Infinite impulse response acoustic echo cancellation in the frequency domain

A system for reducing echo in an audio signal, the system comprising:

• a frequency domain based infinite impulse response (IIR) filter to model longer term echo components of a received audio signal, the received audio signal including a combination of (1) a speech signal and (2) a modified reference signal, the modified reference signal resulting from transformation of a reference signal by an echo path transfer function;
• a frequency domain based finite impulse response (FIR) filter to model shorter term echo components of the received audio signal;
• and
• a summing circuit to subtract an echo correction signal from the received audio signal to reduce the echo in the received audio signal to generate an estimate of the speech signal, the echo correction signal generated by applying, in the frequency domain, the IIR filter and the FIR filter to the reference signal.

At least one non-transitory computer readable storage medium having instructions encoded thereon that, when executed by one or more processors, cause a process to be carried out for reducing echo in an audio signal, the process comprising:

• generating coefficients of an infinite impulse response (IIR) filter and coefficients of a finite impulse response (FIR) filter in the frequency domain, the IIR filter and the FIR filter to estimate an echo path transfer function associated with a received audio signal, the received audio signal including a combination of (1) a speech signal and (2) a modified reference signal, the modified reference signal resulting from transformation of a reference signal by the echo path transfer function, wherein the IIR filter is configured to model longer term echo components and the FIR filter is configured to model shorter term echo components;
• applying, in the frequency domain, the IIR filter and the FIR filter to the reference signal, to generate an echo correction signal;
• and
• subtracting the echo correction signal from the received audio signal to reduce echo in the received audio signal and to generate an estimate of the speech signal.