Back to results
US 9,584,946 B1Granted
Audio diarization system that segments audio input
OwnerPhilip Scott Lyren, Glen A. Norrisinventors
Life of this patent
Estimating the term…
Abstract
Speech and/or non-speech in an audio input are convolved to localize sounds to different locations for a user. An audio diarization system segments the audio input into speech and non-speech segments. These segments are convolved with one or more head related transfer functions (HRTFs) so the sounds localize to different sound localization points (SLPs) for the user.
Claims20
A method executed by one or more electronic devices in a computer system to convolve sounds in a telephone call to different locations for a listener, the method comprising:
- determining, by the one or more electronic devices in the computer system, the telephone call as a source of an audio input;
- segmenting, by an audio diarization system, the audio input of the telephone call into a speech segment that is a voice and a non-speech segment that is music;
- analyzing, by the one or more electronic devices in the computer system, sound localization point (SLP) locations in prior telephone calls where voices in the prior telephone calls localized to the listener;
- determining, by the one or more electronic devices in the computer system and based on analysis of the SLP locations in the prior telephone calls, preferred SLP locations where the listener previously localized the voices in the prior telephone calls;
- storing, in the user preferences of the listener, the preferred SLP locations where the listener desires to localize the voices in the telephone calls;
- convolving, by the one or more electronic devices in the computer system and with a first pair of head related transfer functions (HRTFs), the speech segment so the voice externally localizes to a first SLP that is external to the listener;
- convolving, by the one or more electronic devices in the computer system and with a second pair of HRTFs, the non-speech segment so the music externally localizes to a second SLP that is external to the listener and at a different location than the first SLP;
- and
- providing, by the one or more electronic devices in the computer system, the speech segment to the listener so the voice externally localizes to the first SLP, and the non-speech segment to the listener so the music externally localizes to the second SLP.
6 more elements
A method to convolve a voice in a telephone call to an external location that is familiar to a user, the method comprising:
- storing, in memory of a handheld portable electronic device (HPED), user preferences of the user that include preferred locations where the user desires to localize voices in the telephone call and preferred locations where the user desires to localize music in the telephone call;
- segmenting, with an audio diarization system, the telephone call into a speech segment and a non-speech segment in order to identify a name of a speaker in the speech segment of the telephone call;
- retrieving, from the memory in the HPED, a head related transfer function (HRTF) that is assigned to the speaker in the user preferences of the user;
- convolving, with a digital signal processor (DSP) in the HPED, the speech segment with the HRTF so a voice of the speaker localizes to a sound localization point (SLP) that is away from but proximate to the user;
- and
- providing, by the HPED, the voice of the speaker to the user so the voice of the speaker localizes to a SLP that is familiar to the user as an external location where voices in telephone calls localize to the user;
- analyzing SLP locations in prior telephone calls where voices in the prior telephone calls localized to the user;
- determining, based on analysis of the SLP locations in the prior telephone calls, a preferred SLP location where the user previously localized the voices in the prior telephone calls;
- and
- storing, in the user preferences of the user, the preferred SLP location where the user desires to localize the voices in the telephone calls.
7 more elements
A handheld portable electronic device (HPED) of a user that convolves a telephone call, the HPED comprising:
- an audio diarization system that segments the telephone call into a first speech segment of a first person and a second speech segment of a second person;
- a memory that stores user preferences that include preferred sound localization point (SLP) locations for where the user prefers to localize voices in telephone calls, a first head related transfer function (HRTF) of the user, and a second HRTF of the user;
- a SLP selector that analyzes where voices in prior telephone calls localized to the user, determines the preferred SLP locations based on where the user previously localized the voices in the prior telephone calls, stores the preferred SLP locations in the user preferences, and assigns, based on the user preferences for where the user prefers to localize the voices in the telephone calls, the first HRTF to a voice of the first person and assigns the second HRTF to a voice of the second person;
- and
- a digital signal processor (DSP) that convolves, during the telephone call, the first speech segment with the first HRTF so the voice of the first person externally localizes to the user to a first location that is away from but proximate to the user and convolves the second speech segment with the second HRTF so the voice of the second person externally localizes to the user to a location that is away from but proximate to the user, wherein the first location and the second location are separate from each other.
2 more elements
Dependencies are inferred from the claim text.