JPS59201141A - Input device of sound information - Google Patents

Abstract

PURPOSE:To improve the sound recognizing factor and the reliability for an input device of sound information by generating the guidance in sound, when the words are registered in sound and inputting the sound data while hearing the answer-back sound when a sound information data is inputted. CONSTITUTION:A control circuit 3 controls a sound recognition control circuit 13 of a sound recognizing device 1 to fetch the result of sound recognition or controls the sound output control circuit 21 of a sound output device 2 to output the guidance and answer-back sound through a speaker 7. For registration of the sound, a speaker utters following the guidance sound outputted through the speaker 7 with use of a mike 6. Then the voice of the speaker is stored in a registered sound memory 14 via the circuit 13. For inputting the sound information data, the speaker inputs the voice in reminding own voice of a register mode while hearing the answer-back sound. In such a way of sound registration and input, the sound recognizing factor is improved with the improvement of reliability and operability.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention relates to a voice information input device for inputting information by voice by a human being.

[Background of the invention]

In conventional voice information input devices, when registering a registered word to confirm the voice, a system is adopted in which the word is uttered automatically while looking at characters and symbols displayed on a screen such as a CRT. For this reason, when inputting voice information, the user forgets to pronounce the registered word at the time of registration, resulting in a poor recognition rate and a drawback that the system is unusable. In addition, since the registered words were registered while looking at a screen such as a CRT, it was not possible to re-register the voice unless the speaker had to go to the trouble of entering the voice information to the location where the CB or T screen was located. It was inconvenient.

[Purpose of the invention]

An object of the present invention is to provide an easy-to-use voice information input device with a high recognition rate.

[Summary of the invention]

In the present invention, when registering five registered words by voice, a guidance sound is generated by voice, and when voice information data is input, an answerback sound is generated.

[Embodiments of the invention]

FIG. 1 is a system configuration diagram showing the configuration of a voice information input device according to an embodiment of the present invention.

The voice information recognition device 1 includes an amplifier 11 that amplifies the voice signal of the voice input micro, an A/D converter 12 that converts the voice signal from analog to digital, a registered voice memory 14 that stores registered voice in advance, and input voice. The voice recognition control circuit 13 performs voice recognition by comparing the registered voice and the registered voice.

The voice output device 2 includes a synthesized voice memory 22 that stores voices for voice output, and a voice output control circuit 21 that selects and outputs the stored contents of the synthesized voice memory 22 according to voice registration and voice recognition results. - D/A converter 23 that converts the output signal from digital to analog. An amplifier 24 that amplifies the analog signal and generates guidance and answerback audio from the speaker (!, or earphone) 7
It is made up of. Control circuit 3 is voice recognition device 1
control the voice recognition control circuit 13 of the voice recognition control circuit 13 to take in voice recognition results, control the voice output control circuit 21 of the voice output device 2 to output guidance and answerback sounds from the speaker 7, and display the control status on the display 5. This is a control computer that displays information such as information and voice recognition results. The control circuit 3 is controlled not only by voice but also by a keyboard 4.

FIG. 2 is a list showing an example of audio words used in an embodiment of the present invention. The voice words N011 to N42 in FIG. 2, that is, the range of α, are registered voices for voice recognition, and are registered in advance as registered words in the registered voice memory 14. To register the voice, the speaker uses a micro to register voice words N011~
42 by imitating the sequence using the guidance sound from the speaker 7, the amplifier 11, A/D
It is stored in the registered voice memory 14 via the converter 12 and the voice recognition control circuit 13. The range α in FIG. 2 is also a guidance sound when registering a registered word by voice and a voice output word for answering when voice recognition is performed.

After the speaker completes voice registration, when the speaker inputs voice word No. 1 to 42 using the micro, the voice recognition control circuit 13 registers the voice word in the registered voice memory 14.
A voice word with the same No. is searched out from the voice word memory of , and the NO. is outputted to the control circuit 3. Control circuit 3
In addition to the input data of the audio word No., which is displayed on the display 5, the audio output control circuit 21
issue commands to.

The voice word No. 1 to 55 in FIG.
Record it in advance in step 2. The recording method of audio words in the synthesized speech memory 22 is as follows: Audio words No. 1 to 55 are recorded as they are. Audio words No. 1 to 55 are recorded as they are.
In this method, each voice word is divided into several words, and the voice output control circuit 21 arranges them in the correct order and outputs them.Furthermore, only the phoneme pieces are recorded in the synthesized voice memory 22, and the voice output control circuit 21 outputs them. A method can be adopted in which the spoken words N091 to N095 are synthesized and output as a spoken word,
Either method is fine. When the speaker inputs speech other than the registered word, the speech recognition control circuit 13 determines that the speech is unrecognizable and outputs the unrecognizability level to the control circuit 3. The control circuit 3 takes in the unrecognizable level, displays the fact on the display 5, and issues a command to the audio output control circuit 21. The speech output control circuit 21 receives the command from the control circuit 3, reads out the phoneme pieces in the synthesized speech memory 22 in order, outputs speech words N0051 to N0055, and outputs the speech words N0051 to N0055 through the D/A converter 23 and amplifier 24. synthesized speech is generated from the speaker 7. That is, when a registered word of a speaker is registered as a voice, a guidance sound of the registered word is outputted from the speaker 7, and when voice data for inputting voice information is input, a recognition result is answered back from the speaker 7.

FIGS. 3 and 4 are flowcharts of voice registration showing an embodiment of the present invention.

In FIG. 3, a guidance sound "Zero" is generated from the speaker 7 of the audio output device 2 at 201 (Statement No. 201). Then, in step 202, the speaker pronounces "zero" in return according to the voice guidance, and the registered word "zero" is sent from the micro to the registered speech memory 14 of the speech recognition device 1.
0" is memorized.

Subsequently, at 211, a guidance sound is generated from the speaker 7, saying "1 1". At 212, the speaker utters "ichi" to store the registered word "1" in the registered speech memory 14 of the speech recognition device 1. Thereafter, in the same manner, following the voice guidance from the speaker 7, the speaker repeats the voice input and sequentially inputs the registered words from the micro.
The voice guidance in 291 is “Kyuu”, while in 292 it is “Kyuu”.
``Kyu'' and photons to register the registered numeric words ``0-9''.

FIG. 4 shows an example in which the recognition rate is improved when voice information data is input by having a speaker register the same registered word multiple times and registering the voice. Up to 201 and 202, the third
As in the figure, after "zero" has been registered as a voice once, the speaker 7 again gives voice guidance as "zero" in the statement member 201A.

The speaker again utters "zero" at 202A and repeats the voice registration. Next, at 211 and 212, the first
Register the voice of "Ichi" for the second time, and the statement No.
211A and 212A register the second "Ichi" voice. Hereinafter, in the same manner, 291.292,291
A. At 292A, in response to the voice guidance "Kyu", "
Say ``Kyuu'' and repeat it twice to register ``9'' as a voice. In this way, the voice registration of the registered word is completed while repeating the numbers "0 to 9" a plurality of times. In this case, it is preferable that the voice data stored in the registered voice memory 14 as the actual registered voice words be the voice data uttered at the end of each registered word. By repeating the same word, the speaker's tone quality, accent, and method of enunciation become closer to the guidance sound. As a result, the answerback sound during voice registration and voice input is the same synthesized voice that is completely synthesized from the same phoneme.
When inputting voice information data, the user inputs voice while listening to the answerback sound and remembering his/her own voice at the time of registration, which greatly improves the recognition rate.

In this way, registration words for speech recognition can be easily registered by simply repeating voice input in accordance with the voice guidance from the speaker, and it is possible to register speech with a high speech recognition rate.

FIG. 5 is a system configuration diagram of a voice information input device according to another embodiment of the present invention. A speaker carries a portable radio device 8 consisting of a transmitter 81 and a receiver 82, inputs voice from a micro, and inputs the voice information to the receiver 91 of the wireless fixed station 9 via radio waves, and the voice is transmitted. The receiver 82 receives radio waves from the transmitter 92 of the wireless fixed station 9 for guidance and voice answerback.
and listen to it through the speaker (or earphones) 7.

Since the speaker only needs to carry a small portable wireless device, the speaker can move freely while registering registered words and inputting voice information using both hands. [Effects of the Invention] According to the present invention , the recognition rate is significantly improved, and a highly reliable and easy-to-use voice information input device can be obtained.

[Brief explanation of the drawing]

Fig. 1 is a system configuration diagram of an embodiment of the present invention, Fig. 2 is a list diagram showing an example of voice words used in the present invention, and Figs. 3 and 4 are voice registration diagrams of an embodiment of the present invention. FIG. 5 is a system configuration diagram of another embodiment of the present invention. 21...Speech output control circuit, 22...Synthesized speech memory, broken 3.

Claims (1)

[Scope of Claims] 1. Voice information input characterized by a configuration in which a guidance voice of a registered word is emitted during voice registration, and an answerback voice for confirming voice input data is emitted when voice information is input. Device. 2. The voice information input device according to claim 1, characterized in that a memory for storing the guidance voice and the answerback voice is shared.