JPS59170936A - Input device of voice - Google Patents

Abstract

PURPOSE:To enable a voice input device to be easily used by a user and to shorten the time required for input work by detecting the voice made by the user before a sign promoting reinput to the user. CONSTITUTION:A voice detecting part 3 discriminates the existence of a voice (B) from a voice input terminal 10. When a signal from the voice detecting part 3 indicates the existence (C) of the voice at the time immediately before the generation of an input promoting signal (A) specified by a control part 5, i.e. the generation of the sound of ''pi'', a reinput command part 2 commands the regeneration of the input promoting signal (A) to an input promoting signal generating part 1. A voice recognition part 4 is commanded to start recognition after the generation of the input promoting signal. The recognized result is outputted to an output terminal.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to improvement of a voice input device. A voice input device is a very easy-to-use input means for a single person, and has already been put into practical use. However, when inputting voice, the timing of two utterances becomes a problem. That is, the user must start speaking after the two voice input devices are in a state for accepting input. In general, to notify the user that the device is ready to accept input, a voice response, a beep, or an indication on the display is used. In some cases, such as a skilled operator, it is possible to vocalize in correct synchronization with these prompting cues.

However, in general, the act of making a sound is an unexpected move for the person 44, so they cannot wait for the first signal (
I'm afraid to say this out loud. In such a case, information about the part uttered before input reception is lost, resulting in erroneous recognition or input not being accepted.

An object of the present invention is to realize a voice input means that is easy for the user to use by detecting when the user speaks without waiting for a signal and prompting the user to use the voice again. .

According to the present invention, there are two input prompting signal generation means for generating a signal urging the user to use voice input, a voice detection means for determining whether or not a voice signal is being input, and a time preceding the end of the manual input prompting signal. means for instructing the input prompting signal generating means to re-generate the input prompting signal when a voice input is detected according to the detection signal output by the voice detecting means;

A voice input device including recognition means for recognizing input voice is obtained.

According to this configuration, if the second user cannot wait for the signal and speaks, the user is immediately prompted to re-enter the signal.

You can avoid situations such as erroneous recognition or input not being accepted.
The time required for input work is also shortened.

Furthermore, since the inconvenience caused when uttering a signal without waiting for one signal is alleviated, the user does not have to worry so much about synchronizing with the signal, which improves ease of use.

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

FIG. 1 is a block diagram showing an embodiment of the apparatus according to the invention.

Reference numeral 1 denotes an input promotion signal generating section, which generates an input promotion signal to the output terminal 11 in response to a command signal from the control section 5. The input prompting signal is a "bee" sound.

Voice response, display display, etc. are possible.

In the following explanation, the case where the input prompting signal is a "beep" sound will be considered. Reference number 3 is a voice detection section.

Any device can be used as the voice detection section 2 as long as it can determine whether the voice from the voice input terminal 10 is M-less. Reference numeral 2 is a re-input command section, and the signal from the voice detection section 3 indicates the presence of voice at a time immediately before the input prompting signal specified by the control section 5 is generated, that is, immediately before the "beep" sound is generated.

The input prompting signal generator 1 is instructed to regenerate the input prompting signal. An example of the temporal flow of this operation is shown in FIG.

In the figure, the horizontal axis represents the time axis, (At, @.

(Q represents an input prompting signal, a voice input signal, and a voice detection signal, respectively. The dotted arrow indicates that the input prompting signal is commanded to be regenerated because a voice was detected immediately before the input prompting signal was generated.Reference number 4 is the voice recognition section, and the control section 5
As a result, the start of 1g recognition is commanded after the input prompting signal is generated.

At this time, when the recognition start command is received, even if the recognition unit 4 is executing the recognition process, that process is aborted and a new recognition process is started. The recognition result is output to the output terminal 12. Any existing recognition unit can be used.

In the explanation, a beep sound is used as the input prompting signal, but a similar configuration is possible for voice response and display display.

Which one is appropriate to use as the input promotion signal depends on the application. For example, the display display cannot be used for voice input via a telephone.

Voice responses are suitable for non-specific speakers, where it is difficult to explain how to use them, but voice responses themselves take time.
If the number of users is limited, a signal such as a "beep" sound is suitable. It is also possible to use these input prompting signals in combination. In the explanation, the voice detection section was used to implement the voice detection function, but this function is also included in the voice recognition section, so if the voice detection function can be separated from the voice recognition section, use this function. It is also possible.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment according to the present invention, in which 1...--input promotion signal generating section. 2... Re-input command section, 3...
...Audio detection section. It is a chart. 81 Low 223-

Claims (1)

[Claims] an input prompting signal generating means for generating a signal to prompt the user to use the audio input to the ori state name; a voice detecting means for determining whether or not an audio signal is being input; T according to the detection signal output by the detection means
A voice characterized by comprising means for instructing the input prompting signal generating means to regenerate the input prompting signal when a voice input is detected, and a recognition means for recognizing the inputted voice. input device.