JPS59153238A - Voice input/output system - Google Patents

Abstract

PURPOSE:To attain an improved interface provided with the atmosphere in which the interaction with a person is possible by increasing the comprehension degree of voice information and the understanding degree of the contents of the information, and adjusting the speed of a response voice to be outputted in accordance with a voice generating speed. CONSTITUTION:A voice detector 3 detects voice generating speed and input information inputted from an acoustoelectric converter 1 and transmits the input information to an information processor 5. Response information outputted from the information processor 5 is voice-synthesized at an audio response unit 4. In this case, the voice generating speed is controlled by a standardized parameter from a characteristic transmission function information line 6. On the other hand, the voice detector 3 consisting of a voice recognizing device, voice waveform analyzer, voice frame detector, etc. standardizes the expansion/compression of human's voice in the time axial direction and an interval between syllables. Thus, the speed of a response voice from the andio response unit 4 can be adjusted.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a voice input/output system in which a human inputs voice, processes it in an information processing device, and outputs voice.

The voice response device in the conventional voice input/output system does not have a means to vary the response speed in response to the human speaking speed, and the output voice rate is constant or even if it is variable, it is not set at the initial setting or The function was limited to changing the speaking speed to a speed predetermined by a control computer. FIG. 1 shows an example in which a voice detection device and a voice response device are used as input/output devices of an information processing device. There is no transmission path for information regarding the voice input speed between the voice detection device and the voice response device.

In conventional voice response devices, the response speed is predetermined and a response voice is always returned at a constant speed, which is reasonably fast. The drawback is that it feels like an obligatory notification, rather than a sense of communication.

SUMMARY OF THE INVENTION An object of the present invention is to provide a voice input/output system that improves the intelligibility of the output voice information and the content of the information, gives the atmosphere of dialogue with a human, and establishes a better man-machine interface.

The voice input/output system of the present invention includes: a voice detection means that outputs speech rate information of voice input; an information processing device that outputs response information according to input information of the voice and input; and voice response means having a function of adjusting the speed of the response voice to be output according to the speech rate information.

In the embodiment of the present invention shown in FIG. 2, only the information on the input speed of the voice detected by the voice detection device and the output information coming from the information processing device are transmitted to the voice response device through the characteristic transfer function information line. When synthesizing, the speech rate is varied and output. The speech rate and input information of the voice input from the acoustoelectric transducer 1 are detected by the voice detection device 3, the input information is transmitted to the information processing device 5, and the response information outputted from there is detected by the voice response device 4. When synthesized, the speech rate is controlled by the standardization parameter from the characteristic transfer function information line 6, and the electroacoustic transducer 2 produces a voice.

By varying the time unit (frame) in which the voice signal is generated in the voice response device 4, it is possible to change the rate of speech without changing the pitch, height, or voice quality of the voice. On the other hand, in the voice detection device 3 such as a voice recognition device, a voice waveform analysis device, a voice frame detection device, etc., in order to reduce the influence of the expansion and contraction of human voice in the time axis direction and the interval between syllables on the recognition result. The time axis is standardized, and it is possible to extract the standardized parameters in the time axis direction as information. By feeding back this parameter to the voice response device 4 through the characteristic transfer function information line 6, the speed of the voice to be responded to is adjusted.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing a conventional example, and FIG. 2 is a block diagram showing an embodiment of the present invention. 1...Acoustic electric transducer, 2...Electroacoustic transducer, 3...Audio detection device, 4...
・Voice response device, 5... Information processing device, 6...
...Characteristic transfer function information line. Agent Patent Attorney Susumu Uchihara

Claims (1)

[Claims] a voice detection means that outputs speech rate information of voice input; an information processing device that outputs response information according to the input information of the voice input; A voice input/output system equipped with a voice response means having a function of adjusting according to information.