Detecting acoustic echoes using microphone arrays
View Patent ↗A method is provided for discriminating between the case of a single talker with an acoustic reflection and the case of two talkers, regardless of their power levels. The method is implemented in real time by performing a cross-correlation between pairs of average power signals originating from pairs of beamformers. A detection decision is then made based on the value of the cross correlation and its lag.
1. A method of discriminating between a single signal source with reflection and two individual signal sources, said method comprising:
detecting, by a sensor array, a plurality of signals originating from said signal sources, said sensor array for receiving said signals via respective multiple sensors arranged in paired sectors of said array;
Generating, by a beamformer, a plurality of power signals resulting from an average power of the signal received from each of said sectors;
Computing a cross correlation function of said power signals for respective pairs of said sectors;
For each said cross correlation function, recording a maximum cross correlation and a lag between the pair of power signals at which said maximum cross correlation occurs; and
Identifying said single signal source with reflection in the event that said lag is of high absolute magnitude, and identifying said two individual signal sources in the event that said lag is of low absolute magnitude.
2. The method of claim 1 , further comprising normalizing said power signals prior to computing said cross correlation function.
3. The method of claim 2 , further comprising determining active ones of said sectors, wherein a sector is deemed to be active if the respective power signal exceeds a predetermined threshold, and normalizing the power signals from only the active ones of said sectors.
4. The method of claim 1 , wherein said received signals are audio signals and said sensor array is a microphone array.
5. The method of claim 2 , wherein said received signals are audio signals and said sensor array is a microphone array.
6. The method of claim 3 , wherein said received signals are audio signals and said sensor array is a microphone array.
7. The method of claim 4 , wherein said function is computed via delay and sum beamforming.
8. The method of claim 5 , wherein said function is computed via delay and sum beamforming.
9. The method of claim 6 , wherein said function is computed via delay and sum beamforming.