to this terminal 13. The software 11 on the DSP 1 determines whether or not to perform the voice speed conversion process according to the numerical value of the flag register 14.

The digital voice data which has been subjected to the speech speed conversion processing is held in the output buffer memory 2. The D / A converter 6 converts the data in the output buffer memory 2 from a digital value to an analog value at preset time intervals.
The analog signal obtained by this conversion is input to the analog amplifier 9 through the low-pass filter 8 and is output as sound by the binaural headphones 325 at the listener's favorite sound pressure level.

In the first embodiment, two types of switches are prepared for the PTL switch 4. One is a switch that conducts only while the push button is pressed. The other is a switch that maintains a conductive state even when the hand is released from the push button. The former is used for dialogue, the latter is
It is used when continuously converting the speech rate of a unilaterally given voice such as a radio voice, which is a conventional use method.
In addition, in the first embodiment, in addition to the PTL switch 4, the selector switch 3 is connected to the external interrupt flag terminal 13 of the DSP 1. By switching the selector switch 3, the numerical value of the flag register 14 is changed, and the software 11 changes the expansion rate of the speech speed conversion processing according to this numerical value.

FIG. 2 is a diagram for explaining the speech speed conversion processing executed in the DSP 1 of the first embodiment. Example 1
The speech speed conversion processing is a method of detecting the pitch (basic period) of a voice signal and extending the length of the waveform in units of the detected pitch. This is a set of voice data for several tens of milliseconds (hereinafter referred to as a frame). ) Is a unit of one processing. Therefore, an input buffer having at least two frame lengths is prepared inside the DSP 1, and while the data is input from the A / D converter to one buffer, the data accumulated in the other buffer is processed ( Pipeline processing). The processed data is stored in the output buffer 2 having a sufficiently large capacity. The processing procedure for the data of each frame is as follows.

First, (a) pitch extraction processing (not shown) is applied to the head portion of the frame to detect the pitch of that portion.

(B) Next, the detected data for one pitch length is transferred to the output buffer 2.

(C) Next, for the data for two pitch lengths,
A window function that changes from 0 to 1 and a window function that changes from 1 to 0 are applied.

However, the position on the data where the window function is started is shifted by one pitch. Then, the data obtained by applying the respective window functions are added to create a composite waveform having a time length of 2 pitches, and the synthesized waveform is inserted after the previously transferred data for 1 pitch.

(D) Next, a pitch detection process (not shown) is performed again with the position separated by two pitches from the position on the data which has been subjected to the pitch extraction process first, and the pitch detection at that position. I do. Generally, the pitch of the voice is constantly changing, and a pitch different from the pitch detected previously is obtained by the second detection.

(E) Data for n pitches is transferred to the output buffer in units of the pitch length obtained by the last pitch detection.

The above steps (a) to (e) are repeated over the entire frame.

Since the pitch length depends on the input voice, the number of repetitions in one frame is not constant. In addition, different expansion rates are realized by changing the value of n in the process (e). For example, 3 in the input buffer with n = 1
Since data for 4 pitches is generated from data for pitches, the expansion rate is 4/3 = 1.33 times. Similarly, n =
When it is 0, it is 1.50 times, and when n = 2, it is 1.25 times.

In the first embodiment, the speech speed conversion processing of FIG. 2 is not performed for all frames, but the average power of each frame is calculated and the power is set in advance. Only when the threshold value Th is exceeded, the above-described processing of FIG. 2 is performed. Then, the data of the frame whose power does not exceed the threshold value Th is transferred to the output buffer as it is. FIG. 3 shows the concept of this threshold processing.

In FIG. 3, the portion where the power of each frame exceeds the threshold value Th is shown as an extension section. By this threshold processing, the rising and falling portions of the audio signal are not processed and are output as the original sound.
There is an advantage that the characteristics of the voice included in the rise and fall of the voice, for example, the features of consonants are not destroyed.

Furthermore, in the first embodiment, in the threshold processing for the average power for each frame shown in FIG.
A second threshold value To is provided. Then, when a frame whose power is lower than the second threshold value To continues for 1 second or more, the frame whose power is lower than To lasting 1 second or more is not output. As a result, the amount of data stored in the output buffer can be reduced.

In FIG. 3, this non-output portion is represented as a deletion section. In the output buffer 2, the data after the post-speech speed conversion processing is written once per frame, and at the same time, the data is output to the D / A converter 6 one by one at regular time intervals. The addresses of the output buffer 2 are set in a ring shape, and the last address and the first address are continuous.

Therefore, in this ring-shaped address space, the address pointer Pi pointing to the data write destination after the voice speed conversion processing is operated by the address pointer Po pointing to the data to be sent to the D / A converter.
In the first embodiment, the traveling speed of Pi is faster than the traveling speed of Po, so that Pi will overtake Po anyway. At this point, the information that has been stored in the output buffer 2 until then is rewritten without being output.

Therefore, the time from the start of the voice speed conversion operation until such a state becomes the time length of the input voice that can be handled by the voice speed conversion processing of the first embodiment. The reduction of the data amount by the threshold value To has the effect of increasing the time length that can be handled.

Regarding the signal processing method of the speech speed conversion processing described with reference to FIGS. 2 and 3, the Institute of Electronics, Information and Communication Engineers Technical Report SP92-150 (1993-03)
Reported in "Evaluation of Speech Rate Conversion Method by Hearing-Impaired People" or Proceedings of Acoustical Society of Japan (March 1993) 1-7-6 "Study of Speech Rate Conversion Method Using Portable DSP System" .

FIG. 4 is a diagram showing a form of use of the speech speed conversion apparatus according to the first embodiment. Although the PTL switch 4 is arranged on the upper surface of the device in FIG. 4, the arrangement position may be another part. On the other hand, next to the PTL switch 4, there is provided a selector switch 3 for changing the expansion rate of the voice speed conversion. Similar to the PTL switch 4, the state of the selector switch 3 can be observed from the software on the DSP 1 through the external interrupt flag terminal of the DSP 1, and the state of the selector switch 3 when the PTL switch 4 is pressed Change the n value during the speech speed conversion process. By operating the PTL switch 4 and the selector switch 3 alternately, it is possible to change the expansion rate for each utterance.

FIG. 5 is a flow chart showing the control procedure. In the speech speed conversion processing, for example, a processing unit is a frame having a time length of several tens of milliseconds.
The / D conversion and the D / A conversion are processings performed at fixed time intervals shorter than that, for example, several tens of microseconds.
Therefore, the A / D conversion, the D / A conversion, and the processing associated therewith are realized as interrupt processing, as shown in FIG. While the speech speed conversion process and the interrupt waiting process are being performed, the interrupt process is performed by the interrupt signal from the serial port to which the A / D converter and the D / A converter are connected.

As can be seen from the above description, this embodiment 1
According to the authors, it becomes possible to use a speech speed conversion device not only for voice that is given to the listener unilaterally such as radio speech, but also for situations such as dialogue, so that the listener can speak without changing the speech speed. The listener can select the voice to be converted.

Further, the speech speed converting apparatus of the first embodiment can be used for assisting the deterioration of the ability of the elderly to hear the voice. It goes without saying that it can also be used in situations where a hearing person hears a foreign language that they are not familiar with.

(Embodiment 2) FIGS. 6 to 10 are views showing the external structure of Embodiment 2 of the speech speed conversion apparatus according to the present invention. FIG. 6 is a front plan view seen from the front, and FIG. 7 is a rear view. FIG. 8 is a top plan view seen from above, FIG. 9 is a left side plan view seen from the left side, and FIG. 10 is a right side plan view seen from the right side.

6 to 10, 101 is the main body of the speech speed conversion device, 102 is a back cover, 103 is a dent for finger rest, 104 is a slow switch (slow push button), 10
5 is a repeat switch (repeat push button), 106 is a reset switch (reset push button), 321 is a microphone, 108 is a volume control, 109 is a power switch, 110 is an earphone terminal, 111 is an external input terminal,
Reference numeral 112 is an AV control terminal, and 113 is a speech speed changeover switch (speech speed setting switch).

As shown in FIGS. 6 to 10, the speech speed converting apparatus according to the second embodiment has a slow switch 104 and a repeat switch at a position where it is easy to operate with one hand of the main body 101 of the speech speed converting apparatus, for example, on the front upper side. A switch 105 and a reset switch 106 are provided, and the speech speed changeover switch 11 is shown on the right side plan view.
3 is provided.

The push button of the slow switch 104 is
Since it is pushed frequently, it is made larger than other push buttons. And since I keep pushing the push button slowly, I am tired and can fix it.
For example, a slide lock method that locks by pushing and sliding to the side, a double-click method that locks by clicking twice, and a method that releases when the reset push button is pressed are used.

The speech speed changeover switch (speech speed setting switch) 113 is arranged close to the range where it can be operated with the same finger so that it can be operated alternately with the slow switch 104.

In addition to the position of the above-mentioned embodiment, a ring switch, a slide switch or the like may be used to make the operation easier.

The volume control 108 is also arranged in a range that can be easily adjusted by the same finger so that the user can always listen at an appropriate volume.

Further, the frequently used switches such as the slow switch 104, the repeat switch 105, the reset switch 106, the speech speed changeover switch 113, etc. are switches having a soft touch feeling so that the microphone 321 does not pick up the click sound of the switch. Preference is given to using.
For example, a switch using a conductive rubber or the like is used.

Further, it is preferable that the appearance of each of the switches is formed in a surface state having a different tactile sensation so that the kind of switch can be recognized without looking.

When the finger-holding dent 103 is opened, some switches such as a speech speed selection switch at the time of repeat can be seen.

The internal circuit configuration of the speech speed converter of the second embodiment is similar to that of the first embodiment shown in FIG.

As the PTL switch 4 of the first embodiment,
The slow switch 104, the repeat switch 105,
The reset switch 106 and the like are used. A speech speed changeover switch (speech speed setting switch) 113 is used as the selector switch 3 of the first embodiment.
Then, the speech speed changeover switch (speech speed setting switch) 113
Are connected to the external interrupt flag terminal 13 of the DSP 1. By switching the voice speed changeover switch 113, the numerical value of the flag register 14 is changed, and the software 11 changes the expansion rate of the voice speed conversion processing according to this numerical value.

FIG. 11 is a block diagram showing the functional arrangement of the speech speed conversion apparatus according to the second embodiment, in which 21 is a voice input unit and 2 is a voice input unit.
2 is an input buffer, 23 is a central processing unit (CPU), 24
Is a ring buffer memory (corresponding to the audio memory 2 in FIG. 1), 25 is a function selection unit, 26 is an output buffer, and 27 is an audio output unit.

When the components of the second embodiment are made to correspond to those of the first embodiment with reference to FIG. 1, the voice input unit 21 includes the microphone 321 and the analog amplifier (amplifier) shown in FIG. 10, a low pass filter 7 and an A / D converter 5.

The input buffer 22 is for holding the voice converted into a digital signal by the voice input unit 21, and is large enough to hold one frame of data which is a unit for performing subsequent signal processing. have. The input buffer 22 can be realized by allocating a part of the address of the ring buffer memory (corresponding to the voice memory 2 of FIG. 1) 24.

The central processing unit (CPU) 23 corresponds to the software portion to be executed on the DSP 1 shown in FIG. 1, and includes a voice compression unit 23A and a silence compression unit 23.
B, decompression unit 23C, waveform processing unit (speech rate conversion processing unit) 23
D, and a control unit 23E.

The function selector 25 corresponds to the switches 3, 4 and the external interrupt flag terminal 13 shown in FIG. 1, and corresponds to the slow switch 104 and the repeat switch 1.
05, reset switch 106, speech speed switch 11
It is composed of 3 etc.

The output buffer 26 is for holding the data of the result processed by the waveform processing section 23D. Actually, there are two output buffers, and one frame is expanded by the waveform processing. Minute data is large enough to fit. In the first embodiment, there are two input buffers, and the pipeline processing is realized by alternately using them. In the second embodiment, two output buffers are alternately used and the pipeline processing is performed. Realize processing.

That is, while the waveform processing for one frame is executed and the result is stored in one output buffer, the waveform processing result which has been processed in the previous cycle is output as voice from the other output buffer. Output through the unit 27. The output buffer 26 can be realized by allocating a part of the address of the ring buffer memory (corresponding to the audio memory 2 in FIG. 1) 24.

The data input to the input buffer 22 and the output from the output buffer 26 are performed at the sampling rate intervals of the A / D converter 5 and the D / A converter 6, as in the first embodiment. Therefore, the processing executed by the DSP 1 includes waveform processing in frame units and interrupt processing executed at sampling intervals.

That is, while waveform processing is being performed on one frame of data, interrupt processing is executed many times, and apparently two processings are simultaneously executed.

As the ring buffer memory 24,
A known one is used, and the writing / reading is performed in frame units. The details will be described below.

(Write Operation) In FIG. 11, the voice data input by the voice input unit 21 is held in the input buffer 22. The input buffer 22 is assigned a code length of 16 bits per data, has a capacity to hold the number of data for one frame, and is realized by allocating a part of the address on the audio memory 2 shown in FIG. Has been done.

The control unit 23E shown in FIG. 11 monitors the state of the input buffer 22 and transfers the data for one frame to the audio compression unit 23A every time one frame of audio data is accumulated.

The audio compression unit 23A performs information compression processing on the input audio data for one frame, and stores the compression result data in the ring buffer memory 24. Several methods can be considered for this compression processing. One example is the method of saving the difference data shown in FIG. FIG. 12 is a schematic diagram for explaining the compression process of the audio compression unit 23A of the second embodiment.

In this compression processing, the "difference from the immediately preceding data" is calculated in order from the head data of each frame.
In FIG. 12A, these differences are represented by Δ1, Δ2,
... is expressed. The output data of the compression processing is obtained by dividing the data at the head of the frame into upper 8 bits and lower 8 bits, and then arranging the difference data Δ1, Δ2, ... As one data 8-bit code length. One piece of input data has a digital code length of 16 bits,
In the case of an input signal, which changes sufficiently gently compared to the sampling interval such as an audio signal, the difference from the previous sample value does not become so large, and as shown in FIG. If there is a bit code length, it can be sufficiently expressed. Therefore, the data capacity before and after the compression processing is about half, but the content contained therein is not so large that the difference cannot be expressed by the code length of 8 bits during the processing. There is nothing missing.

The data stored in the ring buffer memory 24 is arranged in the ring buffer memory 24 in such a manner that the data thus compressed to a half capacity on a frame-by-frame basis is maintained in the ring buffer memory 24.

[0129] In addition, so that the break of the frame can be seen,
A frame header is added to the beginning of the compressed data of each frame. The compression processing unit calculates the sum of the absolute values of all data of the frame and records the result as the power value of the frame in the frame header unit in the same manner as the compression processing of FIG. .

The frame to be subjected to the waveform expansion / contraction processing is determined by comparing the power of the frame with the threshold Th. Further, the silent section deletion process is performed by comparing the power of the frame with the threshold value To.

It is desirable to change these threshold values according to the volume of the input voice, rather than using fixed values. For example, when using in a quiet room and when the background noise is large, naturally speaking rate conversion cannot be performed successfully unless these thresholds are adjusted properly.

As a concrete realization method, the maximum / minimum values of the frame power for the past several seconds are stored, and the threshold value is determined based on these values. For example, the time length of one frame is 50 milliseconds (msec),
When it is desired to change these thresholds every 5 seconds, the threshold Th may be changed once every 100 frame processing.

As described above, the power for each frame is
Every time the information is compressed in frame units in the audio compression unit in FIG. 11, all inputs are calculated, and the information is recorded in the frame header and stored in the ring buffer 24.

When the frame power is calculated, the maximum frame power Pmax and the minimum frame power Pmin are compared with each other and updated if necessary. If the maximum frame powers Pmax and Pmin are reset every 5 seconds (every 100 frames), the maximum and minimum frame powers for the past 5 seconds will always remain.

The threshold value is calculated by, for example, Th being 1 between the maximum frame power Pmax and the minimum frame power Pmin.
Set 0% and To to 5%. If expressed by equations, the following equations (Equation 1) and (Equation 2) are obtained.

[0136]

## EQU1 ## Th = | Pmax-Pmin | * 0.10 + Pmin

[0137]

## EQU00002 ## To = | Pmax-Pmin | * 0.05 + Pmin Next, since the method of storing the original (raw) data in the ring buffer memory 24 according to the second embodiment has been described, the details of the silent interval compression will be described. .

As described with reference to FIG. 3 of the first embodiment, the function of the silent interval compression is to delete a silent interval (a period in which the power is smaller than the voice / silent threshold To) for 1 second or longer. is there.

The silence compression processing is performed by the silence compression unit 23B shown in FIG. This silent compression process is a process independent of a process (hereinafter referred to as a main process) executed in units of one frame, which will be described later, and is performed after the main process for one frame is completed. (In FIG. 14, “delay amount = 0” determination (S143) and “PowerON?” Determination (S14
4) (not shown).

In the silence compressor 23B, the input buffer 22
The power is calculated by adding the data accumulated in 1) every fixed unit (for example, 1/4 frame), and the silent compression operation is started when the power "crosses the voice / silence threshold from the bottom to the top". To do. Because when the silent section ends,
This is because it is the time when the power changes from small to large, and unless it is this time, it is not possible to determine whether or not the silent section that had continued until then was for one second or longer.

When the silence compression process is started, first, the frame header of the ring buffer memory 24 is searched backward in the past. The compressed data on the ring buffer memory 24 is compressed in frame units, and the power value of the frame is recorded in the frame header as described above. If a frame with a power lower than To continues for 1 second or more, the silence can be deleted, and the input pointer to the ring buffer memory 24 is returned to the point where the silent section continues for 1 second. The input of the next compressed data is recorded so as to overwrite from the time when it is returned. As a result, the silent section lasting 1 second or more immediately before the present time is always deleted.

(Reading Operation) Main processing of the apparatus of the second embodiment, which will be described later, is performed in frame units. Therefore,
The waveform processing unit 23D shown in FIG. 11 holds the frame data currently being processed, and the reading from the ring buffer memory 24 is performed collectively for each frame. That is, when collectively fetching, the addressing to the ring buffer memory 24 can be performed simply by increasing the address one by one, which is more efficient than fetching data one by one.

Since the data stored in the ring buffer memory 24 is compressed data as described above, it is necessary to decompress this data and restore the original data before performing the waveform processing. The decompression unit 23C shown in FIG. 11 is provided for this purpose. First, the compressed data for one frame is input, and the two leading 8-bit data are arranged in the upper / lower 16 bits to create the leading data. Next, the third value of the compressed data is added to the preceding head data to restore the second data. The next compressed data value is then added to this second data to restore the third data. After that, all the frame data is restored by repeating the operation of adding the compressed data to the previous restored data one after another.

Next, the basic operation of the speech speed conversion apparatus according to the second embodiment will be briefly described.

As shown in FIG. 11, the voice converted into a digital signal by the voice input unit 21 is first input to the input buffer 22. The audio signal read from the input buffer 22 is sent to the audio compression unit 23A in the CPU 23 of the DSP 1 (FIG. 1), the audio signal is subjected to data compression processing, and stored in the ring buffer memory 24. Further, the voice signal is sent to the silence compression unit 23B, and if necessary, the silence compression processing is performed on the data stored in the ring buffer memory 24.

The audio signal data stored in the ring buffer memory 24 is sent to the decompression unit 23C in frame units, and the audio data compressed by the decompression unit 23C is decompressed, and the waveform processing unit (speech speed conversion processing) is performed. Part) 23D. The waveform processing unit (speech speed conversion processing unit) 23D performs a speech speed conversion process and the like based on the conditions set by the function selection unit 25. The digital voice data that has been subjected to the speech speed conversion processing and the like is held in the output buffer 26. The data in the output buffer 26 is read out, and the voice output unit 27 outputs the voice subjected to the voice speed conversion processing and the like.

That is, the data in the output buffer 26 is read out and converted from a digital value to an analog value at time intervals set by the D / A converter 6, as shown in FIG. The analog signal obtained by this conversion is input to the analog amplifier 9 through the low-pass filter 8 and is output as sound by the binaural headphones 325 at the listener's favorite sound pressure level.

Next, processing executed in units of one frame in the second embodiment (hereinafter referred to as main processing)
This will be described with reference to FIGS. 11, 13 and 14.

13 and 14 are flowcharts showing the procedure of the main processing in the second embodiment.

The main processing in the second embodiment is shown in FIG.
As shown in, the power is turned on and the "fade in" process is performed (S131). That is, immediately after the power is turned on,
The data stored in the output buffer 26 is indefinite.
For this reason, there is a possibility that data unrelated to voice may be output immediately after the power is turned on, and if this data is output from the voice output unit 27 as it is, a very large level of noise may occur. In order to prevent this, in the second embodiment, the fade-in process is executed so that the output of the audio output unit gradually becomes loud regardless of the data in the output buffer for a certain period after the power is turned on. The value of the data in the output buffer is adjusted. Specifically, this function is realized by multiplying this data value by a coefficient each time one data is transferred from the output buffer to the D / A converter and changing the coefficient value with time. This operation is executed by the control unit 23E shown in FIG.

Then, the "through mode" process is started. In the through mode process, first, a "reading pointer match" process is performed (S132). The reading pointer matching process is a process of inputting the same data to the output buffer 26 immediately after inputting the data to the input buffer 22 when the data is input from the voice input unit 21. This operation changes the value of the input pointer pointing to the input address on the memory to the input buffer 2
Immediately after inputting the data to 2, it is realized by matching the value of the output pointer indicating the address of the output data on the memory. This operation is performed by the control unit 23E in FIG.

After the reading pointer matches, in the through mode,
The state where the slow switch 104 and the repeat switch 105 are pressed (ON state) is checked (S133, S1).
34) If both are not pressed (OFF state), the process returns to the previous reading pointer matching process (S132), and the through mode is continued. Therefore, in the interrupt processing that occurs while the through mode is continued, the input data is always output as it is, so that the audio output unit 27
Is the same as the input voice.

Each of the slow switch 104, the repeat switch 105, and the reset switch 106 is
In FIG. 11, it is included in the function selection unit 25, and the control unit 23E checks the state.

Repeat switch 1 during the through mode
When 05 is pressed (turned on), a separately prepared repeat flag (not shown) is changed from 0 to 1, and a "read pointer return" routine is performed (S135). FIG. 16 shows a flowchart of the internal processing procedure of this read pointer returning routine. The description of FIG. 16 will be given later.

During the through mode, the slow switch 10
When 4 is pressed, as shown in FIG. 13, the routine jumps to a routine (S136) for performing "parameter setting for expansion processing". FIG. 17 is a flowchart of the processing procedure inside this routine.
Shown in. The description of FIG. 17 will be given later.

After the parameters are set for the expansion process, the waveform expansion or contraction process for one frame is executed (S137). 18 and 19 are flowcharts of the processing means inside the waveform expansion processing for one frame. The description of FIGS. 18 and 19 will be given later.

After the processing for one frame is completed, it is checked whether or not each switch is pressed. By the way, since processing of one frame is completed within the time length of one frame, several tens of milliseconds (msec)
Complete with the order. On the other hand, when the user presses each switch (push button), the switch device is used in the present apparatus so that the pressed state continues for a longer time no matter how short the switch is pressed. Therefore, if the pressed state of the switch is checked every time one frame is processed, the desired operation can be performed with a time delay that does not cause the user to feel a reaction delay.

First, it is checked whether the reset switch 106 is pushed down (S138). if,
The reset switch 106 is pushed down (S13
8)), the mode is forced to shift to the through mode at this point.

If the reset switch 106 is not pushed down (No in S138), as shown in FIG. 14, it is checked whether or not the slow switch 104 is pushed down (S139). If the slow switch 104 is pushed down (if ON: S139
If Yes), the waveform decompression process is continued, and the process returns to the routine for setting parameters in the decompression process so that it is performed in the next frame. If the slow switch 104 continues to be depressed, it will continue to rotate this loop. Further, even when the slow switch 104 is continuously pressed down during the repeat reproduction and the chase reproduction, this loop is continued.

When the slow switch 104 is opened (when OFF: when No in S139), the process goes to the next repeat depressing state determination (S140). At this point, the case where the pressing down of the repeat switch 105 is detected is either “the repeat switch was pressed during the repeat reproduction” or “the repeat switch was pressed during the chasing reproduction”. In either case, the process branches to the read pointer returning routine so that the repeat reproduction is started from the silent portion near the position of the output pointer of the ring buffer memory 24 about 5 seconds after that.

The slow switch 104 is opened,
If the repeat switch 105 has not been pushed down either, the process proceeds to the next repeat end determination (S141). The repeat operation is continuously continued until the output pointer returns from the through mode to the position of the output pointer when the repeat operation is started by pressing down the repeat switch 105. That is, if it is determined in this step that the repeat pointer is currently set and the position of the output pointer does not return to the position of the output pointer at the time when the repeat is started, the waveform expansion / contraction for one frame is performed. A processing loop is formed to return to the processing. The subsequent processing is processing for chasing playback.

After the repeat reproduction and the slow (slow) reproduction are completed, the chase reproduction is started. Chasing playback is a fast-listening playback realized by repeating the waveform shortening process for each frame.
This is an operation to recover the delay from the real time caused by repeat or slow (slow) playback. In the processing of this portion, parameters are set for the waveform shortening processing for the chase reproduction (S142).

The delay amount from the real time increases when the repeat button is pressed and when the waveform expansion processing is executed,
On the contrary, it decreases when the waveform shortening process is executed.

Note that the waveform expansion / compression for one frame, which will be described later,
18 and 19 showing the procedure (flowchart) of the shortening process, the process of increasing / decreasing the delay amount is not shown.

It is determined whether or not there is a delay amount from this real time (S143), and if there is still a delay amount, chasing reproduction is continuously performed, so a processing loop is formed. That is, the operation of continuously performing the chasing reproduction until the delay amount becomes 0 is realized by this determination.

By the way, in the main processing described above,
The time delay from the real time caused by the speech speed conversion or the repeat operation is managed as a "delay amount" using a counter.

The time delay from the real time is the difference between the position on the ring buffer 24 to which the current sample data is input and the position on the ring buffer 24 to which the position of the data being output is input, that is, Although it is possible to manage the difference between the addresses indicated by the two pointers, the present invention uses the method of managing with the delay amount counter as described above. This is because the delay amount may not be correctly expressed by the difference in address between the input / output pointers on the ring buffer 24.

For example, assuming that the memory address space assigned to the ring buffer 24 is from address 0 to address 1000, the ring buffer 24 is realized by treating it as "jump to address 0 after address 1000" in the program. Therefore, when the input / output pointer crosses this address break, it is not possible to express the amount of data in this interval by simply taking the difference between the address values. To know the amount of data between these pointers by calculating the address,
It is necessary to perform address value calculation including complicated case classification based on the process of the two pointers reaching their positions.

In the speech speed conversion apparatus of the present invention, the time delay amount is managed by increasing or decreasing the value of the delay amount counter each time data is read from or written in the ring buffer 24, and the processing amount by the complicated address calculation is controlled. It is preventing the increase.

In the main processing, the power switch is turned off.
Until it is turned on, an infinite loop is performed in which the above processing is repeated (S144).

When the power switch is turned off,
Instead of stopping the processing suddenly, the processing is continued for a certain time and then stopped (mute) (S145). During this time,
Here, processing is performed so that the volume of the output voice gradually decreases.

Specifically, similar to the first fade-in operation, in the interrupt processing, this data value is multiplied by a coefficient every time one data is transferred from the output buffer 26 to the D / A converter 6 shown in FIG. , This function is realized by changing the coefficient value with time. This operation is executed by the control unit 23E shown in FIG.

FIG. 15 is a state transition diagram schematically showing the transition between the modes in the second embodiment described above.
According to this FIG. 15, it will be understood well how the modes are switched in accordance with the switch operation. The standby mode in FIG. 15 will be described in detail later.

The details of the processing operation of each routine of the second embodiment described above will be described below.

FIG. 16 is a flow chart showing the processing procedure of the reading pointer returning routine.

The reading pointer return routine of the second embodiment is a concrete method for changing the value of the output pointer which indicates the position where the data is read from the ring buffer 24, which is necessary to realize the repeat function.

As shown in FIG. 16, first, the current output pointer position is set to Pout (S161). Next, set the delay amount to D from the current real time (set)
Yes (S162).

If the delay amount is already large at this point and the delay amount (B5) is increased by 5 seconds, it is determined whether the size of the ring buffer memory 24 is exceeded (S1).
63), if it is determined that it will be exceeded (Y in S163)
es), leave Pout and D unchanged (S1
69, S170), and this routine ends.

If it is okay to increase the delay amount (B5) by 5 seconds (No in S163), the pointer is returned by 5 seconds (-B5) and the delay amount is increased by 5 seconds (+ B5).
(S164).

Next, a process for searching a silent section in the reverse direction is started so that the start of repeat is a break between words. First, data is accessed in the reverse direction from the position indicated by Pout on the ring buffer memory 24, and the power for one frame is calculated (S165).

At this time, if the output pointer is returned by one frame (F) (-F), the delay amount is further increased by one frame (+ F). However, if the delay amount is increased by one frame, the delay amount is increased. It is determined whether the total amount exceeds the ring buffer memory size (S166). If it is determined that the total amount exceeds the ring buffer memory size (Yes in S166), this silent portion search is stopped, and Pout and Pout D is set as the output pointer value and the delay amount (S169, S1
70) and this routine is completed.

If the total delay amount does not exceed the ring buffer memory size even if the output pointer for one frame is returned (No in S166), Pout is returned for one frame length and the delay amount D is set. Increase by one frame (S
167), the calculated power W for one frame is compared with the sound / silence threshold value (S168). When the power W for one frame is smaller than this threshold value, it is determined that the vicinity of this frame is a word break (in the case of No in S168), and Pout and D at this time are output pointer values. And the delay amount (S169,
(S170), this routine ends.

If the power W for one frame is larger than this threshold value (in the case of Yes in S168), one frame is returned and the silent part search is continued in the same manner to detect a silent part. However, it continues until the delay amount exceeds the ring buffer memory size. In this way, the process of returning the output pointer when the repeat switch 105 is pressed is completed.

FIG. 17 and FIG. 18 are flowcharts showing the waveform expansion / contraction processing procedure for one frame in the second embodiment.

In the waveform expansion / contraction processing for one frame in the second embodiment, as shown in FIGS. 17 and 18, first, the power of the data for one frame this time is calculated (S171). Next, this power value P is compared with a threshold value Th (S172), and for the frame having the power larger than the threshold value Th, the processing described below is performed, and the frame having the power smaller than the threshold value Th is processed. The data is output as it is without any processing and transferred to the ring buffer 24 (S173), or is subjected to the consonant emphasis processing and then transferred to the output buffer 26. Whether to enhance the consonant is determined by the state of the mode switch, which is one of the hidden switches.

As a concrete method of realizing the consonant emphasizing process, for example, a frame having a power equal to or lower than the threshold value Th immediately before a frame having a power equal to or higher than the threshold value Th is regarded as a consonant, and A method of increasing the data value can be considered.

In the frame having the power larger than the threshold value Th in the power judgment (Yes in S172), first, the number of data of one frame is stored in the variable Z representing the unprocessed data amount (S174). , Pitch extraction processing is performed from the beginning of the frame (S175). Several methods are conceivable for the pitch extraction processing. For example, the pitch length at the beginning of the frame is extracted by a well-known algorithm using autocorrelation.

Next, the data amount of the two extracted pitch lengths is compared with the amount of unprocessed data (S176),
If the unprocessed data amount Z is smaller than the data amount of the two extracted pitches, this process is stopped.

When the unprocessed data amount Z is equal to or more than 2 pitches (Yes in S176), first, the pre-transfer process is performed (S178). The pre-transfer process is a process of transferring a part of the input data as it is to the output buffer 26 before the synthetic waveform insertion process described below, and corresponds to the part (b) of FIG. 2 of the first embodiment. To do. The number of data to be transferred in the pre-transfer processing is set in pitch units, but the number differs depending on the waveform expansion rate / shortening rate, and this number Npf is set by the parameter setting routine described later with reference to FIG. 19 (S177). ). After the pre-transfer processing (S178), the unprocessed data amount Z is reduced by the number of transferred data (S179).

Next, according to another parameter Ptri set in the parameter setting routine shown in FIG.
The position to which the Δ window function for generating the composite waveform is applied is determined (S180). The only difference between expansion and contraction is the position on the current waveform on which the window function is applied when the composite waveform is generated using the Δ window function.

That is, in the case of waveform expansion, as shown in FIG. 2, the Δ window function is applied so that a waveform for one pitch is generated from a waveform for two pitches (S181).
On the other hand, in the case of waveform shortening, the Δ window function is applied so that a waveform for 2 pitches is generated from a waveform for 3 pitches or 4 pitches, as shown in FIGS.
By inserting this composite waveform, the delay amount from the real time changes (not shown).

After the synthetic waveform insertion processing, the unprocessed data amount Z is reduced by the number of processed data (S18).
2).

Next, pitch extraction processing is performed again (S183). This is a process corresponding to the fact that the pitch of human voice is constantly changing, and by re-extracting the pitch, the error between the actual pitch length and the pitch length to be processed is reduced, and as a result, This prevents an increase in distortion in the waveform after stretching / shortening.

Next, as shown in FIG. 18, the number of unprocessed data is compared with the data amount of two newly extracted pitch lengths (S184). If the data amount for two pitches does not remain (in the case of No in S184), this process is immediately stopped.

If the data amount of 2 pitches or more remains (Yes in S184), post-transfer processing is performed. The post-transfer process is the same process as the pre-transfer process and corresponds to the part (e) of FIG. 2 in the first embodiment. The number of data to be transferred in the post-transfer process is set in pitch units, but the number differs depending on the waveform expansion rate / shortening rate, and this number Npr is set by the parameter setting routine described in FIG. 19 (S185). After the post-transfer processing (S186), the unprocessed data amount Z is reduced by the number of transferred data (S187).

The above processing is continuously repeated until the processing is interrupted by comparing the data amount for two pitches and the unprocessed amount, which is performed twice during the process.

FIG. 19 is a flow chart showing the processing procedure of a parameter setting processing routine for setting parameters for the decompression processing in the second embodiment.

The parameter setting routine shown in FIG.
Actually, it is used twice in the main processing shown in FIGS. 13 and 14. One of them is performed immediately before the waveform expansion / contraction processing routine for one frame, and the other is
It is used in the "processing for setting parameters for shortening processing" after the end of repeat determination.

By the way, the waveform shortening processing is for realizing a "chasing processing (fast-listening processing)" which is continuously performed after listening slowly or after repeating. If the composite waveform generation using the Δ window function, which was performed during the waveform expansion processing, is performed by shifting the position to which the window function is applied in the opposite direction to the expansion, the waveform is shortened.

In FIG. 19, first, it is determined whether to extend or shorten (S191). This determines either of the above two times.

When setting parameters for the expansion processing, after this determination, the position of the speech speed selection switch is checked (S192), the expansion rate e is set according to the switch position (S193), and the expansion rate e is set. The positions of the parameters Npf and Npr used in the waveform expansion processing are set according to
194, S195), and a parameter P indicating the position where the product-sum calculation with the Δ window performed in the waveform expansion processing is started.
Set tri and end this routine.

On the other hand, when setting parameters for the shortening process, the flow on the right side of FIG. 19 is executed. First, the position of the chase mode switch (one of the hidden switches) is checked (S196), and it is checked whether the chase mode (Mcat) is set to "fly", "quick listening" or "1x" (S197). , S198).

When set to "fly", the chase mode (Mcat) is not actually "chase",
As soon as the slow switch (slow push button) is released, it functions so as to actually fly (S199). Specifically, a branch process forcibly returning to the through mode is performed in this portion.

The chasing mode (Mcat) switch is set to "1".
If it is set to "double", the shortening rate s is set to 1 (S200), and the process proceeds to step S202.

The chasing mode (Mcat) switch is set to "1".
If it is not set to “double”, the shortening rate s is set through the flow in the middle of FIG. 19 in the chase mode (S201), and the parameter Npf used in the waveform shortening process is set according to the shortening rate s. Set the value of Npr (S20
2) Furthermore, the parameter Ptri indicating the position where the product-sum calculation with the Δ window, which is performed during the waveform shortening process, is set (S203), and this routine is ended.

(Embodiment 3) FIGS. 23 and 24 are flowcharts showing the processing procedure of the entire operation of the speech speed conversion apparatus to which the continuous speech speed conversion means of the third embodiment of the present invention is added.

Continuous voice speed conversion in the voice speed control apparatus to which the continuous voice speed conversion means of the third embodiment is added. Basically, the slow switch (slow push button) 104 is continuously pressed and the slow reproduction is continuously performed. That is. However, if waveform expansion is continued at a constant waveform expansion rate, time delays accumulate, and finally, the amount of delay from the actual time exceeds the capacity of the ring buffer 24, and you should continue listening more slowly. Becomes impossible.

Therefore, when performing slow reproduction, it is necessary to mix the period for extending the waveform and the period for shortening the waveform, on the contrary, so that the delay from the real time does not increase steadily. It is a means.

There are several possible methods for switching to the continuous voice speed conversion mode. However, a clear distinction should be made between simply holding the slow switch for a long time and entering the continuous voice speed conversion mode. It is easier to understand, for example, by double-clicking the slow switch (pressing twice at short time intervals) or by pushing the slow switch and sliding it sideways, this can be achieved by using a switch component that locks. .

Each processing in the flowcharts shown in FIGS. 23 and 24 of the third embodiment is exactly the same as the processing procedure in the main processing described in FIGS. 13 and 14 above.

The continuous speech speed converting means of the third embodiment is shown in FIG.
Then, in step S231 in FIG. 24, it is checked whether or not it is the continuous voice speed conversion processing (S231). If it is the continuous speech speed conversion process (Yes in S231), the waveform expansion / contraction process for one frame is performed (S232). next,
It is determined whether the reset switch 106 is pressed (ON) (S233), and the reset switch 1
If 06 is not pressed (OFF), one frame number is counted up (S234), it is determined whether it is the extension period (S235), and if it is the extension period (Y of S235).
es), the process returns to step S232. If it is the extension period (No in S235), parameters are set for the shortening process (S236). Next, it is checked whether the delay amount is 0 (S237), and if the delay amount is 0 (S237).
If Yes), the process returns to step S232. 0 delay
If not (No in S237), parameters are set for the decompression process (S238), the slam counter is reset (S239), the process returns to step S232, and the continuous speech speed conversion process operation is repeated. Step S23
If the continuous voice speed conversion processing is not performed in No. 1 (No in S231), the above-described main processing routine (through mode) is performed.

That is, the continuous speech speed converting means of the third embodiment is a method in which slow reproduction and catch-up reproduction are alternately repeated at a predetermined time interval. According to this method, it is always possible to catch up with the real time at regular time intervals. The management of switching between waveform expansion and waveform shortening is performed by counting the number of frames. For example, after extending the number of frames for about 5 seconds, the shortening process is repeated, and when the delay amount becomes 0, the frame count is set to 0.
Then, the expansion process is repeated again.

To get out of the continuous speech speed conversion mode, the reset switch 106 is pushed down to return to the through mode.

(Embodiment 4) FIGS. 25 and 26 are flow charts showing the processing procedure of the entire operation of the speech speed conversion apparatus to which the continuous speech speed conversion means of the present embodiment 4 different from the embodiment 3 according to the present invention is added. Is.

The continuous speech speed conversion in the speech speed conversion apparatus to which the continuous speech speed conversion means of the fourth embodiment is added is an operation in which a waveform having a high power is subjected to waveform expansion and a frame having a low power is subjected to waveform shortening.

The continuous speech speed converting means of the fourth embodiment is shown in FIG.
Also, in step S251 in FIG. 26, it is checked whether or not it is the continuous voice speed conversion process. If it is the continuous speech speed conversion process (Yes in S251), it is determined whether or not the reset switch 106 is pressed (ON) (S252), and the reset switch 106 is not pressed (OFF). If so, the power for one frame is calculated (S253). Next, it is checked whether the calculated power for one frame is larger than the threshold Th (S254), and if the calculated power for one frame is smaller than the threshold Th (No in S251). In the case of), parameters are set for the shortening process (S256), and the process proceeds to step S257. If the calculated power for one frame is larger than the threshold value Th (Yes in S254), a parameter is set for the decompression process (S25).
5) Perform waveform expansion / contraction processing for one frame (S2
57), the process returns to step S252, and the continuous speech speed conversion processing operation is repeated. In step S251, if it is not the continuous speech speed conversion process (No in S251), the process proceeds to the main process routine (through mode) described above.

That is, when the continuous speech speed conversion mode is entered,
The power for one frame is calculated, and the expansion or contraction is performed for each frame by comparison with the threshold value Th.
To get out of the continuous speech speed conversion mode, the reset switch 106 is pushed down.

According to the fourth embodiment, the voice becomes slow or fast depending on the power of the voice.

Generally, in a normal conversation, the voice tends to be louder in the important portions that the other party wants to hear, and the voices tend to be smaller in the less important portions. The characteristic is that a more natural output voice can be obtained.

However, since the appearance probabilities of the high power portion and the low power portion are not necessarily equal, there is no guarantee that they will always catch up with the real time at a constant time interval as in the case of the third embodiment.

In order to enter the continuous speech speed conversion mode,
As an instruction method from the user, after pushing the slow switch (slow push button) 104, slide it sideways to lock it, or double click the slow switch (slow push button) 104 (continue twice at short time intervals). And push down) is possible. If these methods are used, the difference between the intention of pushing the slow switch (slow push button) 104 to "play slowly" and the intention of "making the operation continuous" and the push method of the same push button are different. Since it can be expressed by, it becomes possible to provide a more intuitive and easy-to-understand operation method as compared with providing a push button for continuous speech speed conversion separately.

In the second, third, and fourth embodiments so far, the "expansion rate" of the waveform when "slowly" is reproduced by the waveform expansion process is determined by the setting of the "speech speed setting switch" provided on the apparatus. As for the "shortening rate" of the waveform when "fast listening" is reproduced by the waveform shortening process, it is assumed that the "deflate value" determined in advance (in the program) is used.
I've explained.

However, if it is possible to change the waveform expansion / contraction ratio by the "accelerator type switch" shown in FIG. 27, the function "to freely switch the voice on the time axis" provided by this device is provided. The user can use it more intuitively.

When the accelerator type switch is set at the center, the through mode of the above-mentioned Embodiments 2, 3 and 4 is executed. When the slide switch is pulled to the near side, waveform expansion processing starts and "slow playback" is executed with a delay from the actual time. Then, press the slide switch to the opposite side,
Conversely, the waveform shortening process is performed (until the delay from the actual time becomes 0), and the fast-listening reproduction is executed.

At this time, the control unit changes the waveform expansion and contraction rate according to the distance from the center of the slide switch. However, as can be seen from the description of FIGS. 20 to 22 of the above-described embodiment, the expansion rate / reduction rate can be set only to some values that can be represented by the ratio of integers. Depending on the distance from the center of
It may be set so that values of several levels can be selected.

When the user releases his / her finger from the accelerator type switch, the lever is automatically returned to the center so that the user keeps the slide switch at a position other than the center. This makes it easier to implement, and thus a more convenient operating method can be realized. In addition,
In order to generate force so that the lever returns to the center, it is possible to provide two springs inside the switch device and to provide mechanical means such as pulling the lever with equal force from both sides. it can.

(Embodiment 5) FIG. 28 is a block diagram showing the functional structure of a speech speed conversion apparatus to which the AV control means of embodiment 5 according to the present invention is added, and FIG.
FIG. 30 is a diagram for explaining the operation of the V control means.
31 and FIG. 31 are flowcharts showing the operation procedure of the main processing of the speech speed conversion apparatus to which the AV control means according to the third embodiment is added.

As shown in FIG. 28, the speech speed conversion apparatus to which the AV control means of the fifth embodiment is added is the same as that shown in FIG.
The AV speed control unit 28 is added to the function structure of the speech speed conversion apparatus of the first embodiment shown in FIG.

The control section 28 judges whether or not the condition for outputting the AV control signal is satisfied, and the AV control section 28 is operated.
Starts / stops the output of control signals.

As shown in FIG. 29, the AV control means is a slow (slow) reproduction or a repeat reproduction so that the delay amount from the real time is a constant amount (3 in FIG. 29).
If it exceeds 0 seconds), the AV control signal is output and the delay amount becomes 0 after chasing playback.
It is software that stops the output of the same signal when it becomes.

The AV control signal is taken out of the apparatus and used for temporarily stopping the reproducing operation of the recording / reproducing apparatus such as a tape recorder or a video recorder. By this means, it becomes possible to continuously and slowly listen to a continuous input voice for a long time that exceeds the capacity of the ring buffer 24 of the present apparatus.

In FIGS. 30 and 31, the portion surrounded by the dotted line is the step showing the operation procedure of the AV control means added to the flow charts of FIGS. 12 and 13, and whether the conditions for outputting the AV control signal are satisfied or not. It is determined whether or not (S301). The AV control signal output is determined by determining whether or not the delay amount from the real time is 30 seconds or more in a loop in which the wavelength expansion / reduction processing for one frame of slow reproduction or repeat reproduction is repeated. (S301), if there is a delay of 30 or more from the real time, the AV control signal output is started (S301).
302).

On the other hand, the process of stopping the AV control signal is a decision to exit the loop of the chase reproduction process. Immediately after the "delay amount = 0" determination (S30
3).

(Sixth Embodiment) FIG. 32 is a view for explaining the arrangement of microphones in a speech speed conversion apparatus according to a sixth embodiment of the present invention, in which 101 is a speech speed conversion apparatus main body, 321 is a microphone, and 322 is a microphone. Stretchable columns for supporting 321 are flexible columns for supporting the microphone 321, flexible columns for supporting the microphone 321, and 324 are electrical cords for electrically connecting the microphone 321 and the speech speed conversion device main body 101 by wire. .

FIG. 33 is a diagram showing a modification of the sixth embodiment, in which 101 is a speech speed converting apparatus main body, 104 is a slow switch, 105 is a repeat switch, 106 is a reset switch, 321 is a microphone, and 324 is a microphone 3.
21 is an electric cord for electrically connecting 21 to the speech speed conversion device main body 101, 325 is an earphone, and 300 is a connecting member.

As for the arrangement of the microphones of the speech speed converting apparatus of the sixth embodiment, as shown in FIG. 32 (a), the microphones 321 are supported by the stretchable columns 322. By supporting the microphone 321 in this manner, the microphone 321 is separated from the speech speed conversion device main body 101, so that it is possible to prevent a cloth rubbing sound when the main body is put in a chest pocket and used.

Further, as shown in FIG. 32 (b), a flexible support column 323 supports the microphone 321. By supporting the microphone 321 in this way, the microphone 32 can be moved from the speech speed conversion device main body 101.
Since 1 is separated and twisted in a free direction,
It is possible to prevent the cloth rubbing noise when the main body is put in the chest pocket and used.

Further, as shown in FIG. 32 (b), the microphone 321 and the speech speed conversion device main body 101 are electrically connected by wire (or wirelessly). In this way, the microphone 321 and the speech speed conversion device main body 101 are electrically connected by wire (or wirelessly), the microphone 321 is separated from the speech speed conversion device main body 101, and the microphone 321 is arranged near the speaker. The S / N ratio can be improved.

Further, as shown in FIG. 33, the speech speed conversion device main body 101, the microphone 321, and the earphone 325 are electrically connected by an electric cord with the connecting member 300 interposed therebetween. Then, operation switches such as the slow switch 104, the repeat switch 105, and the reset switch 106 are provided on the connection member 300. By doing so, it is possible to prevent the cloth rubbing noise when the main body is put in the chest pocket and used, and it is possible to improve the S / N ratio and further improve the usability. .

(Embodiment 7) FIG. 34 is a diagram for explaining a delay time display means of a speech speed conversion apparatus according to Embodiment 7 of the present invention, in which 341 is a display unit and 342 is a display screen.

The delay time display means in the seventh embodiment is
As shown in FIG. 34, it is displayed how much the voice of the speaker lags behind the actual speech speed during the slow reproduction and the repeat reproduction. For example, in FIG. 34, it is assumed that one person image is delayed by 10 seconds, and the time delayed from the present is displayed by the number of person display images. By doing so, the amount of time delay from the present can be visually recognized, and it is possible for both the speaker and the listener to easily adjust the speech speed conversion, so that the speaker can be used conveniently.

For visual display of time delay, for example, a liquid crystal display is installed in the center of the front of the speech speed conversion apparatus main body shown in FIG. 6, and a display screen as shown in FIG. 34 is displayed on the liquid crystal display. Will be realized in. Then, this display unit is controlled by a “liquid crystal display driver” connected to the control unit 23E of FIG. 11 (not shown).

The displayed time delay amount is as shown in FIGS.
In the main process shown by, the delay amount counter is always managed, so the numerical value of the delay amount counter may be converted into units of 10 seconds and the corresponding number of human images may be displayed on the display. . This display operation is performed by the control unit 23E of FIG. 11 through the display driver, but the timing for rewriting the display is sufficient if the processing of one frame is completed. For example, S1 in FIG.
This display process is performed between 37 and S138.

(Embodiment 8) FIG. 35 is a diagram for explaining a power supply device of a speech speed converting apparatus according to an eighth embodiment of the speech speed converting apparatus of the present invention, in which 1000 is a device relating to the speech speed converting apparatus. Part 1, 1 DSP, 5 A / D converter, 6 D / A
Converter, 9 is an analog amplifier, 10 is an analog amplifier,
Reference numeral 1001 is a power source, 1002 is a power supply line, and 1003 is a changeover switch.

The speech speed conversion apparatus according to the eighth embodiment is provided with a standby mode in addition to the through mode as shown in the state transition diagram of FIG. 15, and when the through mode continues for a certain period of time, the standby mode is automatically entered. I am trying. That is,
When either the slow (slow) switch or the repeat switch is pressed (turned on), the clock frequency becomes high and each process is performed.

In the through mode, the DSP 1 operates at a fast clock, but the power is wasted because no processing such as voice speed conversion is performed. Therefore, in the standby mode, the operating clock of DSP1 is dropped,
Power consumption is reduced by only inputting / outputting data.
Then, only save to memory. As a result, the voice memory function is realized.

Further, as shown in FIG. 35, in the analog through mode, the changeover switch 1003 is connected to the contact side for disconnecting the power supply line 1002, and the contact point between the analog amplifier 10 and the analog amplifier 9 is directly connected. Connected to the side, DSP1, A / D converter 5, D / A converter 6
Also, power is not supplied to peripheral digital circuits. At this time, it is not stored in the memory either. That is, the input / output analog system is directly connected to operate as a simple analog amplifier. The changeover switch is a three-stage switch which is on (ON), off (OFF), or intermediate between on (ON) and off (OFF) as shown in FIG. 35, and an analog through mode is provided.

As can be seen from the above description, according to the eighth embodiment, on (ON), off (OFF), on (ON).
Since the switch has a three-stage switch between the OFF state and the OFF state and the analog through mode is provided, it is possible to reduce the power consumption and to expand the range of use of the power source.

(Embodiment 9) FIG. 36 is a diagram for explaining an embodiment 9 in which the voice speed converting means according to the present invention is applied to a telephone, and 2000 is the voice speed converting means according to the present invention.
00 is a telephone main body, 3001 is a handset, and 3002 is a telephone line.

As shown in FIG. 36, the telephone according to the ninth embodiment has a voice speed converting means 2000 according to the present invention inserted between a handset 3001 and a telephone main body 3000. The speech speed conversion means 2000 is, for example, the telephone body 30.
It is configured in a shape like a table on which 00 is placed.

Further, in the case of the handset 3001 of a cordless handset or cordless handset, the voice speed converting means 2000 is inserted between the handset 3001 and the telephone body 3000 by a wireless system.

The speech speed converting means according to the present invention can be used as a speech speed converting means in the exchange and operated by a request from the user.

With this structure, the voice of the telephone can be heard slowly. Also, when calling the old man etc. by feeding back the through voice to the speaker side so that the listener can hear the voice slowly, the speaker side can speak normally, so it is not difficult to talk .

If it is a digital circuit, the A / D means is not required inside the speech speed converting means.

(Embodiment 10) FIG. 37 is a diagram for explaining the embodiment 10 in which the speech speed conversion means according to the present invention is applied to a local broadcast. 2000 is a speech speed conversion means, 321 is a microphone, and 325 is Earphones, 4003 is an amplifier, 4
It is a 004 speaker.

As shown in FIG. 37, the telephone according to the tenth embodiment has a speech speed converting means 2000 according to the present invention inserted between a microphone 321, an earphone 325 and an amplifier 4003 of a speaker 4004.

With this configuration, the listener can listen at an appropriate speech speed even if the speaker does not control the speech speed conversion operation. For example, even if the speaker speaks freely at a high speech speed (already speed), the listener can listen at an appropriate speech speed.

It is also possible to allow the listener to listen to the speaker at an appropriate speaking speed even when speaking slowly.

As can be seen from the above description, the present invention is
The present invention can be applied to telephone fields, telephone exchanges, technical fields requiring speech rate conversion other than in-house broadcasting, such as hearing aids, language learning, overseas travel, and music.

For example, language learning and overseas travel goods can be applied in the following cases.

(1) Listen slowly to the recorded voice.

(2) The expansion rate is changed according to the level improvement.

(3) Listen at a normal speed, and listen slowly by repeating the part you do not understand.

(4) After listening slowly, listen again at the original speed.

(5) Slowly repeat and then imitate.

(6) Imitate and compare the pronunciation with the original voice.

(7) Multiple people can use one source at the same time.
Listen at your favorite speed.

In combination with a digital audio device such as a tape recorder, a CD, an MD, a device having a digital output does not require an A / D converter in the speech speed converter.

Also, for music, the following points can be applied by making changes.・ Do not judge by the power of the decompressed frame (because the tempo comes).・ Make the pitch extraction range wider than that of voice. -Waveform expansion processing is performed at a fixed pitch. In the case of this voice, the pitch is detected and processed at the detected pitch. -Allows conversion operation with a foot switch. This allows you to control while playing the instrument. Although the present invention has been specifically described above based on the embodiments, the present invention is not limited to the embodiments and various modifications can be made without departing from the scope of the invention.

[0270]

The effects obtained by the typical ones of the inventions disclosed in the present application will be briefly described as follows.

(1) Since not only the voice given to the listener unilaterally such as radio voice but also the speech speed conversion device can be used in a situation such as dialogue, the listener's own utterance can be prevented. , The listener can select the voice for which the speech speed conversion is performed.

Further, in a hearing aid, a foreign language learning device, a telephone, etc., it is possible to listen at a slow speech speed without changing the characteristics of the speaker's voice.

(2) The storage device (memory) can be effectively used, the original voice repeat function, the voice memory function, the repeat voice talk speed conversion function, and the fast-listening playback function can be provided.

(3) Since the means for changing the voice speed selected by the voice speed selection switch is provided, the voice speed of the voice heard by the listener can be selected.

(4) Repeat switch is on (ON)
Since the means for repeating the reproduced voice is provided during the operation, the voice speed conversion of the repeat voice can be performed.

(5) Since the chase means for chasing the information stored in the speech speed conversion device to a desired point is provided, the application range of the speech speed conversion device is expanded, the operation time is shortened, and the usability is improved. be able to.

(6) At least one of the voice speed conversion processing switch, the voice speed selection switch, the repeat switch, and the reset switch is provided on one easily-operable peripheral portion on one side of the voice speed conversion device. Therefore, the application range of the speech speed conversion device can be expanded, the operation time can be shortened, and the usability can be improved.

(7) The efficiency of speech speed conversion processing can be improved.

(8) The waveform expansion processing, the shortening processing, and the silent section deletion processing in the speech speed conversion processing are determined by comparing the power of the frame with the threshold value, and the threshold value is input to the speech. Since it is changed according to the size of, the voice speed conversion processing can be performed according to the usage environment condition.

(9) Since the microphone does not pick up the click sound of the switch, the reproduced sound can be heard accurately.

(10) Since the surfaces have different tactile sensations so that it is possible to recognize which switch without looking, the operability can be improved.

(11) Since the means for preventing the rubbing noise of the microphone is provided, the intrusion of noise can be reduced.

(12) Since the display means for visually observing the time delay amount from the present is provided at a predetermined position of the speech speed conversion device, the operation time can be shortened and the usability can be improved.

(13) Since the ring buffer is used as the storage means and the means for managing the delay time is provided by the counter indicating the time delay on the ring buffer, the repeat processing, the chasing processing, etc. are calculated by complicated pointer address calculation. Can be done easily.

(14) Since the standby mode and the analog through mode are provided in addition to the through mode, low power consumption can be achieved.

(15) Since the power supply switch has three stages of ON (ON), OFF (OFF), and ON and OFF intermediate and the analog through mode is provided, it is possible to reduce the power consumption and use the power supply. The range can be expanded.

(16) Since the speech speed converting means is provided between the handset of the telephone and the main body of the apparatus, the listener can select the voice to be subjected to the speech speed conversion without disturbing the listener's own speech.

(17) With the telephone, it is possible to listen at a slow speaking speed without changing the characteristics of the speaker's voice.

(18) Since the speech speed converting means is provided in the telephone exchange, the listener can select the voice to be subjected to the speech speed conversion without disturbing the listener's own speech.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an internal circuit of a first embodiment according to the present invention.

FIG. 2 is a diagram for explaining a voice speed conversion process executed in the DSP according to the first embodiment.

FIG. 3 is a diagram for explaining the concept of threshold value processing according to the first embodiment.

FIG. 4 is a diagram showing a usage pattern of the speech speed conversion apparatus according to the first embodiment.

FIG. 5 is a flowchart showing a control procedure of the speech speed conversion apparatus according to the first embodiment.

FIG. 6 is a front plan view seen from the front of a speech speed conversion device according to a second embodiment of the present invention.

FIG. 7 is a rear plan view of the speech speed converting apparatus according to the second embodiment as viewed from the rear side.

FIG. 8 is an upper plan view seen from above the speech speed conversion apparatus according to the second embodiment.

FIG. 9 is a left side plan view of the speech speed conversion apparatus according to the second embodiment as viewed from the left side.

FIG. 10 is a right side plan view of the speech speed conversion apparatus according to the second embodiment as viewed from the right side.

FIG. 11 is a block diagram showing a functional configuration of a speech speed conversion apparatus according to the second embodiment.

FIG. 12 is a schematic diagram for explaining compression processing of an audio compression processing unit according to the second embodiment.

FIG. 13 is a flowchart illustrating a procedure of main processing according to the second embodiment.

FIG. 14 is a continuation of the flowchart of FIG.

FIG. 15 is a state transition diagram schematically showing a transition between each mode in the second embodiment.

FIG. 16 is a flowchart illustrating a processing procedure of a reading pointer returning routine according to the second embodiment.

FIG. 17 is a waveform length for one frame in the second embodiment.
It is a flowchart which shows the shortening process procedure.

FIG. 18 is a continuation of the flowchart of FIG.

FIG. 19 is a flowchart showing a parameter setting processing procedure in the second embodiment.

FIG. 20 is a diagram for explaining a data compression process according to the second embodiment.

FIG. 21 is a diagram for explaining a data compression process according to the second embodiment.

FIG. 22 is a diagram for explaining a data compression process according to the second embodiment.

FIG. 23 is a flowchart showing the processing procedure of the overall operation of the speech speed conversion device to which the continuous speech speed conversion means according to the third embodiment of the present invention is added.

FIG. 24 is a continuation of the flowchart of FIG. 23.

FIG. 25 is a fourth embodiment different from the third embodiment according to the present invention.
6 is a flowchart showing a processing procedure of the overall operation of the voice speed conversion device to which the continuous voice speed conversion means of FIG.

FIG. 26 is a continuation of the flowchart of FIG.

FIG. 27 is a schematic diagram for explaining an accelerator type switch used in the continuous voice speed converting means in the fourth embodiment.

FIG. 28 is a block diagram showing a functional configuration of a speech speed conversion device to which an AV control means according to a fifth embodiment of the present invention is added.

FIG. 29 is a diagram for explaining the operation of the AV control means of the fifth embodiment.

FIG. 30 is a flow chart showing a main processing procedure of the speech speed conversion apparatus to which the AV control means of the fifth embodiment is added.

FIG. 31 is a continuation of the flowchart of FIG. 30.

FIG. 32 is a diagram for explaining the arrangement of microphones in the speech speed conversion device according to the sixth embodiment of the present invention.

FIG. 33 is a diagram showing a configuration of a modified example of the sixth embodiment.

FIG. 34 is a diagram for explaining a delay time display means of the speech speed conversion apparatus according to the seventh embodiment of the present invention.

FIG. 35 is a diagram for explaining a power supply device of a speech speed conversion device according to Example 8 of the present invention.

FIG. 36 is a diagram for explaining the ninth embodiment in which the speech speed converting means according to the present invention is applied to a telephone.

[Fig. 37] Fig. 37 is a diagram for explaining the tenth embodiment in which the speech speed conversion means according to the present invention is applied to a local broadcast.

[Explanation of symbols]

1 ... DSP (Digital Signal Processor), 11 ...
Software for performing speech rate conversion processing, 12 ... Serial port, 13 ... External interrupt flag terminal, 14 ... Flag register, 2 ... Voice memory, 3 ... Selector switch, 4 ...
PTL switch, 5 ... A / D converter, 6 ... D / A converter, 7 ... Low-pass filter, 8 ... Low-pass filter, 9
... Analog amplifier, 10 ... Analog amplifier, 321, ... Microphone, 325 ... Binaural headphones (earphones),
101 is the main body of the speech speed conversion device, 102 is a back cover, 103 ...
Depression for finger rest, 104 ... Slow switch (slow push button), 105 ... Repeat switch (repeat push button), 106 ... Reset switch (reset push button), 108 ... Volume control volume, 109 ... Power switch, 110 ... Earphone terminal, 111 ... External input terminal, 1
12 ... AV control terminal, 113 ... Speech speed changeover switch (speech speed setting switch), 21 ... Voice input section, 22 ... Input buffer, 23 ... Central processing section (CPU), 24 ... Ring buffer memory, 25 ... Function selection section , 26 ... Output buffer, 27 ... Audio output section, 28 ... AV controller.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Yasunori Kawauchi 1410 Inada, Katsuta-shi, Ibaraki Hitachi Ltd. AV Equipment Division (72) Inventor Nobuo Hataoka 1-280, Higashi Koikeku, Kokubunji, Tokyo Hitachi Central In the laboratory (72) Inventor, Juichi Morikawa, 292, Yoshida-cho, Totsuka-ku, Yokohama, Kanagawa

Claims (32)

[Claims] 1. A voice speed conversion method for inputting voice and changing only the speed of voice without changing the pitch of the input voice, wherein a voice speed conversion is performed at a time specified when a listener needs the voice speed conversion. The speech speed conversion method is characterized in that the speech speed conversion processing of the input voice is performed only during the period, and the speech speed conversion is not performed during the other time. 2. A voice speed conversion having a means for inputting a voice, a voice speed conversion processing means for changing the speed of the input voice, and a means for outputting the output of the voice speed conversion processing means to a listener's ear. In the apparatus, the speech speed conversion device is provided with a speech speed conversion processing switch, and the speech speed conversion processing switch is turned on (O
N) The voice speed of the input voice is changed and output only during the period of time N), and is output without changing the voice speed of the input voice while the switch for the voice velocity conversion processing is off (OFF). A speech speed conversion device comprising means. 3. A speech speed conversion method for encoding and accumulating original speech, reading out the accumulated encoded speech, and changing only the speed of the speech without changing the pitch of the original speech, A speech speed conversion method characterized in that the speech speed conversion processing of the input voice is performed only during a designated time when the speech speed conversion is required, and the speech speed conversion is not performed during the other time. 4. A means for inputting an original voice, a storage means for encoding and accumulating the input voice, and a speech speed conversion processing means for reading out the accumulated encoded voice and changing the speed of the input voice. A voice speed conversion device having a means for outputting the output of the voice speed conversion processing means to a listener's ear as a voice, wherein the voice speed conversion device is provided with a voice speed conversion processing switch. The voice speed of the input voice is changed and output only while the switch is on (ON), and the voice speed of the input voice is changed while the voice speed conversion processing switch is off (OFF). A speech speed conversion device characterized in that a means for outputting without being provided is provided. 5. The voice speed conversion apparatus according to claim 4, wherein the storage unit has a unit for storing in frame units. 6. A method for determining waveform expansion / contraction processing in the speech speed conversion processing by comparing frame power with a threshold value, wherein the threshold value is variable. The speech speed conversion device according to claim 5. 7. The voice speed conversion device is provided with a voice speed selection switch for selecting a voice speed, and means for changing the voice speed selected by the voice speed selection switch is provided. 7. The speech speed conversion device according to any one of 2, 4 to 6. 8. The voice speed control apparatus according to claim 7, wherein said voice speed control apparatus is provided with means (AV control) for controlling an audio / video device. 9. The speech speed conversion device is provided with a repeat switch, and means for repeating the reproduced voice while the repeat switch is on (ON) is provided. 9. The speech speed conversion device according to any one of items 8 to 8. 10. The repeat means is means for backing up for several seconds each time it is pressed, means for occasionally generating an intermittent sound while returning, means for preventing further returning after reaching the end of the ring buffer, and repeat time. 10. The speech speed conversion apparatus according to claim 9, further comprising at least one means for selecting the speech speed of the above. 11. The means for selecting the speech speed at the time of repeat includes at least two or more of a default value of repeat, a slow repeat, a fast-listening repeat, and a repeat gradually increasing. Item 10. The speech speed conversion device according to item 10. 12. In the speech speed conversion device, when a delay from real time occurs due to a speech speed conversion or a repeat operation, while reproducing the stored information,
12. The speech speed conversion apparatus according to claim 2, further comprising a chasing unit that adjusts the delay amount. 13. The chasing means is means for starting chasing when the slow reproduction mode ends, a means for starting chasing when reproducing to a repeat start time after a repeat, a means for selecting a speech speed at the time of chasing, and a case of catching up. 13. The speech speed conversion apparatus according to claim 12, further comprising at least one of a unit that automatically shifts to a through mode that outputs the input voice as it is and a unit that generates a notification signal sound (message) when catching up. . 14. The means for selecting the speech speed at the time of chasing includes at least one of a means for skipping to the reality at once, a means for chasing the reality by listening fast, and a means for moving in parallel with a delay. Claim 13
The speech speed conversion device described in. 15. At least one of the voice speed conversion processing switch, the voice speed selection switch, the repeat switch, and the reset switch is provided on one easily accessible peripheral portion on one side surface of the voice speed conversion device. The speech speed conversion device according to any one of claims 2, 4 to 14, characterized in that. 16. The reset switch has means for stopping the operation when the switch is turned on during the repeat operation or the chase operation, skipping the operation, and then shifting to the through mode. The speech speed conversion device described in. 17. The speech rate conversion processing means is provided as software executed by a digital signal processor having a terminal for inputting an interrupt request signal from the outside, and the speech rate conversion processing switch switches the speech rate. 17. Control of conversion processing or switching of speech speed conversion speed is given to a digital signal processor through a terminal for inputting the interrupt request signal, according to any one of claims 2, 4 to 16. Speech speed converter. 18. The voice speed conversion apparatus according to claim 2, further comprising means for listening to an output voice of the voice speed conversion apparatus through binaural headphones. 19. A microphone for converting an acoustic signal into an electric signal, an analog amplifier for amplifying the microphone output, a low-pass filter for removing high frequency components of the output of the analog amplifier, and an analog signal for the output of the low-pass filter as a digital signal. To an A / D converter, a digital signal processor that executes a process of changing a voice speed by digital signal processing, a storage unit that stores input voice data and data of a signal processing result, and a digital signal processor of the digital signal processor. A means for controlling the processing for changing the speed of the sound to be performed, a means for changing the processing parameters, a D / A converter for converting the digital sound data into an analog value, and a high frequency component of the output of the D / A converter. The second low-pass filter that removes the A speech speed conversion device comprising: a wide second analog amplifier; and a headphone which converts an output of the second analog amplifier into an acoustic signal and applies it to both ears. 20. A microphone for converting an acoustic signal into an electric signal, an analog amplifier for amplifying the microphone output, a low-pass filter for removing high frequency components of the output of the analog amplifier, and an analog signal for the low-pass filter output as a digital signal. To A / D converter, storage means for storing input voice data and data of signal processing result, and digital signal for executing processing for reading the stored information and changing the speed of voice by digital signal processing. A processor, a means for controlling the processing for changing the speed of sound performed by the digital signal processor, a means for changing processing parameters, a D / A converter for converting digital sound data into an analog value, and the D
A second low-pass filter that removes high-frequency components of the output of the A / A converter, a second analog amplifier that amplifies the output of the second low-pass filter, and an output of the second analog amplifier that is converted into an acoustic signal. A speech speed conversion device having headphones for both ears. 21. The speech speed conversion processing means is performed by pipeline processing in frame units using a plurality of input frame buffers, and for each frame of data, first, pitch extraction processing is performed on the beginning of the frame. Then, the pitch of that portion is detected, the detected data of one pitch length is transferred to the output buffer, and the window function that changes from 0 to 1 and the data of 1 pitch to 0 for the data of two pitch lengths. Multiply the changing window function, add the data of the results of applying each window function to create a composite waveform with a time length of 2 pitches, insert it after the 1 pitch of data that was transferred earlier, and The pitch detection process is performed again starting from the position separated by 2 pitches from the position on the data subjected to the pitch extraction process, the pitch is detected at that position, and the pitch length obtained by the last pitch detection is taken as a unit. 21. The speech speed conversion apparatus according to claim 19, wherein a series of procedures for transferring data for n (n is an integer) pitch to the output buffer is repeatedly performed over the entire frame. 22. The speech speed conversion processing means calculates an average power of data in an input frame, is executed only when the average power is larger than a preset threshold value, and when it is smaller, 22. The speech speed conversion apparatus according to claim 21, wherein the data included in the frame is transferred to the output buffer as it is. 23. A threshold value is provided in the threshold processing for the average power of data in the input frame, and a frame having an average power smaller than the second threshold value has a preset time. When the data continues for a time longer than the threshold value, the transfer of data of a frame having a smaller average power than the second threshold value that continues beyond the time threshold value to the output buffer is prohibited. 23. The speech speed conversion device according to claim 22, wherein: 24. The switches according to claim 2, 4 or 1, wherein each of the switches is a switch having a soft touch so that a microphone does not pick up a click sound of the switch.
8. The speech speed conversion device according to any one of 8. 25. The speech speed conversion device according to claim 24, wherein each of the switches has a surface form having a different tactile sensation so that the switch can be seen without looking. 26. A cloth rubbing noise preventing means is provided for changing the distance between the microphone and the apparatus main body so that the microphone and the clothes do not come into direct contact when the apparatus main body is put in a chest pocket for use. The speech speed conversion device according to any one of claims 2, 4 to 25. 27. The display means for visually observing a time delay amount from the present is provided at a predetermined position of the speech speed conversion device, according to any one of claims 2, 4 to 26. The described speech speed conversion device. 28. A ring buffer is used as the storage means, and means for managing the delay time is provided by a counter indicating a time delay on the ring buffer. The speech speed conversion device described in the item. 29. In addition to the through mode, a standby mode for lowering the clock cycle of the processor and performing the same processing as in the through mode is provided, according to any one of claims 19 to 28. Speed converter. 30. The power switch has three stages of ON, OFF, ON and OFF intermediate, and when the switch is adjusted to the intermediate position, the analog input / output systems in the audio signal processing system are short-circuited to each other. 30. A power supply means for operating in an analog through mode for stopping power supply to a digital processing system between analog input / output systems is provided. Speed converter. 31. A telephone set, comprising the speech speed converting means according to claim 2, provided between the handset of the telephone set and the main body of the telephone set. 32. A telephone exchange, wherein the speech speed converting means according to any one of claims 2, 4 to 30 is provided in the telephone exchange.