1
$\begingroup$

What is currently the most accurate Voice Activity Detection algorithm with high levels of background noise? I would also appreciate if it does not require any manual calibration. Bonus points for examples.

EDIT: I am referring to accuracy as yielding the lowest amount of false positives and negatives.

$\endgroup$
  • $\begingroup$ What sort of background noise? What sort of voices (talking? singing?) ? $\endgroup$ – Peter K. Nov 17 '13 at 18:58
  • $\begingroup$ Talking. Ambient background noise. $\endgroup$ – Skylion Nov 17 '13 at 19:03
  • $\begingroup$ As the fact that the question is put on hold, I disagree. Accuracy is a quantitative not qualitative measurement. -_- $\endgroup$ – Skylion Nov 17 '13 at 19:03
  • $\begingroup$ Please re-word (edit) your question to make the accuracy quantitative. What measure are you using? $\endgroup$ – Peter K. Nov 17 '13 at 19:04
  • $\begingroup$ Is that better? $\endgroup$ – Skylion Nov 17 '13 at 19:06
1
$\begingroup$

If you have more than 1 detectors, you may try blind separation based on the power spectrum difference of voice and background noise

$\endgroup$
  • $\begingroup$ Define more than 1 detectors? As in stereo vs mono? $\endgroup$ – Skylion Nov 27 '13 at 17:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.