US 11,817,113 B2
Systems and methods for filtering unwanted sounds from a conference call
Rajendran Pichaimurthy, Karnataka (IN); and Madhusudhan Seetharam, Karnataka (IN)
Assigned to Rovi Guides, Inc., San Jose, CA (US)
Filed by Rovi Guides, Inc., San Jose, CA (US)
Filed on Sep. 9, 2020, as Appl. No. 17/015,830.
Prior Publication US 2022/0076689 A1, Mar. 10, 2022
Int. Cl. G10L 21/028 (2013.01); H04M 3/56 (2006.01); G10L 21/0232 (2013.01); G10L 15/26 (2006.01); H04M 3/22 (2006.01); G10L 25/84 (2013.01); G10L 21/0208 (2013.01)
CPC G10L 21/0232 (2013.01) [G10L 15/26 (2013.01); G10L 21/028 (2013.01); G10L 25/84 (2013.01); H04M 3/2236 (2013.01); H04M 3/568 (2013.01); G10L 2021/02087 (2013.01); H04M 2203/5072 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A method for filtering unwanted sounds from a conference call, the method comprising:
generating a first voice profile of a first user of a media device based on a first voice signal captured by the media device during a first conference call; and
during a second conference call:
analyzing a second voice signal captured by the media device during the second conference call;
determining, using the first voice profile of the first user, that the second voice signal includes a voice of a second user different from the first user by determining that one or more portions of the second voice signal do not match the first voice profile of the first user;
in response to determining that the second voice signal includes the voice of the second user:
determining, using the first voice profile of the first user, portions of the second voice signal spoken by the second user by determining the one or more portions of the second voice signal that do not match the first voice profile of the first user;
separating the second voice signal into: (a) only portions spoken by the first user, and (b) only the portions spoken by the second user;
modifying a second voice profile of the second user based on the portions of the second voice signal captured by the media device spoken by the second user;
identifying a subject of the second conference call; and
in response to determining, using the first voice profile of the first user, that the first user is familiar with the subject of the second conference call:
allowing the portions of the second voice signal spoken by the first user to be transmitted into the second conference call; and
in response to determining, using the modified second voice profile of the second user, that the second user is not familiar with the subject of the conference call:
preventing the portions of the second voice signal spoken by the second user from being transmitted into the second conference call.