JP2002304192A - Voice recognition device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce the time required for recognition by reducing the time required to read voice data from a disk. SOLUTION: When discrimination is made to determine that a PTT switch has been operated (a step 101), a beep sound is generated (a step 102) and a command to start rotating the disk is outputted to a disk driver (a step 103). Then, audio signals of voice uttered by a user are taken in from a microphone to conduct voice recognition process (a step 104). When a recognition result is obtained, the driver is controlled, talk-back data, corresponding to the recognition result are obtained from the disk and talk-back voice is outputted from a loudspeaker (a step 105). Since the rotation of the disk is started prior to the voice recognition process, the time required to read the talk-back data is reduced, even though the disk is stopped.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech recognition apparatus for recognizing speech uttered by a user and producing talkback speech.

[0002]

2. Description of the Related Art Various types of voice recognition apparatuses of this type have been proposed, such as voice recognition used in car navigation systems.

In such a voice recognition device, voice data for notifying a user of a recognition result is stored in a large-capacity disk such as a DVD. Data stored on the disk is read using a disk driver.
The disk driver reads the data by rotating the disk at a high speed and moving the head perpendicular to the rotation direction. Further, in order to improve the durability and reduce the power consumption, the disk driver stops the rotation of the disk when there is no access to the disk for a certain time or more.

If there is a request to read audio data while the disk is stopped, it takes about 1-2 seconds for the rotation of the disk to stabilize and for the read head to be in focus. For this reason, the recognition time in the voice recognition device becomes long. In other words, if the recognition time is defined as the time from the completion of the utterance of the user to the time when the recognition result is obtained and the talk-back sound is generated, the time until the recognition result is obtained and the talk-back sound is generated becomes longer. Time will be long.

The present invention has been made in view of the above problems, and has as its object to reduce the time required for reading audio data from a disk to shorten the recognition time.

[0006]

In order to achieve the above object, according to the first aspect of the present invention, a voice signal of a voice uttered by a user after a user operates a switch for starting voice recognition is transmitted from a microphone. In the voice recognition device that captures and performs voice recognition processing, obtains voice data according to the recognition result from the disk, and generates talkback voice from a speaker, after the switch is operated,
At a predetermined timing before the start of the voice recognition process, there is provided a means for performing a process for starting the rotation of the disk.

According to the present invention, since the rotation of the disk is started first, and then the audio data corresponding to the recognition result is obtained from the disk, the time required for reading the audio data from the disk is obtained. And the recognition time can be shortened.

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below with reference to an embodiment shown in the drawings. FIG. 1 shows the overall configuration of the speech recognition system. The voice recognition system is PTT (pusu to talk)
It comprises a switch 1, a microphone 2, a speaker 3, a voice recognition device 4, and a disk driver 5. The disk driver 5 includes a disk 6 in which data such as talkback data is stored.

The PTT switch 1 is operated by the user when starting speech recognition. The voice recognition device 4
A voice signal of a voice uttered by the user after the PTT switch 1 is operated is taken in from the microphone 2, voice recognition processing is performed, and the disk driver 5 is controlled to generate voice data (talkback data) corresponding to the recognition result. 6, and the talkback sound is uttered from the speaker 3. The voice recognition device 4 can be realized not only in a single device but also in a control device of a navigation device for a vehicle, for example. In this case, the control device in the navigation device is configured using a computer, and has various functions for executing functions as navigation and also has a function for performing voice recognition.

Hereinafter, a specific operation of the speech recognition apparatus 4 in this embodiment will be described with reference to the processing shown in FIG.

The speech recognition device 4 repeatedly executes the processing shown in FIG. First, it is determined whether or not the PTT switch 1 has been operated (step 101). When it is determined that the PTT switch 1 has been operated, a beep sound is generated from the speaker 3 to prompt the user to utter (Step 102). Then, a command to start the rotation of the disk 6 is output to the disk driver 5 (step 10).
3).

Next, a voice signal of a voice uttered by the user is taken in from the microphone 2 and voice recognition processing is performed (step 104). That is, based on the voice signal taken in from the microphone 2, analysis and dictionary illumination are performed to obtain a recognition result. At this time, the CP of the speech recognition device 4 is compared with the recognition processing time.
If the processing capability of U is not sufficient, the voice signal is buffered, and the voice recognition process is continuously performed based on the buffered voice signal even after the utterance of the user is completed.

When the recognition result is obtained, the disk driver 5 is controlled to obtain talkback data corresponding to the recognition result from the disk 6, and the talkback sound is uttered from the speaker 3 (step 105). in this case,
In step 103, since the rotation of the disk 6 is started first, the rotation of the disk 6 is stable before the talkback data is read even if the disk 6 has been stopped. Data read time can be shortened.

In the above embodiment, the disk 6
The process for starting the rotation of the disk 6 is described as starting the rotation regardless of whether the disk 6 is rotating or stopped. However, it is determined whether or not the disk 6 is rotating, and the process is stopped. The rotation may be started only when necessary.

Further, in the above-described embodiment, a case is described in which it is determined that the PTT switch 1 is operated, and after the beep is generated, the rotation of the disk 6 is started. Other timings may be used as long as the timing is after the switch 1 is operated and before the voice recognition processing is started. However, assuming that the voice to be recognized is very short, immediately after the PTT switch 1 is operated (before or after the generation of the beep).
Is preferred.

[Brief description of the drawings]

FIG. 1 is a diagram showing an overall configuration of a speech recognition system according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a process of the speech recognition device 4 in FIG.

[Explanation of symbols]

1 ... PTT switch, 2 ... microphone, 3 ... speaker, 4 ...
Voice recognition device, 5: disk driver, 6: disk.

──────────────────────────────────────────────────続 き Continued on the front page (51) Int.Cl. ⁷ Identification symbol FI Theme coat ゛ (Reference) G11B 19/20 G10L 3/00 R F-term (Reference) 5D015 KK02 KK04 LL10 LL12 5D045 AB17 5D109 EA20

Claims (1)

[Claims] 1. A user operates a switch for starting voice recognition and then takes in a voice signal of a voice uttered by the user from a microphone, performs voice recognition processing, and obtains voice data according to a recognition result from a disk. Then, in a voice recognition device that generates talkback voice from a speaker, a process for starting rotation of the disc at a predetermined timing after the switch is operated and before the voice recognition process is started. A speech recognition device, comprising: