×

Separation of target acoustic signals in a multi-transducer arrangement

  • US 7,366,662 B2
  • Filed: 08/09/2006
  • Issued: 04/29/2008
  • Est. Priority Date: 07/22/2004
  • Status: Active Grant
First Claim
Patent Images

1. A speech separation process, comprising:

  • providing a plurality of microphones with respect to a speech source so that each respective microphone generates a signal having a speech component and a noise component in different mixing ratios;

    receiving each of the signals generated by the microphones into a separation process;

    separating the received signals into a first channel and a second channel using a blind-source separation process, wherein one of the channels provides a noise signal that is substantially noise components and the other channel provides a combination signal that is a combination of noise components and speech components;

    identifying which of the first or second channels has the combination signal;

    processing the combination signal with the noise signal; and

    generating a speech signal indicative of the speech from the speech source.

View all claims
  • 6 Assignments
Timeline View
Assignment View
    ×
    ×