Acoustic voice activity detection (AVAD) for electronic systems
Acoustic Voice Activity Detection (AVAD) methods and systems are described. The AVAD methods and systems, including corresponding algorithms or programs, use microphones to generate virtual directional microphones which have very similar noise responses and very dissimilar speech responses. The ratio of the energies of the virtual microphones is then calculated over a given window size and the ratio can then be used with a variety of methods to generate a VAD signal. The virtual microphones can be constructed using either an adaptive or a fixed filter.
1. An acoustic voice activity detection system comprising:
a first virtual microphone comprising a first combination of a first signal and a second signal, wherein the first signal is received from a first physical microphone and the second signal is received from a second physical microphone;
a second virtual microphone comprising a second combination of the first signal and the second signal, the first and second virtual microphones having substantially identical responses to noise; and
an adaptive filter configured to reduce speech response of the second virtual microphone relative to the first virtual microphone, wherein a magnitude of the adaptive filter is limited to predetermined expected speech values to preclude the adaptive filter from accidental training on noise when noise is present during training of the adaptive filter; and
wherein acoustic voice activity of a speaker is determined to be present when an energy ratio (R) of energies of the first virtual microphone and the second virtual microphone is greater than a threshold value.
2. The system of claim 1 , wherein the magnitude of the adaptive filter is limited to between approximately 0.82 and approximately 0.88 to preclude the adaptive filter from accidental training on noise when noise is present during training of the adaptive filter.
3. The system of claim 1 , wherein the threshold value is approximately two.
4. The system of claim 1 , wherein training of the adaptive filter only occurs when a current value of R is larger than a smoothed history of R values.
5. The system of claim 1 , wherein generating of the energy ratio comprises generating the energy ratio for frequencies in a range from approximately 250 Hz to 1250 Hz.
6. The system of claim 1 , wherein generating of the energy ratio comprises generating the energy ratio for frequencies in a range from approximately 20 Hz to 3000 Hz.