US 11,705,101 B1
Irrelevant voice cancellation
Akinobu Morishima, Tokyo (JP); Akio Oka, Kawasaki (JP); Sho Ayuba, Tokyo (JP); Eri Kimura, Setagaya (JP); and Yukiko Furusho, Tachikawa (JP)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Mar. 28, 2022, as Appl. No. 17/656,670.
Int. Cl. G10K 11/178 (2006.01); G10L 25/78 (2013.01)
CPC G10K 11/17823 (2018.01) [G10K 11/17873 (2018.01); G10L 25/78 (2013.01); G10K 2210/3027 (2013.01); G10K 2210/3044 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method, comprising:
receiving first audio data from a first computing device via a first microphone;
receiving second audio data from a second computing device via a second microphone;
detecting a first user utterance having a first user utterance local decibel measure in the first audio data and the first user utterance having a first user utterance remote decibel measure in the second audio data;
detecting a second user utterance having a second user utterance local decibel measure in the second audio data and the second user utterance having a second user utterance remote decibel measure in the first audio data;
determining a first cancelling coefficient based on the second user utterance detected at the first microphone and the second user utterance detected at the second microphone;
applying the first cancelling coefficient to the first audio data; and
generating first output audio data comprising the first user utterance at the first computing device.