Voice processing device, voice processing method and voice processing system
This voice processing device is provided with: an utterer's position detection unit which specifies, as position microphones of an utterer, microphones that receive a voice signal of WuW on the basis of the characteristics of each voice signal for a prescribed time, when the WuW voice is detected, the voice signal being held in a voice signal buffer unit; and a CTC unit (one example of a voice processing unit) which outputs a voice uttered by the utterer and suppress a voice uttered by an occupant, who is not the utterer, by using the voice signal for the prescribed time, which is held in the voice signal buffer unit, and information relating to the utterer's position microphones.
1. A voice processing device installed in a vehicle and in which plural different microphones are arranged so as to correspond to respective seats, the voice processing device comprising:
at least one memory that stores instructions and voice signals from the plural different microphones; and
a processor that, when executing the instructions stored in the at least one memory, performs a process, wherein the process includes:
storing, in the at least one memory, the voice signals collected by the plural different microphones, respectively, during a prescribed period before a present time, the voice signals being repeatedly stored as buffered voice signals;
detecting whether a prescribed word is uttered by a speaker in the vehicle based on the voice signals collected by the plural different microphones, respectively;
determining, in response to the prescribed word being uttered and as a speaker position microphone, a microphone closest to the speaker by referring to the buffered voice signals stored in the at least one memory; and
in response to the determining of the speaker position microphone:
setting coefficients of the plural different microphones other than the speaker position microphone to suppress the voice signals collected by the plural different microphones other than the speaker position microphone;
stopping the storing of the buffered voice signals and the detecting of whether the prescribed word is uttered;
setting coefficients of the speaker position microphone to suppress voices uttered by persons other than the speaker; and
outputting a voice uttered by the speaker while suppressing the voices uttered by the persons other than the speaker.
2. The voice processing device according to claim 1 , wherein the process further includes:
suppressing reproduction sound signals collected by the plural different microphones by a reproduction sound of a music reproduction device installed in the vehicle.
3. The voice processing device according to claim 1 , wherein the process further includes:
forming voice signal directivity that is directed to the speaker corresponding to a sound source of a voice signal collected by the speaker position microphone, wherein:
each of the plural different microphones is a micro-array having plural microphone elements.
4. The voice processing device according to claim 1 , wherein utterance of the prescribed word by the speaker is detected, by the processor, based on a voice signal collected by a particular one of the plural different microphones.
5. The voice processing device according to claim 1 , wherein the process further includes:
changing an operation mode of the voice processing device from a particular speaker voice output mode for outputting the voice uttered by the speaker while suppressing the voices uttered by the other persons to a prescribed word detection standby mode for detecting whether the prescribed word is uttered, when having detected satisfaction of a prescribed condition.
6. The voice processing device according to claim 5 , wherein:
presence/absence an end word that is different from the prescribed word is detected, by the processor, based on a voice signal collected by the speaker position microphone; and
the operation mode of the voice processing device is changed to the prescribed word detection standby mode by judging that the prescribed condition has been satisfied when a voice of the end word that is different from the prescribed word has been detected.
7. The voice processing device according to claim 5 , wherein the operation mode of the voice processing device is changed, by the processor, to the prescribed word detection standby mode by judging that the prescribed condition has been satisfied when a prescribed period has elapsed from acquisition of a recognition result of the voice uttered by the speaker.
8. A voice processing method employed in a voice processing device, the voice processing device configured to be installed in a vehicle in which plural different microphones are arranged so as to correspond to respective seats, the voice processing method comprising:
storing, in a memory, voice signals collected by the plural different microphones, respectively, during a prescribed period before a present time, the voice signals being repeatedly stored as buffered voice signals;
detecting whether a prescribed word is uttered by a speaker in the vehicle based on the voice signals collected by the plural different microphones, respectively;
determining, in response to the prescribed word being uttered and as a speaker position microphone, a microphone closest to the speaker by referring to the buffered voice signals stored in the memory; and
in response to the determining of the speaker position microphone:
setting coefficients of the plural different microphones other than the speaker position microphone to suppress the voice signals collected by the plural different microphones other than the speaker position microphone;
stopping the storing of the buffered voice signals and the detecting of whether the prescribed word is uttered;
setting coefficients of the speaker position microphone to suppress voices uttered by persons other than the speaker; and
outputting a voice uttered by the speaker while suppressing the voices uttered by the persons other than the speaker.
9. A voice processing system comprising a voice processing device installed in a vehicle in which plural different microphones are arranged so as to correspond to respective seats and a control device which controls a vehicular device installed in the vehicle, wherein:
the voice processing device
is configured to store, in a memory, voice signals collected by the plural different microphones, respectively, during a prescribed period before a present time, the voice signals being repeatedly stored as buffered voice signals;
is configured to detect whether a prescribed word is uttered by a speaker sitting in the vehicle based on the voice signals collected by the plural different microphones, respectively;
is configured to determine, in response to the prescribed word being uttered and as a speaker position microphone, a microphone closest to the speaker by referring to the buffered voice signals stored in the memory; and
in response to the determining of the speaker position microphone:
is configured to set coefficients of the plural different microphones other than the speaker position microphone to suppress the voice signals collected by the plural different microphones other than the speaker position microphone;
is configured to stop storing of the buffered voice signals and detecting of whether the prescribed word is uttered;
is configured to set coefficients of the speaker position microphone to suppress voices uttered by persons other than the speaker;
is configured to output a voice uttered by the speaker while suppressing the voices uttered by the persons other than the speaker; and
is configured to acquire a recognition result of the voice uttered by the speaker; and
the control device is configured to control operation of the vehicular device according to the recognition result of the voice uttered by the speaker.