Next: Frequency to eigendomain transform
Up: Speech Enhancement Summaries
Previous: Bayes optimal decision rule
This involves deriving the MMSE STSA estimator using a complex Gaussian model of the a priori probability distribution of speech and noise Fourier expansion coefficients.
If
and
, then the MMSE estimator of
is
With the assumption of Fourier coefficients having a Gaussian distribution, the polar form of the coefficients have the following marginal distribution
and
The prior pdf is
The joint pdf is
The posterior density can be worked out to be
where
The authors use the first moment of the posterior distribution giving
They also extend the amplitude estimator under signal presence uncertainty (see for example Maximum Likelihood estimator) but this is beyond the scope of this summary.
Next: Frequency to eigendomain transform
Up: Speech Enhancement Summaries
Previous: Bayes optimal decision rule
Vinesh Bhunjun
2004-09-17