A Multimicrophone Voice Activity Detection System Based on Mutual Information
In many applications, such as hands-free telephony and smart environments, there is a need to distinguish voice from background noise and reverberation. A proposed system computes an information-theoretic coherency between multiple microphones in order to detect near-field speech. The system performance was compared to that of a reference case in both simulations and experiments. While all such systems degrade with increasing noise and reverberation, the proposed system proved to be more robust under adverse conditions.
Click to purchase paper 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 $20 for non-members, $5 for AES members and is free for E-Library subscribers.