Voice-identification-based signal processing for multiple-talker applications
US7995713B2 · kind B2 · utility
7Cited by
17References
21Claims
0Family size
Assignee
Inventor
Key dates
| Filing date | Apr 3, 2006 |
| Grant date | Aug 9, 2011 |
| Priority date | — |
| Expiry date | Apr 28, 2030 |
Classification
- Technology area (CPC H)Electricity
- CPC primaryH04M2201/41
- WIPO fieldTelecommunications
- WIPO sectorElectrical engineering
Abstract
The audio signals associated with different co-located groups of talkers in a teleconference are detected (e.g., by comparing the voiceprint for the current talker group with stored voiceprints corresponding to all of the co-located teleconference participants) and processed using different and appropriate automatic gain control (AGC) levels, where each group has a corresponding stored AGC level. Depending on the embodiment, each group may have one or more participants.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.