Patent · US Active

Separating audio sources

US10176826B2 · kind B2 · utility

5Cited by
9References
19Claims
0Family size

Assignee

Inventor

Key dates

Filing dateFeb 11, 2016
Grant dateJan 8, 2019
Priority date
Expiry dateFeb 11, 2036

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG10L2021/02166
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

Example embodiments disclosed herein relate to source separation in audio content. A method for separating sources from audio content is disclosed, the audio content being of a multi-channel format based on a plurality of channels. The method comprises performing a component analysis on the audio content for each of the plurality of channels to generate a plurality of components, each of the plurality of components comprising a plurality of time-frequency tiles in full frequency band; generating at least one dominant source with at least one of the time-frequency tiles from the plurality of the components and separating the sources from the audio content by estimating spatial parameters and spectral parameters based on the dominant source. Corresponding system and computer program product are also disclosed.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.