JP2000081899A - Speech recognizing device and speech recording and reproducing device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain good speech recognition performance with recorded speeches without deteriorating the sound quality of reproduced sounds by executing control in such a manner that a post filter means is not driven at the time of executing speech recognition processing and that the post filter means is driven at the time of reproduction. SOLUTION: A post filter control section 55 judges whether a reproducing button is clicked or not and if the section judges that the button is clicked, the section controls a changeover switch 56 and connects the output of a synthesizing filter 53 to the post filter 54. As a result, the post filter 54 is driven. The speeches are then reproduced by a speaker in accordance with the post filter output. When the speech recognition button is clicked, the post filter control section 55 controls the changeover switch 56 and connects the output of the synthesizing filter 53 as it is to a terminal. As a result, the post filter 54 is not driven. The speech recognition processing is thus executed in accordance with the signal not passing the post filter 54.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a voice recognition device and a voice recording / reproducing device, and more particularly to a voice recognition device and a voice recording / reproducing device capable of decoding voice data and shaping the noise.

[0002]

2. Description of the Related Art A so-called voice word processor or a so-called dictation system that automatically creates a document based on the voice data when inputting voice data by dictation and displays it on a screen or the like, This is one of the goals in the development of a conventional speech recognition system, and is currently being actively researched.

[0003] With the recent advances in speech recognition technology and computer technology, a microphone is connected to a personal computer, and voice input using the microphone is documented on the personal computer and displayed on a screen. It has been developed and is generally commercially available.

On the other hand, conventionally, when a document is created, the content of the document to be created is once dictated and recorded on a recording device such as a tape recorder, and a secretary or a typist or the like later reproduces the dictated content, and a typewriter. The use of a recording device such as a tape recorder or the like has become common as one of effective forms of using a recording device such as a tape recorder.

[0005] In such dictation recordings,
It is strongly desired to realize a technique for automatically converting a recorded content into a document.

As an example of a technique for realizing such a demand, for example, Japanese Patent Publication No. 63-33174 discloses that a recording device such as a tape recorder is connected to a voice recognition device, and a reproduction signal of the recording device is inputted. This document describes a technology related to an apparatus for creating sentences.

With the recent development of computer technology and digital signal processing technology, so-called digital recorders have been developed which convert recorded data into digital data and record it on a writable / erasable recording medium such as a flash memory. As a result, the digitized recorded contents can be transferred to a personal computer, and the recorded contents can be reproduced on the personal computer.

The present applicant has developed an audio data processing control device which enables the recorded data transferred from such a digital recorder to be handled by a simple operation on a personal computer. 9-149
No. 728 makes the proposal.

Further, the present applicant has developed a dictation system in which digitally recorded voice is passed from the voice data processing control device to a voice recognition device for voice recognition and displayed on a screen as a document. Hei 9-
No. 149729 makes such a proposal.

According to such a dictation system, a recording is temporarily made on a digital recorder, the recording data is transferred to a computer later, and the recording data transferred to the computer is subjected to voice recognition to create a document. It is not necessary to sit down in front of the computer and input voice directly, so that voice input can be performed in various environments, and playback operation of the recording device when performing voice recognition. There is a great advantage that a troublesome operation such as is unnecessary.

By the way, in the above-mentioned digital recorder, in order to secure a longer recordable time, the input audio signal is subjected to highly efficient encoding, so that the data recorded on the recording medium can be recorded. Attempts have been made to minimize the amount.

As means widely used for efficiently compressing such a speech signal, a speech signal is encoded using a linear prediction parameter representing a spectral envelope and an excitation parameter corresponding to the linear prediction residual signal. There is a method to do. Since the speech coding method using such a linear prediction method can obtain relatively high quality synthesized speech with a small transmission capacity, various application methods are actively studied in conjunction with recent advances in hardware technology. Is being developed.

Among them, as a method for obtaining good sound quality, for example, “Improved speech quality” by Kleijin et al.
and efficient vector quantization in SELP "(ICASS
P'88 s4.4, pp. 155-158, 1988), a CELP (Code Excited Linear Predictive C) using an adaptive codebook obtained by repeating past sound source signals.
The oding) scheme is well known.

In speech decoding for such speech encoding, it is known to use a post filter for the purpose of improving subjective sound quality. This post filter performs noise shaping by utilizing the auditory masking characteristic. Noise shaping refers to a function of transforming quantization noise, which originally has almost flat spectral characteristics, so as to be close to the spectral characteristics of the speech to be processed, and suppressing the perception of the quantization noise by masking. Such a post filter
Generally, it is arranged immediately after the audio decoding unit. The detailed configuration of the post filter is described, for example, in Japanese Patent Application Laid-Open No. 8-63196.

[0015]

The above-mentioned post-filter performs a spectral transformation on the quantization noise so as to be close to the spectral characteristic of the speech to be processed, and suppresses the perception of the quantization noise by masking. Therefore, it is intended to improve the subjective sound quality and does not faithfully reproduce the spectrum of the original sound. Therefore, even if the subjective sound quality is improved as a result of the post-filtering, the degree of waveform approximation between the spectrum of the original sound and the spectrum of the reproduced sound is rather small in many cases. As a result, there is a problem that the recognition performance is reduced when the reproduced sound is subjected to voice recognition.

SUMMARY OF THE INVENTION The present invention has been made in view of the above circumstances, and has been made in view of the above circumstances. It is an object of the present invention to provide a device and an audio recording / reproducing device.

[0017]

In order to achieve the above object, a voice recognition apparatus according to a first aspect of the present invention provides a voice recognition device for capturing voice data from a recording medium in which voice data is recorded in an encoded state. Data acquisition means, decoding means for decoding audio data acquired by the audio data acquisition means, post-filter means for performing noise shaping on the audio data decoded by the decoding means, Discriminating means for discriminating whether to perform speech recognition processing or reproduction, and when the discriminating means determines that speech recognition processing is to be performed, the post-filter means is controlled not to be driven, and discrimination is made by the discriminating means to reproduce. And a post-filter drive control means for controlling the post-filter means to drive the post-filter means.

The audio recording / reproducing apparatus according to the second aspect of the present invention comprises: an audio input means for converting an input audio into an electric signal to generate audio data; and an encoding means for encoding the audio data from the audio input means. A recording medium for recording audio data encoded by the encoding means, a decoding means for decoding the encoded audio data, and a noise shaping of the audio data decoded by the decoding means. Post-filtering means, output means for outputting audio data decoded by the decoding means, and output of audio data via the output means, and the audio data performs a voice recognition process. In other words, the post-filter means is controlled not to be driven, and otherwise, the post-filter means is controlled not to be driven. It is obtained by a preparative filter drive control means.

Further, the audio recording / reproducing apparatus according to the third invention is the audio recording / reproducing apparatus according to the second invention, wherein the output means is an earphone jack, and the audio data output via the earphone jack is audio data. The apparatus further comprises a detecting means for detecting whether or not the recognition processing is performed.

[0020]

Embodiments of the present invention will be described below with reference to the drawings. FIGS. 1 to 8 show a first embodiment of the present invention, and FIG. 1 is a diagram showing an overall configuration of a system including a voice recording / reproducing device and a voice recognition device.

As shown in FIG. 1, a system including the voice recording / reproducing device and the voice recognition device converts a voice into an electric signal, converts it into voice data, and reproduces the voice data as necessary. Recorder 1, miniature card 2 which is detachably attached to digital recorder 1 and is used as a recording medium for recording the audio data, and PC card slot 20 (see FIG. 3) for inserting miniature card 2 ), A PC card adapter 3 for enabling connection and a display 5 for displaying a speech recognition result, a keyboard 6 for inputting, a mouse 7 and the like. A sound to be processed by the control program 8 and the voice recognition program 9 on the voice data obtained from the miniature card 2. It is configured to include a personal computer 4 as the recognition device.

Next, FIG. 2 shows the digital recorder 1 described above.
FIG. 3 is a block diagram mainly showing an electrical configuration of FIG.

As shown in FIG. 2, the digital recorder 1 includes a microphone (hereinafter, abbreviated as a microphone) 21 as a sound input means for inputting a sound and converting it into an electric signal.
A preamplifier 22 for amplifying the audio signal from the microphone 21 to an appropriate level; and a coding means (CODEC) 25 for removing unnecessary high frequency components from the audio signal amplified by the preamplifier 22. And a low-pass filter (LPF) 26 for converting an analog audio signal output from the low-pass filter 26 into digital data.
An A / D converter 27 constituting the digital audio signal 25 and encoding means for encoding (compressing) the digitalized audio signal during a recording operation and decoding (expanding) the encoded data during a reproducing operation. Based on address information obtained from an encoding / decoding unit (DSP) 31 also serving as a decoding unit and a system control unit 34 serving as a post-filter drive control unit described later, audio is sent to a flash memory card 33 serving as a recording medium described later. A memory control unit 32 configured by, for example, a 16-bit CPU for controlling information recording / reproduction, a flash memory card 33 configured by the miniature card 2 or the like as a recording medium, and the encoding / decoding unit 31 Of the quantization noise in the audio signal output during playback operation so that it is close to the spectral characteristics of the audio to be processed A post filter unit serving postfilter 30 which deforms, constitute what is and the CODEC25 for converting the digital audio signal outputted from the post-filter 30 into an analog signal D /
An A converter 28, a low-pass filter 29 for removing unnecessary high-frequency components from the audio signal converted into an analog signal by the D / A converter 28, and constituting the CODEC 25; Amplifier 2 for amplifying the analog audio signal output from
3, a speaker 24 driven by the power amplifier 23 to emit sound, an operation input unit 36 including various operation buttons and the like, and the digital recorder 1
A display unit 35 such as an LCD for displaying various information related to the digital recorder 1, a power supply 37 such as a detachable battery for supplying power to each circuit of the digital recorder 1, a CODEC 25 described above, encoding / decoding, etc. Unit 31, memory control unit 32, flash memory card 33, display unit 3
5, a system control section 34 for controlling the whole of the digital recorder 1 including the power supply 37 and having an output of the operation input section 36 connected thereto. Have been.

FIG. 3 is a block diagram showing an electrical configuration of the personal computer 4.

The personal computer 4 performs sound reproduction, information display, and the like according to the control program 8.
The CPU 11 performs document creation and the like according to the voice recognition program 9 and performs various processes according to other various programs, a main memory 12 as a recording medium serving as a work area of the CPU 11, a hard disk, a floppy disk, and the like. An internal recording medium 13 as a recording medium in which the control program 8 and the voice recognition program 9 are recorded, an external port 14 for connecting to various external devices, and an interface (hereinafter, IF) for connecting the display 5 15), an IF 16 for connecting the keyboard 6 and the mouse 7, a speaker 18 for emitting sound based on audio data, an IF 17 for connecting the speaker 18, and the PC card slot 2.
0, and an IF 19 for connecting the CPU 11, the main memory 12, the internal recording medium 13, the external port 14, the IFs 15, 16, 17, and 19.
Are connected to each other via a bus.

The audio data may be read directly from the miniature card 2 attached to the PC card adapter 3 via the PC card slot 20.
The information may be temporarily recorded on the internal recording medium 13 and read out from the internal recording medium 13, or
You may make it read directly from the digital recorder 1 via communication means etc.

FIG. 4 is a diagram showing an example of a screen display when the control program is executed in the personal computer 4. As shown in FIG.

FIG. 4 shows a main screen 41.
A menu bar 42 for selecting an operation related to a file or an operation related to editing, and a tool button bar 4 for displaying various operations in an easily understandable manner using icons.
3 and a list of information such as the name of the audio file transferred from the miniature card 2, the length of the recording time, the recording date and time, and the priority. An audio file list box 44 for highlighting an audio file is displayed, and a reproduction control 45 for performing processing such as reproduction by a reproduction button 46 and stop, fast forward, and fast rewind by other buttons are displayed. The tool button bar 43 is provided with a voice recognition button 47 for starting voice recognition processing.

FIG. 5 is a block diagram showing a configuration of a voice decoding process (program) for voice data reproduction and voice recognition processing performed in the personal computer 4.

As shown in FIG. 5, the coded audio data input to the coded data input section 51 as the audio data capturing means is converted into a driving sound source signal generating section 52 constituting a decoding means and a decoding section. A synthesizing filter 53 constituting an optimizing means;
Each is input to a post filter 54 which is a post filter means.

The driving excitation signal generating section 52 generates and outputs a driving excitation signal from the input coded data.

The synthesis filter 53 generates a synthesized signal based on the coded data input from the coded data input section 51 and the driving excitation signal input from the driving excitation signal generation section 52. Output.

The output of the synthesizing filter 53 can be switched between output to the post filter 54 or output as it is by the changeover switch 56 as the post filter drive control means.

The operation of the changeover switch 56 is controlled by a post filter control unit 55 which is a post filter drive control unit and also serves as a discrimination unit. That is, the post filter control unit 55 The switching of the drive of the control unit 54 is controlled.

The post-filter 54 generates a post-filter output signal that has been subjected to noise shaping from the input synthesized signal, and outputs the post-filter output signal from an output terminal.

The operation of the post filter control section 55 will be described with reference to FIG. FIG. 6 is a flowchart showing the flow of the operation of the voice recognition process and the reproduction process performed in the personal computer 4 shown in FIG.

When the operation is started, it is first determined whether or not the play button 46 as shown in FIG. 4 has been clicked by the mouse 7 or the like (step S1). The post filter control unit 55 controls the changeover switch 56 to connect the output of the synthesis filter 53 to the post filter 54. Thus, the post filter 54 is driven (Step S6). Then, based on the output of the post filter, the speaker 1
8 is reproduced as audio (step S7).

If it is determined in step S1 that the play button 46 has not been clicked, it is determined whether the voice recognition button 47 has been clicked (step S2). If not clicked here,
The process returns to step S1 to wait for a button operation.

On the other hand, if the voice recognition button 47 is clicked in step S2, the post-filter control unit 55 controls the changeover switch 56 to connect the output of the synthesis filter 53 to the output terminal as it is.
As a result, the post filter 54 is not driven (Step S3).

In this way, the speech recognition processing is performed based on the signal that does not pass through the post filter 54 (step S).
4) The recognition result is displayed on the display 5 (step S5), and the process ends when the recognition of all the sentences is completed.

FIG. 7 is a block diagram showing another configuration example of the audio decoding process (program) for reproducing the audio data and performing the audio recognition process performed in the personal computer 4. In FIG. 7, the same reference numerals are given to portions that perform the same functions as the configuration shown in FIG. 5, and description thereof will be omitted.

FIG. 7 shows an audio reproducing decoding unit 63,
The speech recognition decoding unit 64 is provided independently.

Based on the control of the decoding control unit 61 which is the post-filter drive control means and also serves as the discrimination means, the connection destination of the changeover switch 62 which is the post-filter drive control means is changed to the audio reproduction decoding unit 63 Recognition decoding unit 64
Is switched.

The audio reproduction decoding section 63 includes an encoded data input section 51, a driving excitation signal generating section 52, a synthesis filter 53, and a post filter 54. The output is the driving sound source signal generation unit 5
2, the synthesis filter 53, and the post filter 54.

The speech recognition decoding section 64 includes an encoded data input section 51, a drive excitation signal generating section 52, and a synthesis filter 53. The output of the encoded data input section 51 is , The driving sound source signal generation unit 52 and the synthesis filter 53.

As described above, the independent provision of the audio reproducing decoding unit 63 and the audio recognizing decoding unit 64 enables the audio to be reproduced from the speaker even during the audio recognition processing. have.

FIG. 8 shows the operation of the decoding control unit 61.
This will be described with reference to FIG. FIG. 8 is a flowchart showing the flow of the operation of the voice recognition processing and the reproduction processing. This FIG.
In FIG. 7, the same reference numerals are given to the portions indicating the same processes as those in the configuration shown in FIG. 6, and description thereof will be omitted.

In step S1, the reproduction button 4
When the button 6 is clicked, the decoding control section 61 controls the changeover switch 62 to connect the output to the decoding section 63 for audio reproduction. Thus, the audio reproduction decoding unit 63 including the post filter 54 is driven (Step S).
12). Thereafter, based on the output of the audio reproduction decoding unit 63, the audio is reproduced by the speaker 18 in step S7.

When the speech recognition button 47 is clicked in step S2, the decoding control section 61 controls the changeover switch 62 to connect the output to the speech recognition decoding section 64. Thus, the speech recognition decoding unit 64 not including the post filter 54 is driven (step S11). The subsequent steps S4 and S5 are the same as described above.

According to the first embodiment, in the voice recognition device, whether or not to drive the post filter is appropriately switched between when voice data is recognized and when voice data is reproduced. In addition, it is possible to subjectively hear a sound of good sound quality during reproduction, and improve the performance of voice recognition during voice recognition.

FIGS. 9 to 12 show a second embodiment of the present invention. FIG. 9 is a diagram showing an overall configuration of a system including a voice recording / reproducing device and a voice recognition device, and FIG. 10 is a diagram showing one use mode of the recording / reproducing device, and FIG. 10 is a diagram showing another two use modes of the audio recording / reproducing device.

In the second embodiment, the first
The description of the same parts as those of the embodiment is omitted,
Only the differences will be mainly described.

In the first embodiment described above, the reproduction process or the voice recognition process is performed on the audio data transferred to the personal computer, whereas the second embodiment is different from the first embodiment. A voice recognition process is performed by inputting a reproduction signal from a voice recording / reproducing apparatus to a personal computer via a line, which is equipped with a sound board (not shown) and capable of executing a voice recognition processing program.

In the second embodiment, the output form of the reproduced signal of the audio recording / reproducing apparatus can take the following three forms depending on the use mode of the user.

First, a sound is emitted from a built-in speaker provided in the audio recording / reproducing apparatus.

That is, as shown in FIG. 10A, a digital recorder 71 as an audio recording / reproducing apparatus is provided with a built-in speaker 71a as an output means, and the recorded sound is reproduced by the built-in speaker 71a. Sounding.

Next, in a second mode, an earphone or a headphone or the like is connected to an earphone terminal (earphone jack) of the audio recording / reproducing apparatus, and sound is emitted by the earphone or the headphone.

That is, as shown in FIG. 10B, the digital recorder 71 has the built-in speaker 71.
Earphone jack 7 which is output means together with a
1b is provided, and the earphone jack 71b
Is connected to an earphone plug 76. The earphone plug 76 is connected to an earphone 74 via an earphone cord 75, and emits sound using the earphone 74. In this case, sound is not emitted by the built-in speaker 71a.

Third, the earphone terminal of the sound recording / reproducing device is connected to the microphone terminal (or line terminal) of the sound board of the personal computer on which the voice recognition program is installed by connecting the reproduction sound of the sound recording / reproducing device. Is input to the personal computer, and the personal computer executes voice recognition processing.

That is, as shown in FIG. 9, in this digital recorder 71, a plug 73 provided at one end of a connection cord 72 is connected to the earphone jack 71b, and the other end of the connection cord 72 is connected to a personal computer. By connecting to the personal computer 4, the reproduced sound is inputted to the personal computer 4. Then, the voice recognition processing is performed by executing the voice recognition program in the personal computer 4, and the result is displayed on the screen. Alternatively, sound reproduction processing is performed in the personal computer 4, and the sound reproduced by the speaker 18 is emitted.

In order to determine which of the above-described embodiments is employed, it is determined whether or not the earphone jack 71b is connected, and if so, the connected one is What is necessary is just to be able to detect what it is.

FIG. 11 is a block diagram mainly showing an electrical configuration of the digital recorder 71 which can determine such an aspect.

The output of the power amplifier 23 is output to the built-in speaker 71a via the switch 79 or to the earphone jack 71b via the detection circuit 80 and the attenuation resistor 78 as detection means. Has become.

The switch 79 is turned on / off in conjunction with the contact of the earphone jack 71b.

That is, when nothing is connected to the earphone jack 71b, since the switch 79 is on, the reproduced signal amplified by the power amplifier 23 is output to the speaker via the switch 79. Is done.

On the other hand, when a plug such as an earphone is connected to the earphone jack 71b, the switch 79 is turned off in conjunction with the contact point of the earphone jack 71b, whereby the conduction to the built-in speaker 71a is cut off.
At this time, the reproduced signal amplified by the power amplifier 23 is output from the earphone jack 71b.

The detection circuit 80 is a circuit for detecting the connection status to the earphone jack 71b, and the detection result is output to the system control unit 34. Then, a voltage drop or the like at both ends of the resistor is detected.

The impedance of an earphone or a headphone connected to the earphone jack 71b is usually about several Ω to several tens Ω. On the other hand, since the impedance of the sound board of the personal computer is usually 2 kΩ or more, it is easy to detect whether the earphone or the headphone or the personal computer is connected to the earphone jack 71b. Is possible.

The presence or absence of connection to the earphone jack 71b may be determined by checking for a short circuit in the earphone jack 71b.

Further, the digital recorder 71 is provided with a changeover switch 77 serving as a post-filter drive control means between the encoding / decoding section 31 and the post-filter 30. It is possible to switch whether or not to pass the decoded output through the post filter 30.

Thus, based on the detection result from the detection circuit 80, the system control section 34 as the post-filter drive control means controls the on / off of the switch 77.

FIG. 12 is a flowchart showing the operation of the digital recorder 71 at the time of decoding.

When the operation is started, first, it is determined whether or not the earphone jack 71b has been short-circuited by detecting whether or not the earphone jack 71b is short-circuited (step S21), and the connection to the earphone terminal is made. If yes, it is further determined whether or not the connected earphone (or headphone) is based on the output of the detection circuit 80 (step S22).

If it is determined that the earphone (or headphone) is connected, or if it is determined in step S21 that the earphone jack 71b is not connected, the system control unit 34 By controlling the changeover switch 77, the encoding /
The post filter 30 is driven by connecting the output of the decoding unit 31 to the post filter 30 (step S26).

After that, the output through the post filter 30 is emitted as sound by the built-in speaker 71a or earphone (or headphone) through amplification or the like by the power amplifier 23 (step S27).

If it is determined in step S22 that the earphone (or headphone) is not connected, the system control unit 34 controls the changeover switch 77 so that the encoding / decoding unit 31 Is bypassed so that the output of
The post filter 30 is not driven (step S
23).

This output is output from the earphone jack 71.
b, the data is input to the personal computer 4 connected by the plug 73 and the connection code 72, and the voice recognition program 9 of the personal computer 4 performs voice recognition processing (step S24). (Step S2
5).

According to the second embodiment, in the audio recording / reproducing apparatus, when the audio data is transferred to the personal computer side for voice recognition, and when the reproduced signal is listened to by the built-in speaker, earphone, or the like. By appropriately switching whether or not to drive the post filter, it is possible to subjectively hear a sound of good sound quality during reproduction, and to improve the performance of speech recognition during speech recognition. .

Note that the present invention is not limited to the above-described embodiment, and it is needless to say that various modifications and applications are possible without departing from the gist of the invention.

[0080]

As described above, according to the speech recognition apparatus of the present invention, the post-filter means is not driven when performing the speech recognition processing, and the post-filter means is driven when reproducing. , It is possible to obtain a good voice recognition performance without deteriorating the sound quality of the reproduced sound with respect to the recorded voice.

According to the audio recording / reproducing apparatus of the present invention, when outputting audio data for performing audio recognition processing, the post filter means is not driven, and otherwise, Is controlled so as to drive the post-filter means, so that it is possible to obtain good voice recognition performance without deteriorating the sound quality of the reproduced sound for the recorded voice.

Further, according to the third aspect of the present invention, the same effects as those of the second aspect can be obtained, and the voice data output via the earphone jack by the detecting means can be recognized by voice. It is possible to detect whether or not it is for processing.

[Brief description of the drawings]

FIG. 1 is a diagram showing an overall configuration of a system including a voice recording / reproducing device and a voice recognition device according to a first embodiment of the present invention.

FIG. 2 is a block diagram mainly showing an electrical configuration of the digital recorder of the first embodiment.

FIG. 3 is an exemplary block diagram showing an electrical configuration of the personal computer according to the first embodiment;

FIG. 4 is an exemplary view showing an example of a screen display when a control program is executed in the personal computer of the first embodiment.

FIG. 5 is a block diagram showing a configuration of a voice decoding process (program) for voice data reproduction and voice recognition processing performed in the personal computer of the first embodiment.

FIG. 6 is a flowchart showing a flow of an operation of voice recognition processing and reproduction processing performed in the personal computer having the configuration of FIG. 5;

FIG. 7 is an exemplary block diagram showing another configuration example of the audio decoding process (program) for audio data reproduction and audio recognition processing performed in the personal computer of the first embodiment.

FIG. 8 is a flowchart showing a flow of operations of voice recognition processing and reproduction processing performed in the personal computer having the configuration of FIG. 7;

FIG. 9 is a diagram illustrating an overall configuration of a system including a voice recording / reproducing device and a voice recognizing device according to a second embodiment of the present invention, illustrating one usage mode of the voice recording / reproducing device.

FIG. 10 is a diagram showing another two usage modes of the audio recording / reproducing device in the second embodiment.

FIG. 11 is a block diagram mainly showing an electrical configuration of the digital recorder of the second embodiment.

FIG. 12 is a flowchart showing the operation of the digital recorder of the second embodiment at the time of decoding.

[Explanation of symbols]

1, 71: Digital recorder (voice recording / reproducing device) 2: Miniature card 3: PC card adapter 4: Personal computer (voice recognition device) 11: CPU 20: PC card slot 21: Microphone (voice input means) 30, 54: Post Filter (post-filter means) 31 ... Encoding / decoding section (encoding means, decoding means) 33 ... Flash memory card (recording medium) 34 ... System control section (post-filter drive control means) 51 ... Encoded data input Unit (audio data capturing unit) 52 driving excitation signal generation unit (decoding unit) 53 synthesis filter (decoding unit) 55 post-filter control unit (determination unit, post-filter driving control unit) 56, 62, 77 ... Changeover switch (post-filter drive control means) 61 ... Decoding control section (determination means, PO Strike filter drive control means) 63: decoding section for voice reproduction 64: decoding section for voice recognition 71a: built-in speaker (output means) 71b: earphone jack (output means) 80: detection circuit (detection means)

Claims (3)

[Claims] An audio data fetching means for fetching audio data from a recording medium in which the audio data is recorded in an encoded state, and a decoding means for decoding the audio data fetched by the audio data taking means. Means, post-filter means for performing noise shaping of the audio data decoded by the decoding means, determination means for determining whether the audio data is subjected to voice recognition processing or reproduction, and voice recognition by the determination means. A post-filter drive control means for controlling not to drive the post-filter means when it is determined to perform the processing, and controlling to drive the post-filter means when it is determined to be reproduced by the determination means. A speech recognition device characterized by the following. 2. An audio input means for converting an input audio into an electric signal to generate audio data, an encoding means for encoding the audio data from the audio input means, and an audio encoded by the encoding means. A recording medium for recording data; decoding means for decoding the encoded audio data; post-filter means for noise shaping the audio data decoded by the decoding means; An output unit for outputting decoded audio data; and a post filter for outputting audio data via the output unit, and the audio data is for performing a voice recognition process. A post-filter drive control means for controlling not to drive the means, and otherwise controlling to drive the post-filter means. Audio recording and reproducing apparatus, characterized in that the. 3. The output means is an earphone jack, and further comprises a detection means for detecting whether or not the audio data output via the earphone jack is for performing a voice recognition process. The audio recording / reproducing apparatus according to claim 2, wherein: