A new multiple-input multiple-output frame-processing algorithm is introduced that exploits blind deconvolution for acoustic echo cancellation. The channel deconvolution filters, which can be blindly estimated as either finite impulse or infinite impulse responses, are optimized by maximizing the information flow through several non-linear neurons. The algorithm requires that for every system audio output there be a corresponding microphone for effective feedback signal separation/cancellation. The desired talker signal from the algorithm outputs is recognized and transmitted while the identified feedback signals are discarded.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.