Next: Bayesian Framework Application
Up: BAYESIAN TWO SOURCE MODELING
Previous: DUET and DASSS Review
As suggested above, the DASSS data produced by equation 4
may in fact reveal which two sources are active at a
particular point in time-frequency space if exactly two sources
are active. This is very useful information, because once the
active sources are known, they may be demixed by solving for
and in:
which follows directly from equations 1 and 2 when
only two sources are active.
Formally, we express the two most likely sources given some DUET
or DASSS data as those maximizing . Applying
Bayes' rule, we can express this as
We can see by inspection of equations 9
through 11 that the STFT frequency under
consideration, , affects DASSS data (the
values). So, we are mindful that the problem is a different one
for each of the frequency values under consideration.
Since is not a function of or , it is possible to
discard it from the maximization (though we may wish to use it
later if a confidence measure is sought). The quantity is
largely estimated with musical knowledge. For example we may know
that the clarinet (source for example) tends to play at the
same time as and less loudly than the violin (say, source ),
whose frequency components tend to be at harmonics of frequencies
above 200 Hz, and rarely throughout time. Though very useful, such
information is not within our current signal processing interest,
and is not considered now. (For now, we will treat all as
equally likely.)
We are left, then, to consider for each , the
probability that particular DASSS data is produced when sources
and (but no others) are active at frequency . To
this end, we next explicitly return to the distributions suggested
by equations 9 through 11. By doing so, we
identify the necessary values of where represents
DASS scores .
Next: Bayesian Framework Application
Up: BAYESIAN TWO SOURCE MODELING
Previous: DUET and DASSS Review
Aaron S. Master
2003-10-30