Robust speaker localization in presence of strong noise interference systems and methods
Systems and methods include a plurality of audio input components configured to generate a plurality of audio input signals, and a logic device configured to receive the plurality of audio input signals, determine whether the plurality of audio signals comprise target audio associated with an audio source, estimate a relative location of the audio source with respect to the plurality of audio input components based on the plurality of audio signals and a determination of whether the plurality of audio signals comprise the target audio, and process the plurality of audio signals to generate an audio output signal by enhancing the target audio based on the estimated relative location. The logic device is further configured to use relative transfer-based covariance to construct directional covariance matrix aligned across frequency bands and find a direction that minimizes beam power subject to distortionless criteria.
1. A method comprising:
receiving a multichannel audio signal from a plurality of audio input components;
determining whether the multichannel audio signal includes target audio associated with an audio source;
transforming the multichannel audio signal to a time-frequency domain vector associated with a plurality of frequency sub-bands;
computing a noisy-speech covariance matrix as a function of the time-frequency domain vector where the target audio is present;
computing a noise-only covariance matrix as a function of the time-frequency domain vector where the target audio is absent;
estimating a location of the audio source relative to the plurality of audio input components based on the noisy-speech covariance matrix and the noise-only covariance matrix; and
processing the multichannel audio signal to generate an audio output signal by enhancing the target audio in the multichannel audio signal based on the estimated relative location.
2. The method of claim 1 , wherein the estimating of the relative location of the audio source comprises:
estimating a target speech relative transfer function based on an eigenvalue decomposition of the noisy-speech covariance matrix and the noise-only covariance matrix.
3. The method of claim 1 , further comprising computing a modified covariance-based localization to identify time difference of arrival.
4. The method of claim 1 , further comprising determining whether an input audio frame is a speech frame or a non-speech frame.
5. The method of claim 1 , further comprising constructing a steering matrix for each of a plurality of frequency bands, using one of the audio input components as a reference.
6. The method of claim 1 , further comprising computing a relative transfer function-based covariance when speech is active.
7. The method of claim 1 , further comprising constructing directional covariance matrix coherently aligned across frequency bands.
8. The method of claim 1 , further comprising determining a direction that minimizes beam power subject to distortionless criteria; and picking a time difference of arrival that yields a maximum likelihood of the target audio associated with the audio source.
9. A system comprising:
a plurality of audio input components configured to generate a plurality of audio input signals; and
a logic device configured to:
determine whether the plurality of audio signals includes target audio associated with an audio source;
transform the plurality of audio signals to a time-frequency domain vector associated with a plurality of frequency sub-bands;
compute a noisy-speech covariance matrix as a function of the time-frequency domain vector where the target audio is absent;
compute a noise-only covariance matrix as a function of the time-frequency domain vector where the target audio is present;
estimate a location of the audio source relative to the plurality of audio input components based on the noisy-speech covariance matrix and the noise-only covariance matrix; and
process the plurality of audio signals to generate an audio output signal by enhancing the target audio based on the estimated relative location.
10. The system of claim 9 , wherein the logic device is to estimate the location of the audio source by:
estimating a target speech relative transfer function based on an eigenvalue decomposition of the noisy-speech covariance matrix and the noise-only covariance matrix.
11. The system of claim 9 , wherein the logic device is further configured to compute a modified covariance-based localization to identify time difference of arrival.
12. The system of claim 9 , wherein the logic device is further configured to determine whether an input audio frame is a speech frame or a non-speech frame.
13. The system of claim 9 , wherein the logic device is further configured to construct a steering matrix for each of a plurality of frequency bands, using one of the audio input components as a reference.
14. The system of claim 9 , wherein the logic device is further configured to compute a relative transfer function-based covariance when speech is active.
15. The system of claim 9 , wherein the logic device is further configured to construct a directional covariance matrix coherently aligned across frequency bands.
16. The system of claim 9 , wherein the logic device is further configured to determine a direction that minimizes beam power subject to distortionless criteria; and picking a time difference of arrival that yields a maximum likelihood of the target audio associated with the audio source.