Acoustic signal processing system capable of detecting double-talk and method

1. A method for processing an acoustic signal, comprising:

receiving a first acoustic signal in the time domain;

converting the first acoustic signal in the time domain into a digital audio signal in a sub-band frequency domain;

reserving a first frequency band of the digital audio sub-band frequency signal having a plurality of frequency bands using a frequency band reservation module that includes a digital signal processor and a plurality of control multipliers, wherein a first control multiplier includes a first sub-band frequency input coupled to a first input of the digital signal processor and a second input coupled to a first output of the digital signal processor and a second control multiplier includes a first input coupled to a second input of the digital signal processor and a second input coupled to a second output of the digital signal processor;

transforming the sub-band frequency digital audio signal including the first frequency band into a second acoustic signal using an acoustic transducer;

transforming a third acoustic signal derived from the second acoustic signal and a near end signal into a mixed sub-band frequency signal;

transforming the sub-band frequency digital audio signal into a modified sub-band frequency signal;

generating a sub-band frequency error signal in response to the modified sub-band frequency signal by subtracting the modified sub-band frequency signal from the mixed sub-band frequency signal;

generating a speech activity-indicator in a first reserved frequency band of the mixed sub-band frequency signal in response to the presence of speech activity;

generating an echo activity indicator by subtracting the sub-band frequency error signal from the mixed sub-band frequency signal;

logically ANDing the speech activity indicator and the echo activity indicator to generate a double-talk activity indicator that indicates the presence or absence of double-talk;

generating a second modified digital audio sub-band frequency signal in response to the double-talk activity indicator;

converting the second modified digital audio sub-band frequency signal into a modified time domain signal; and

transmitting the modified time domain signal to a receiver for use by a user.

2. The method of claim 1 , wherein reserving the first frequency band of the digital audio signal includes suppressing speech content in the first frequency band.

3. The method of claim 1 , further including generating a sub-band frequency filter signal in response to the acoustic signal and wherein generating the sub-band frequency error signal includes subtracting the filter signal from the mixed sub-band frequency signal.

4. The method of claim 1 , wherein searching for the echo in the mixed sub-band frequency signal includes determining an auto-correlation metric of the sub-band frequency error signal and an auto-correlation metric of the mixed sub-band frequency signal.

5. The method of claim 1 , wherein reserving the first frequency band includes permanently suppressing an audio signal in the first frequency band.

6. The method of claim 1 , wherein reserving the first frequency band includes configuring a band reservation module to permanently select which frequency band is the first frequency band.

7. The method of claim 1 , wherein reserving the first frequency band includes one of pseudo-randomly selecting which band is the first frequency band or psycho-acoustically selecting which band is the first frequency band.

8. A method for processing acoustic signals, comprising:

receiving an acoustic signal in the time domain;

converting the acoustic signal in the time domain into a digital audio signal in a sub-band frequency domain;

dividing the digital audio signal in the sub-band frequency domain into a plurality of frequency bands;

suppressing speech content of at least one frequency band of the plurality of frequency bands to form a modified first acoustic signal using a frequency band reservation module that includes a digital signal processor and a plurality of control multipliers, wherein a first control multiplier includes a first input coupled to a first input of the digital signal processor and a second input coupled to a first output of the digital signal processor and a second control multiplier includes a first input coupled to a second input of the digital signal processor and a second input coupled to a second output of the digital signal processor;

generating a sub-band frequency error signal from a mixed sub-band frequency signal, wherein the mixed sub-band frequency signal comprises a representation of the mixture of the first acoustic signal and another acoustic signal;

generating a speech detection signal using a speech detection module in response to detecting speech in the mixed sub-band frequency signal and detecting echo in the mixed sub-band frequency signal in response to a difference between the mixed sub-band frequency signal and the sub-band frequency error signal, wherein an error indication signal is generated;

logically ANDing the echo indication signal and the speech detection signal to generate a double-talk indicator signal;

generating a modified digital audio signal in the sub-band frequency domain from the double-talk indicator signal;

converting the modified digital audio signal in the sub-band frequency domain into a modified acoustic signal in the time domain; and

transmitting the modified acoustic signal in the time domain to a receiver for use by an user.

9. The method of claim 8 , further including subjecting the modified first acoustic signal to a filtering process of a sub-band frequency adaptive filter to generate a sub-band frequency filter signal and generating the sub-band frequency error signal by subtracting the sub-band frequency filter signal from the mixed sub-band frequency signal.

10. The method of claim 8 , wherein suppressing the speech content includes temporarily suppressing the speech content.