JP2001142482A - Device for converting voice to caption - Google Patents

Abstract

PROBLEM TO BE SOLVED: To present a voice of a preliminarily determined unit at proper timing. SOLUTION: A voice recognition part 102 performs voice recognition processing on the basis of feature parameters and transmits the obtained recognition result to a recognition result correction part 104 and simultaneously transmits recognition result output timing. A voice presenting part 103 receives a voice divided to a proper unit from a sound analysis and division part 1 and receives recognition result output timing from the voice recognition part 102 and presents the voice of the pertinent unit synchronously with recognition result output timing.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an audio captioning apparatus for recognizing and processing audio to make captions in real time.

[0002]

2. Description of the Related Art Hitherto, when a speech such as a news speech is converted into subtitles in real time by using a speech recognition process, a result of the speech recognition is manually corrected. When correcting speech recognition results by hand, to assist the corrector,
Presenting the original voice provided for speech recognition to the corrector is performed. For example, as the voice presenting method, for example, the original voice provided for voice recognition is presented with a certain time delay. There are known ways to do this.

[0003]

However, according to this method, the time required for the speech recognition process varies,
The output timing of each recognition result is shifted, and as a result, the timing of the presentation of the recognition result to the corrector and the presentation of the voice become non-uniform and do not coincide with each other. That is, the voice is presented to the corrector at an appropriate timing. Could not do so, and it was virtually impossible to assist the corrector.

On the other hand, it is preferable for the viewer to display the subtitles with a minimum delay time from the image. Therefore, the correction of the recognition result and the presentation of the voice for the correction are performed in real time so as to follow the image. Had to be done in.

An object of the present invention is to solve the above-mentioned problems and to provide an audio captioning apparatus capable of presenting audio in predetermined units at an appropriate timing.

[0006]

According to a first aspect of the present invention, there is provided a speech recognition apparatus comprising: speech recognition means for recognizing speech for each predetermined unit; and correction means for correcting a speech recognition result by the speech recognition means. In a voice captioning device having a recognition / correction unit, the voice recognition / correction unit presents a voice for each predetermined unit in synchronization with a timing at which a voice recognition result is output by the voice recognition unit. Means are provided.

In the first aspect, the voice recognition / correction means may further include a delay means for delaying the voice presented by the voice presentation means by a predetermined delay time.

[0008] In the first aspect, the voice recognition / correction means may further include a voice speed changing means for changing a voice speed of the voice presented by the voice presenting means.

In the first aspect, the correction means may include a stop / restart means for stopping and restarting the presentation of the sound by the sound presentation means at an arbitrary timing.

[0010] In the first aspect, the modifying means designates a repetitive presentation part of the voice, and a first resumable capable of resuming the speech presentation from the first part of the repetition presentation part designated by the designation means. Means.

[0011] In the first aspect, the correction means includes a designation means for designating a repeated presentation part of the voice, and a speech presentation from a part which is a predetermined time before the first part of the repetition presentation part designated by the designation means. And a second resumable means that can be restarted.

[0012] In the first aspect, the modifying means designates a repetitive presentation part of the speech, and provides the speech presentation from a part after a predetermined time from the last part of the repetition presentation part designated by the designation means. And a third restart means that can be restarted.

[0013] In any one of the fourth to seventh aspects, the voice recognition / correction means increases the speech speed from the point in time when the presentation of the voice is resumed until the time when the voice is returned to the original voice presentation timing, and the voice is spoken quickly. The speed-up means can be further provided.

According to a ninth aspect of the present invention, there is provided a dividing means for dividing an input sound into predetermined units, a distributing means for distributing different predetermined units of sound obtained by the division by the dividing means, Speech recognition described in items 1 to 8
It is characterized by comprising at least two identical or different voice recognition / correction means of the correction means, and integration means for integrating correction results respectively obtained from the at least two voice recognition / correction means.

The control program recorded on the computer-readable recording medium stores in the computer a speech recognition procedure for recognizing speech for each predetermined unit and a modification procedure for modifying the speech recognition result by the speech recognition procedure. A control program for causing a computer to execute a voice presentation procedure for presenting a voice for each predetermined unit in synchronization with a timing at which a voice recognition result by the voice recognition procedure is output.

In the tenth aspect, the control program recorded on the computer-readable recording medium causes the computer to further execute a delay procedure for delaying the voice presented by the voice presentation procedure by a predetermined delay time.

In a tenth aspect, the control program recorded on the computer-readable recording medium causes the computer to further execute a speech speed variable procedure for changing the speech speed of the voice presented by the voice presentation procedure.

[0018]

Embodiments of the present invention will be described below in detail with reference to the drawings.

<First Embodiment> FIG. 1 shows a first embodiment of the present invention.
An embodiment will be described. In FIG. 1, reference numeral 1 denotes an acoustic analysis / separation unit, which calculates a characteristic parameter from an input voice and divides the input voice into appropriate units. Reference numeral 10 denotes a voice recognition / correction unit, which includes a voice recognition unit 102, a voice presentation unit 103, and a recognition result correction unit 10.
And 4. The voice recognition unit 102 is a sound analysis and classification unit 1
The voice recognition processing is performed based on the characteristic parameters from the above. The speech presentation unit 103 receives the speech divided into appropriate units by the acoustic analysis and division unit 1 and the recognition result output timing from the speech recognition unit 102, and presents the speech to the recognition result correction unit 104 at an appropriate timing. Is what you do. The recognition result correcting unit 104 compares the voice recognition result from the voice recognizing unit 102 with the voice presented by the voice presenting unit 103 and transmits the result of the manual correction (correction result) to the voice recognition result. Is what you do.

Next, the operation will be described. A characteristic parameter is calculated by the acoustic analysis / separation unit 1 from the input speech,
The obtained feature parameters are transmitted to the speech recognition unit 102, and the input speech is divided into appropriate units by the acoustic analysis / separation unit 1 and transmitted to the speech presentation unit 103.
The speech recognition unit 102 that has received the feature parameters from the acoustic analysis / classification unit 1 performs speech recognition processing based on the feature parameters, transmits the obtained recognition result to the recognition result correction unit 104, and simultaneously sets the recognition result output timing. Send. The voice presenting unit 103 receives the voice divided into appropriate units from the acoustic analysis / classification unit 1 and
02, the recognition result output timing is received, and the sound of the unit is presented in synchronization with the recognition result output timing.

The corrector inputs the voice presented by the voice presenting unit 103 and the recognition result correcting unit 10 from the voice recognizing unit 102.
4, and compares the recognition result with the recognition result transmitted to the recognition result correction unit 1.
By operating the button 04, the speech recognition result can be manually corrected. When the correction result is determined, the determined correction result is transmitted by the recognition result correction unit 104.

<Second Embodiment> FIG. 2 shows a second embodiment of the present invention.
An embodiment will be described. This embodiment differs from the first embodiment in the configuration of the speech recognition / correction unit.

That is, the voice recognition and speech recognition of the first embodiment
In the correction unit 10, the speech presentation unit 103 receives the speech divided into appropriate units from the acoustic analysis and classification unit 1, receives the recognition result output timing from the speech recognition unit 102, and synchronizes with the recognition result output timing. Then, the sound of the unit is presented.

On the other hand, the speech recognition and
In the correction unit 20, the sound of the unit presented by the speech presentation unit 103 in synchronization with the recognition result output timing is
The sound is delayed by a predetermined time by the audio delay unit 205 and presented.

<Third Embodiment> FIG. 3 shows a third embodiment of the present invention.
An embodiment will be described. This embodiment differs from the first embodiment in the configuration of the speech recognition / correction unit.

That is, the speech recognition and the speech recognition of the first embodiment
In the correcting unit 10, the voice presenting unit 103 receives the unit voice from the acoustic analysis / classification unit 1, receives the recognition result output timing from the voice recognition unit 102, and synchronizes the unit with the recognition result output timing. Minute voice.

On the other hand, the voice recognition and
In the correcting unit 30, the speech presented by the speech presenting unit 103 in synchronization with the recognition result output timing is converted into the speech speed variable unit 3
05, the speech speed is changed to a predetermined speech speed.

<Fourth Embodiment> FIG. 4 shows a fourth embodiment of the present invention.
An embodiment will be described. This embodiment differs from the first embodiment in the configuration of the speech recognition / correction unit.

That is, the speech recognition and the speech recognition of the first embodiment
In the correcting unit 10, the voice presenting unit 103 receives the unit voice from the acoustic analysis / classification unit 1, receives the recognition result output timing from the voice recognition unit 102, and synchronizes with the recognition result output timing. The sound of the unit was presented.

On the other hand, the speech recognition and
In the correction unit 40, the recognition result correction unit 404 enables a stop request and a restart request for voice presentation to be made at any timing. At the timing, when the stop request and the restart request of the voice presentation are received, the stop and the restart of the voice presentation to the recognition result correcting unit 404 are performed according to each request.

<Fifth Embodiment> FIG. 5 shows a fifth embodiment of the present invention.
An embodiment will be described. This embodiment differs from the fourth embodiment in the configuration of the speech recognition / correction unit.

That is, the speech recognition and the speech recognition of the fourth embodiment
In the correction unit 10, the recognition result correction unit 404 enables a stop request and a restart request for voice presentation to be made at an arbitrary timing. At the timing, when the stop request and the restart request of the voice presentation are received, the stop and the restart of the voice presentation to the recognition result correcting unit 404 are performed according to each request.

On the other hand, the voice recognition and
The correction unit 50 enables (1) the speech presentation unit 503 to restart the speech presentation from the first part of the repeatedly presented portion designated by operating the recognition result modification unit 504, and (2) the recognition result modification. By operating the unit 504, the voice presentation can be resumed by the voice presentation unit 503 from a portion that is a predetermined time before the first portion of the designated repeated presentation portion, and (3) the recognition result correction unit 504 is operated. The voice presentation unit 503 can restart the voice presentation from a part that is a predetermined time after the last part of the designated repeated presentation part.

<Sixth Embodiment> FIG. 6 shows a sixth embodiment of the present invention.
An embodiment will be described. This embodiment differs from the fifth embodiment in the configuration of the speech recognition / correction unit.

That is, the speech recognition and the speech recognition of the fifth embodiment
The correcting unit 50 operates the speech presenting unit 5 from the first part of the repetitively presented part specified by operating the recognition result correcting unit 504.
03, the speech presentation can be resumed, and the speech presentation unit 503 resumes the speech presentation from a part that is a predetermined time before the first part of the designated repeated presentation part by operating the recognition result correction unit 504. In addition, the voice presentation unit 503 can restart the voice presentation from a part that is a predetermined time after the last part of the repeatedly presented part specified by operating the recognition result correction unit 504 by operating the recognition result correction unit 504. did.

On the other hand, the voice recognition and
In the correcting unit 60, the speech presentation restarting part is designated by the recognition result modifying unit 604 essentially in the same manner as in the fifth embodiment, and when the speech presenting unit 503 performs the speech presentation, the speech presenting unit 503 Variable speed section 3 that receives voice from
According to 05, the speech speed of the received voice is increased, that is, the voice is presented in a fast voice.

<Seventh Embodiment> A speech captioning apparatus according to the present embodiment is provided with the speech recognition / correction units 10, 20, 30, 40, 50, and 60 according to the first to sixth embodiments. FIG. 7 shows a voice recognition / correction unit 10 according to the first embodiment, and a voice recognition / correction unit 10 according to the second embodiment. Unit 20 and a speech recognition / correction unit 60 according to the sixth embodiment.
Is shown. These at least two speech recognition / correction units may have the same configuration or different configurations.

The acoustic analysis / speech classification / distribution unit 701 calculates feature parameters from the input speech, transmits the obtained feature parameters to the speech recognition unit of each speech recognition / correction unit, and converts the input speech into an appropriate speech. It is divided into units and distributed to the voice presentation unit of each voice recognition / correction unit. Then, the correction result integration unit 703 integrates the correction results obtained by the correction result correction units of the voice recognition / correction units and outputs subtitles.

As described above, the storage medium storing the program codes of the software for realizing the functions of the above-described embodiments is supplied to the system or the apparatus, and the computer or the CPU (central processing unit) of the system or the apparatus is supplied.
ral processing unit) or MPU (microprocessor uni
Needless to say, the object of the present invention can be achieved also by reading and executing the program code stored in the storage medium in t).

In this case, the program code itself read from the storage medium realizes the novel function of the present invention, and the storage medium storing the program code constitutes the present invention.

As a storage medium for supplying the program code, for example, a floppy disk, hard disk, magnetic tape, optical disk, magneto-optical disk, CD
-ROM (compact disk ROM), CD-R (compact disk
recordable), non-volatile memory card, ROM (read
only memory), and so on.

When the computer executes the readout program code, not only the functions of the above-described embodiment are realized, but also the OS (opera- tion) running on the computer based on the instruction of the program code.
Needless to say, a case in which the functions of the above-described embodiments are realized by performing part or all of the actual processing by a ting system or the like is also included.

Further, after the program code read from the storage medium is written into a memory provided in a function expansion board inserted into the computer or a function expansion unit connected to the computer, based on the instruction of the program code, It goes without saying that the CPU provided in the function expansion board or the function expansion unit performs part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

[0044]

As described above, according to the present invention,
With the configuration as described above, it is possible to present the sound for each predetermined unit at an appropriate timing, and thus to assist the correction.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is a block diagram showing a second embodiment of the present invention.

FIG. 3 is a block diagram showing a third embodiment of the present invention.

FIG. 4 is a block diagram showing a fourth embodiment of the present invention.

FIG. 5 is a block diagram showing a fifth embodiment of the present invention.

FIG. 6 is a block diagram showing a sixth embodiment of the present invention.

FIG. 7 is a block diagram showing a seventh embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Acoustic analysis / speech classification part 10, 20, 30, 40, 50, 60 Speech recognition / correction part 102 Speech recognition part 103, 403, 503 Speech presentation part 104, 404, 504, 604 Recognition result correction part 205 Speech delay part 305 Variable speech rate unit 701 Sound analysis / speech classification distribution unit 703 Correction result integration unit

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Toru Imai 1-10-11 Kinuta, Setagaya-ku, Tokyo Japan Broadcasting Corporation Japan Broadcasting Research Institute (72) Inventor Toru Toki 1-1-10 Kinuta, Setagaya-ku, Tokyo 11 Japan Broadcasting Corporation Broadcasting Research Institute (72) Inventor Akio Ando 1-10-11 Kinuta, Setagaya-ku, Tokyo Japan Broadcasting Research Institute F-term (reference) 5D015 KK02 LL04 LL05 9A001 HH17 HH33 KK60

Claims (12)

[Claims] 1. An audio captioning apparatus comprising: a voice recognition unit for recognizing voice for each predetermined unit; and a correction unit for correcting a voice recognition result by the voice recognition unit. Wherein the voice recognition / correction means comprises voice presentation means for presenting voice of each predetermined unit in synchronization with a timing at which a voice recognition result is output by the voice recognition means. Device. 2. The captioning system according to claim 1, wherein the voice recognition / correction unit further includes a delay unit that delays the voice presented by the voice presentation unit by a predetermined delay time. apparatus. 3. The audio captioning apparatus according to claim 1, wherein the voice recognition / correction means further comprises a voice speed variable means for changing a voice speed of the voice presented by the voice presenting means. . 4. The audio captioning device according to claim 1, wherein the correction means includes a stop / restart means for stopping and restarting the presentation of the sound by the sound presenting means at an arbitrary timing. . 5. The method according to claim 1, wherein the modifying means comprises: designating means for designating a repeated presentation part of the voice; and the speech presentation can be resumed from the first part of the repeated presentation part designated by the designating means. An audio captioning apparatus comprising: a first resuming unit. 6. The method according to claim 1, wherein said correcting means comprises: designating means for designating a repetitive presentation part of the voice; and a predetermined time before the first part of the repetition presentation part designated by the designation means. An audio captioning device, comprising: second resuming means capable of resuming audio presentation. 7. The method according to claim 1, wherein the correcting means comprises: a designating means for designating a repeated presentation part of the voice; and a part after a predetermined time from the last part of the repeated presentation part designated by the designation means. An audio captioning device comprising: a third resuming unit capable of resuming audio presentation. 8. The method according to claim 4, wherein
The voice recognition / correction unit further includes a voice speed increasing unit that speeds up the voice speed and changes the voice to a quick voice during a period from when the voice presentation is resumed to when the voice presentation timing is restored. An audio captioning apparatus characterized by the above-mentioned. 9. A dividing means for dividing the input sound into predetermined units, and a distributing means for distributing sounds of different predetermined units obtained by the division by the dividing means. And at least two of the same or different voice recognition / correction means, and integration means for integrating correction results obtained from the at least two voice recognition / correction means. Characterized audio captioning device. 10. A control program for causing a computer to execute a voice recognition procedure for recognizing voice of each predetermined unit and a correction procedure for correcting a voice recognition result by the voice recognition procedure. A computer-readable recording medium in which a control program for causing a computer to execute a voice presentation procedure for presenting a voice for each predetermined unit in synchronization with a timing at which a voice recognition result is output by the voice recognition procedure is recorded. 11. The computer-readable recording medium according to claim 10, further comprising a control program for causing a computer to further execute a delay procedure for delaying a sound presented by the sound presentation procedure by a predetermined delay time. 12. The computer-readable recording medium according to claim 10, further comprising a control program for causing a computer to further execute a speech speed changing procedure for changing a speech speed of the voice presented by the voice presenting procedure.