JPH1152985A - Fast listening device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide the fast listening device in which voice recognition is easily conducted against the voice, that is fast forward reproduced with a two- fold speed, a two and half-fold speed, a three-fold speed and an N-fold speed (where N is a positive discrete value), and the contents of the reproduced voice is accurately understood and the broadcasts, that are conducted in two languages, are simultaneously listened. SOLUTION: In the device, a voice analysis digital signal processing device 11 connected to an analog/digital converter 53 conducts a voice analysis with a frame period T and the analyzed data are temporarily stored in an input FIFO memory buffer 54. A microcontroller 55 discriminates whether the data stored in the buffer 54 are for voiced sound or not and conducts an N-fold high speed converting process. If the data are for voiced sound, the data are stored in an output FIFO memory buffer 12. A voice synthesis digital signal processor 13 conducts a voice synthesis with a T/N frame period and a digital signal processor 14 corrects the data to produce continuous data.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fast hearing device used in electronic equipment for performing audio processing.

[0002]

2. Description of the Related Art Three conventional hearing devices using the conventional hearing technology will be described below with reference to the accompanying drawings. << Conventional Example 1 >> FIG. 4 is a block diagram showing a fast-listening tape recorder of Conventional Example 1. The tape recorder shown in FIG. 4 has a fast listening technique realized by turning the tape recorder fast. FIG. 4 shows an example in which the tape recorder runs at twice the speed of the magnetic tape 42 by the servomotor 41 at a normal speed, and reproduces at a double speed of the normal reproduction. In the conventional example 1, the magnetic tape 42 as the recording medium runs at twice the speed as compared with the normal reproduction.
Is doubled per unit time. The magnetic change 42m is converted by the head 43 into an electric signal 43e. The electric signal 43 e is amplified by the analog amplifier 44 and input to the audio output device 45. The audio output device 45 outputs a signal from the analog amplifier 44 to a sound wave 45a audible to humans.
And output this. At this time, the duration 45t of the sound wave 45a is 1 / compared to the duration during normal reproduction.
It becomes 2. For example, if the word “good morning” takes t seconds during normal playback, t /
2 seconds. As described above, the double speed reproduction enables the double speed listening.

FIG. 5 is a block diagram showing a video tape recording / reproducing apparatus capable of fast listening according to a conventional example 2 and a waveform diagram showing an example of an input audio signal and an output audio signal. In the video tape recording / reproducing apparatus of the conventional example 2, the head 5
2 receives a video magnetic signal 51mv (video signal) and an audio magnetic signal 51ma (audio signal) from the video tape 51, and receives a corresponding video electric signal 52ev and audio electric signal 5ma.
2ea. The audio electric signal 52ea output from the head 52 is input to an analog / digital converter 53, and is converted into a digital signal 53e sampled at a sampling frequency T. This digital signal 53e is supplied to a FIFO (first-in first-out) memory buffer 54.
Is temporarily stored. The microcontroller 55 connected to the FIFO memory buffer 54 stores audio data 54 d in the first memory of the FIFO memory buffer 54.
Is read. Further, the microcontroller 55 receives a speed instruction signal 56e which is an instruction of normal reproduction or double speed reproduction set by the user in the input / output control device 56. The digital data 55d output from the microcontroller 55 is converted into a digital / analog converter 57
And converted into an analog signal 57e. The converted analog signal 57e is input to the amplifier 58 and amplified. The amplified signal 58e is converted into an audible sound wave 59a in an audio output device 59 and transmitted to human hearing.

Next, the operation of the video tape recording / reproducing apparatus of the prior art 2 at the time of normal speed reproduction will be described. Upon receiving a normal reproduction instruction set by the user on the input / output device 56, the microcontroller 55
The data in the FO memory buffer 54 is output to the digital / analog converter 57 at the sampling frequency T. In other words, the microcontroller 55 outputs the digital data 55d to the digital / analog converter 57 so that the analog signal 57e exactly similar to the audio electric signal 52ea picked up by the head 52 is output from the digital / analog converter 57. . Therefore, the audio output device 59 via the amplifier 58
The audible sound wave 59a output from the audio electric signal 52e
This is a signal similar to a. As described above, at the time of normal speed reproduction, an audible sound wave 59a similar to that at the time of general video reproduction is output from the audio output device 59 as a reproduction sound.

Next, the operation of the video tape recording / reproducing apparatus of the second prior art at the time of reproducing at double speed will be described.
In the case of 2 × speed reproduction, the video from the video tape 51 is reproduced at 2 × speed and transmitted to human vision. On the other hand, the audio from the video tape 51 picked up by the head 52 is transmitted to an analog / digital converter 53.
And is converted into a digital signal 53e sampled at the sampling frequency T. The digital signal 53e is temporarily stored in the FIFO memory buffer 54. When the input / output control device 56 issues a double speed instruction, the input / output control device 56 outputs a speed instruction signal 56 e to the microcontroller 55.

[0006] First, the microcontroller 55
The digitized audio data 54d temporarily stored in the O memory buffer 54 is read, and it is determined whether or not there is no audio. That is, if the audio data 54d is equal to or smaller than the threshold value (threshold value), the microcontroller 55 regards the absence of audio as "no audio", discards the audio data 54d, and immediately reads the next data. go. At that time, the microcontroller 55 does not output the data to the digital / analog conversion device 57.
This operation is repeated until data having a value larger than the threshold value is read. When the microcontroller 55 detects the audio data 54d having a value larger than the threshold set value, the microcontroller 5
Reference numeral 5 denotes digital data 55d as audio data of 2T /
The data is output to the digital / analog conversion device 57 at a triple sampling period. FIG. 5B shows an input audio signal during normal reproduction, and FIG. 5C shows an output audio signal during normal reproduction. FIG. 5D shows an input audio signal at the time of double speed reproduction, and FIG. 5E shows an output audio signal at the time of double speed reproduction. Thus, at the time of 2 × speed reproduction, the sound is 1.5 times.
It is configured to reproduce at the same time and further cut off the silent section and output the next sound continuously. For this reason,
The video recording / reproducing apparatus of Conventional Example 2 has realized pseudo double-speed fast-forward reproduction.

<< Conventional Example 3 >> FIG. 6 is a block diagram showing an example of an audio multiplex television receiver of Conventional Example 3. In FIG. 6, when the audio multiplex television receiver 61 receives the audio multiplex broadcast, the audio multiplex television receiver 61 outputs the Japanese audio 61a.
J and English voice 61aE are simultaneously output in two languages, so that the user can listen.

[0008]

The prior art 1 shown in FIG.
When performing double-speed playback on a tape recorder capable of fast listening, the audio playback time during double-speed playback is t / 2, where t is the audio playback time during normal playback, and the audio playback time during normal playback. Half of The sound wave frequency output from the audio output device 45 is 2 fkHz, assuming that the sound wave frequency during normal reproduction is fkHz. For this reason, in the fast listening device of the conventional example 1, when listening to the sound at the time of double speed reproduction, the frequency is twice the frequency of a normal human voice, and most people cannot recognize the language from the sound, There was a problem that the meaning could not be understood.

The video tape recording / reproducing apparatus of the second conventional example shown in FIG. 5 corresponding to high-speed listening is not a simple double-speed reproduction but a general one that can be heard by humans in order to solve the above-mentioned problem of the first conventional example. .5 times reproduction is performed first. That is, assuming that the normal reproduction time is t, reproduction at 1.5 times speed is performed with a reproduction time of 2t / 3. Further, the video tape recording / reproducing apparatus of the second conventional example is configured to cut a non-voice section and continuously reproduce the next sound in the section. Thus,
1.5 to 2 times speed is realized. Thus, for example,
When audio having no silent section is input from the video tape 51, the video is played back at twice, but the audio is played back only at 1.5 times, and the audio data is temporarily stored in the FIFO memory buffer 54. Exceeds capacity and overflows. In such a case, there is a problem that the video tape recording / reproducing device of the second conventional example outputs a sound cut off from the sound output device.

Also, in the audio multiplex television receiver of the third conventional example shown in FIG. 6, when a mode is set to simultaneously listen to the Japanese voice 61aJ and the English voice 61aE, even if a user who has the ability to hear Japanese and English hears. Even so, there is a problem that Japanese and English voices are output at the same time, and both languages cannot be heard, or only one of the languages can be heard. The present invention provides N ×
An object of the present invention is to provide a fast-listening device that can easily recognize output sound and accurately understand the content when fast-forward reproduction is performed at double speed (N is a positive discrete number). Another object of the present invention is to provide a fast listening device capable of simultaneously listening to bilingual broadcasts.

[0011]

To achieve the above object, a fast hearing device of the present invention is provided with an analog signal processing device for converting an input audio signal into an analog signal, and an analog signal processing device connected to the analog signal processing device. An analog / digital conversion device, a voice analysis digital signal processing device connected to the analog / digital conversion device for performing voice analysis at a frame period T, and an input for temporarily storing data from the voice analysis digital signal processing device. A FIFO memory buffer, a microcontroller for comparing data recorded in the input FIFO memory buffer with a predetermined value and selecting desired data, an output FIFO memory buffer for recording data selected by the microcontroller, The data recorded in the output FIFO memory buffer is A voice synthesizing digital signal processor for synthesizing voice in T / N (N is a positive discrete number); a digital signal processor for correcting data from the voice synthesizing digital signal processor into continuous waveform data; A digital / analog converter connected to a digital signal processor; an amplifier for amplifying an output signal of the digital / analog converter; and an audio output device for converting a signal from the amplifier to an audible sound wave. ing.

[0012] Also, a fast hearing device of the present invention comprises a plurality of analog signal processing devices for converting a plurality of types of input audio signals into analog signals, and a plurality of analog signal processing devices connected to the plurality of analog signal processing devices, respectively. A / D converter, a plurality of voice analysis digital signal processors connected to each of the plurality of analog / digital converters and analyzing a voice signal at a frame period T, and the plurality of voice analysis digital signal processors A plurality of input FIFO memory buffers in which data analyzed in the above are temporarily recorded, a control panel for performing desired settings,
A microcontroller that analyzes each data primarily recorded in the plurality of input FIFO memory buffers, converts the data according to a command of the control panel, and switches the plurality of converted audio signals in a time-division manner;
An output FIFO memory buffer for temporarily recording data formed in the microcontroller, and a speech synthesis digital signal for speech synthesis of data recorded in the output FIFO memory buffer at a frame period T / N (N is a positive discrete number) A processing device; a digital signal processing device for correcting data from the voice synthesis digital signal processing device into continuous waveform data; a digital / analog conversion device connected to the digital signal processing device;
An amplifier for amplifying an output signal of the digital / analog converter, and an audio output device for converting a signal from the amplifier to an audible sound wave are provided.

Further, in the fast hearing device of the present invention, the voice analysis digital signal processing device performs PARCOR analysis (Percoll analysis) at a frame period T, and the voice synthesis digital signal processing device performs PARCOR voice synthesis at a frame period T /. Going in N.

The rapid hearing device of the present invention uses a DSP (Digital Signal Processor) or HSP (Host Signal Processor) to halve the frame period without frequency change, thereby doubling the frame period. By synthesizing the fast-talking voice, the double-speed voice without omission can be realized. The technique of the fast hearing device of the present invention also enables reproduction of an integer multiple or a positive discrete multiple of a triple speed, a quadruple speed, or higher.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments 1 and 2 which are embodiments of the fast hearing device of the present invention will be described below with reference to FIGS. Specific examples of products to which the fast hearing technology of the fast hearing device of the present invention is applied include a tape recorder, a music CD (compact disc) recording / reproducing device, and the like.
MD (mini-disc) recording and playback device, radio receiver, terrestrial TV receiver and audio multiplex TV receiver, satellite and analog TV receiver and digital TV receiver, video recording and playback device, musical sound An electronic musical instrument, a wired telephone, a wireless portable telephone, a PHS (personal handy phone system), a wireless communication device, an Internet telephone, a video telephone conference system, and the like for the purpose of outputting can be given. Further, the rapid hearing technology in the fast hearing device of the present invention is a technology that will be applied to all audio devices and musical sound devices in the future.

<< First Embodiment >> A fast listening device according to the first embodiment is a video tape recording / reproducing device capable of fast listening.
And a fast listening device that solves the problem in Conventional Example 2. That is, the fast hearing device according to the first embodiment is a fast hearing device that can easily recognize the content from the sound at the time of double speed reproduction, and outputs the sound at the time of double speed reproduction without interruption. FIG. 1 is a block diagram showing a video tape recording / reproducing apparatus capable of fast listening according to the first embodiment. In the video tape recording / reproducing apparatus of the first embodiment, the head 52 receives the video magnetic signal 51mv (video signal) and the audio magnetic signal 51ma (audio signal) from the video tape 51 and converts them into the corresponding video electric signal 52ev and audio electric signal 52ea. Convert. The audio electric signal 52ea output from the head 52 is input to an analog / digital converter 53, and is converted into a digital signal 53e sampled at a sampling frequency T.

An analog / digital conversion device (hereinafter referred to as A
A / D converter 53 is connected to the audio analysis digital signal processor 11, and the digital signal 53 e output from the A / D converter 53 is converted into a frame cycle of the sampling frequency T by the audio analysis digital signal processor 11. Analyzed and converted to pitch (fundamental frequency) 11p
And amplitude 11A and PARCOR coefficient (Percoll coefficient) 1
It is converted to 1k1 to 11ki. These data are temporarily stored in an input FIFO (first-in first-out) memory buffer 54 as a pitch 54p, an amplitude 54A, and PARCOR coefficients (Percoll coefficients) 54k1 to 54ki. FIF
The microcontroller 55 connected to the O memory buffer 54, for example, when instructed by the input / output control device 56 to cut off the reproduction during the silent period, when the pitch 54p, the amplitude 54A, and the PARCOR coefficient 54k1
Only the amplitude 54A is read out of one pair of data consisting of .about.54ki and is compared with a threshold set value (threshold). If the amplitude 54A is smaller than the threshold setting value, the FIFO memory buffer 54 determines that there is no sound, and does not output the data of this pair to the output FIFO memory buffer 12. Conversely, if the amplitude 54A is larger than the threshold value, the microcontroller 5
5 outputs the data of the pair to the output FIFO memory buffer 12. At this time, the output FIFO memory buffer 12 stores the pitch 54 of the input FIFO memory buffer 54.
p, amplitude 54A, PARCOR coefficient 54k1-54ki
Are temporarily stored with the pitch 12p, the amplitude 12A, and the PARCOR coefficients 12k1 to 12ki corresponding to.

When the microcontroller 55 informs the voice synthesizing digital signal processing device 13 of the double speed fast forward according to the instruction of the input / output control device 56, the pitch 12p temporarily stored in the output FIFO memory buffer 12, Amplitude 12A, PARCOR coefficient 12k1-12
ki is input to the speech synthesis digital signal processing device 13 in the order of chronological order. At this time, the voice synthesizing digital signal processing device 13 performs voice synthesis at a frame period of T / 2 according to an instruction from the microcontroller 55. Since the output of the voice synthesis digital signal processing device 13 has a waveform including discontinuous points, the digital signal processing device 14 corrects the output into data having a continuous and smooth waveform. The output signal 14d of the digital signal processor 14 is input to a digital / analog converter 57 and converted into an analog signal 57e. The converted analog signal 57
e is amplified in the amplification device 58. The amplified signal 58e is converted into an audible sound wave 59a in an audio output device 59 and transmitted to human hearing. The sound that is the audible sound wave 59a at this time is reproduced as a twice-speed fast sound synchronized with the double speed of the screen.

Embodiment 2 Next, Embodiment 2 of the fast hearing device of the present invention will be described. The fast hearing device according to the second embodiment is an audio multiplex television receiver capable of fast listening, and is a fast hearing device that solves the above-described problem of the third conventional example. That is, the fast hearing device of the second embodiment can easily recognize the content from the sound at the time of double speed reproduction, and the sound at the time of double speed reproduction is output without interruption. Is a sound multiplex television receiver that can hear both sounds. 2 and 3 are block diagrams showing a typical example of a fast listening device of an audio multiplex television receiver. The audio multiplex television receiver according to the second embodiment is described by dividing into FIG. 2 and FIG. FIG.
, The Japanese broadcast input device 31 of the audio channel 1
Contains the audio signal 3 of the Japanese broadcast separated from the TV signal
1e is input, and the English broadcast audio signal 21e separated from the television wave is input to the English broadcast input device 21 of the audio channel 2. Figure 2 shows an example of each signal waveform
It describes in. Next, the audio signal 31e from the Japanese broadcast input device 31 is input to an analog signal processing circuit 32 such as a low-pass filter, and high-frequency noise is cut.
The signal 21e from the English broadcast input device 21 is input to an analog signal processing circuit 22 such as a low-pass filter, and high-frequency noise is cut.

Next, the signal from the analog signal processing circuit 32 of the audio channel 1 is input to an 8-bit analog / digital conversion circuit (hereinafter abbreviated as A / D conversion circuit) 33, and the A / D conversion circuit At 33, it is converted into a digital signal 33e sampled at a sampling period S. On the other hand, a signal from the analog signal processing circuit 22 of the audio channel 2 is input to an 8-bit analog / digital conversion circuit (hereinafter abbreviated as an A / D conversion circuit) 23, and the A / D conversion circuit 23 performs a sampling cycle. It is converted into a digital signal 23e sampled by S. The digital signal 33e of the A / D conversion circuit 33 of the audio channel 1 is input to the audio analysis digital signal processing device 34, where it is subjected to PARCOR analysis conversion (Percoll analysis conversion). Voice analysis digital signal processor 34
3 obtained by PARCOR analysis conversion in
4e (pitch 34p, amplitude 34A, PARCOR coefficient 3
4k1 to 34ki) are input FIFO memory buffers 3
5 is temporarily stored. On the other hand, A / D of audio channel 2
The digital signal 23e of the conversion circuit 23 is input to the voice analysis digital signal processing device 24, and is subjected to PARCOR analysis conversion. Data 24e obtained by PARCOR analysis conversion in the voice analysis digital signal processor 24
(Pitch 24p, amplitude 24A, PARCOR coefficient 24k
1 to 24 ki) are temporarily stored in the input FIFO memory buffer 25. When the user sets on the control panel 36 to switch between Japanese broadcasting and English broadcasting without sound, the microcontroller 37 receiving the instruction sets the speech synthesizing digital signal processing device 39.
, An instruction for double-speed frame conversion is issued.

Further, the microcontroller 37 reads out the oldest data among the data stored in the input FIFO memory buffers 35 and 25. For example, when the signal level value of the FIFO memory buffer signal 35d becomes equal to or less than a predetermined threshold value, it is determined that there is no sound, the reading of the FIFO memory buffer signal 35d is stopped, and the other FIFO memory buffer signal 25d is stopped.
d is read and written to the output FIFO memory buffer 38. The microcontroller 37 continuously reads the FIFO memory buffer signal 25d of the input FIFO memory buffer 25 and outputs the data to the output FIFO memory buffer 38.

When the FIFO memory buffer signal 25d becomes data equal to or less than the threshold value, the microcontroller 37 determines that there is no voice, stops reading the FIFO memory buffer signal 25d, and stops the other FIFO memory buffer signal 25d.
O memory buffer signal 35d is read out and output FIFO
Write to the memory buffer 38. Here, the microcontroller 37 performs high-speed reading and discarding as long as the FIFO memory buffer signal 35d is equal to or smaller than the threshold value. As described above, the signals of Japanese broadcast and English broadcast output are written to the output FIFO memory buffer 38 while alternately switching channels with voiced and unvoiced as one pair.

The voice analysis digital signal processor 34 on the Japanese broadcast side, which is the voice channel 1, is a program in which a PARCOR (partial correlation) voice analysis algorithm is programmed. The result calculated at the frame period T is a pitch 34p and an amplitude 34A. And the PARCOR coefficients become 34 ki data from 34 k 1. Also, the audio analysis digital signal processing device 24 on the English broadcast side, which is the audio channel 2, has a PAR similar to the audio analysis digital signal processing device 34 on the Japanese broadcast side.
The COR (partial correlation) speech analysis algorithm is programmed, and the result of the calculation at the frame period T is data of pitch 24p, amplitude 24A, and PARCOR coefficients 24k1 to 24ki. The three parameter data in each of the audio analysis digital signal processors 34 and 24 are temporarily stored in an output FIFO memory buffer 38. The voice synthesizing digital signal processing device 39 includes a pitch 38p, an amplitude 38A, and a PAR of the output FIFO memory buffer 38.
Using COR coefficients 38k1 to 38ki, PARCO
Perform R speech synthesis. At this time, the microcontroller 3
Receiving the double-speed instruction signal 39c from the voice signal generator 7, the voice synthesizing digital signal processing device 39 performs voice synthesis by setting the frame period to T / 2, which is 倍 times that of the voice analysis.

The above-mentioned voice synthesizing digital signal processor 39
Will be described in more detail. The pitch 38p applied to the voice synthesis digital signal processor 39 is
The pitches are the same as the pitches 35p and 25p in the input FIFO memory buffers 35 and 25, and the amplitude 38A is
A, same as 25A, PARCOR coefficient 38k1-
38ki is a PARCOR coefficient of 35k1 to 35ki, 25
Same as k1 to 25ki. The difference is that the frame period 34c of each of the voice analysis digital signal processors 34 and 24 is different.
And 24c have a period T, whereas the frame period of the speech synthesis digital signal processing device 39 is T / 2. As described above, according to the fast hearing device of the second embodiment, the pitch of the input audio signal and the output audio signal does not change, and the frequency of the audio signal does not change. Furthermore, since the amplitude of these signals does not change, the loudness of the sound does not change. Also, PARCOR
Since the coefficients do not change, the sound quality does not change. However, since the frame period of the input voice signal and the output voice signal is halved, the sound generation period of the voice is not longer than that of the voice analysis digital signal processor 34,2.
4 compared to when entering the speech synthesis digital signal processor 3
When it comes out from 9, it is halved. Therefore, in the fast listening device of the second embodiment, the input audio signal is converted into digital data equivalent to twice as fast as the fast voice.

The output data 39e of the voice synthesizing digital signal processing device 39 is input to the digital signal processing circuit 26, where the continuity of the waveform is corrected, and the corrected data 26e is output to a digital / analog conversion device (hereinafter, referred to as a digital / analog conversion device). D
/ A converter). The output 27e of the D / A converter 27 is sent to the amplifier 28. The signal 28e output from the amplifying device 28 is sent to the audio output device 29, where an audible sound wave 29a is formed, and reaches the human hearing. The sound wave waveform output from the sound output device 29 is similar to the digital signal processing signal 26e, and the sound generation period is reduced to half. Also, the audible sound wave 29 having a waveform in which the portion of the audio signal 31e of the Japanese broadcast and the audio signal 21e of the English broadcast excluding the non-voice section is shortened to １ ／.
a. That is, as shown in FIG.
9a has English 1, Japanese 1, English 2, Japanese 2, English 3
・ English and Japanese languages are included alternately as Japanese 3. Thus, the user will hear every word entered twice as fast.

[0026]

According to the fast hearing device of the present invention, the speed is twice as high,
2.5 times speed, 3 times speed: Even when performing fast forward reproduction at N times speed, 1 /
Fast listening is possible in the time of 2, 2/5, 1/3, and 1 / N, so that the output voice can be easily recognized and its contents can be accurately understood. Further, if the fast hearing device of the present invention is used, simultaneously generated sounds of different languages are time-divisionally double-speed, 2.5-speed, 3-times-speed,.
, And it is possible to obtain a fast listening device capable of simultaneously listening to language broadcasts of many countries.

[Brief description of the drawings]

FIG. 1 is a block diagram illustrating a configuration of a fast hearing device according to a first embodiment of the present invention.

FIG. 2 is a block diagram showing a partial configuration of a fast hearing device according to a second embodiment of the present invention.

FIG. 3 is a block diagram showing a configuration of a part of a fast hearing device according to a second embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration of a fast hearing device of Conventional Example 1.

FIG. 5 is a block diagram showing the configuration of a prior art hearing aid device of a second conventional example, and a waveform diagram of the fast hearing device of the second conventional example.

FIG. 6 is a block diagram showing a configuration of a voice multiplexing apparatus of a third conventional example.

[Explanation of symbols]

 Reference Signs List 11 voice analysis digital signal processing device 12 output FIFO memory buffer 13 voice synthesis digital signal processing device 14 digital signal processing device 52 head 53 A / D conversion device 54 input FIFO memory buffer 55 microcontroller 56 input / output control device 57 D / A conversion Device 58 Amplifying device 59 Audio output device

Claims (3)

[Claims] 1. An analog signal processor for converting an input audio signal into an analog signal, an analog / digital converter connected to the analog signal processor, and a frame cycle connected to the analog / digital converter. A voice analysis digital signal processor for analyzing a voice signal at T; an input FIFO memory buffer in which data from the voice analysis digital signal processor is temporarily recorded; and a data recorded in the input FIFO memory buffer as a predetermined value. A microcontroller for comparing and selecting desired data, an output FIFO memory buffer for recording the data selected by the microcontroller, and a data stored in the output FIFO memory buffer for a frame period T / N (N is a positive value). Voice synthesis digit) Signal processing device, a digital signal processing device for correcting data from the voice synthesis digital signal processing device into continuous waveform data,
A fast hearing device comprising: an analog conversion device; an amplification device for amplifying an output signal of the digital / analog conversion device; and a sound output device for converting a signal from the amplification device to an audible sound wave. 2. A plurality of analog signal processors for converting a plurality of types of input audio signals into analog signals; a plurality of analog / digital converters connected to each of the plurality of analog signal processors; A plurality of audio analysis digital signal processors connected to each of the plurality of analog / digital converters for analyzing an audio signal at a frame period T; and temporarily storing data analyzed by the plurality of audio analysis digital signal processors. A plurality of input FIFO memory buffers to be set, a control panel for performing desired settings, and an analysis of each data temporarily recorded in the plurality of input FIFO memory buffers, and conversion of data according to a command of the control panel, Microcontroller that switches multiple converted audio signals in a time-division manner A roller, an output FIFO memory buffer for temporarily recording data formed in the microcontroller, and a voice for voice-synthesizing the data recorded in the output FIFO memory buffer at a frame period T / N (N is a positive discrete value). A synthetic digital signal processing device; a digital signal processing device for correcting data from the speech synthesis digital signal processing device into continuous waveform data; and a digital / digital converter connected to the digital signal processing device.
A fast hearing device comprising: an analog conversion device; an amplification device for amplifying an output signal of the digital / analog conversion device; and a sound output device for converting a signal from the amplification device to an audible sound wave. 3. The speech analysis digital signal processor performs PARCOR analysis at a frame cycle T, and the speech synthesis digital signal processor performs PARCOR speech synthesis at a frame cycle T / N.
Or the quick hearing device according to 2.