Adaptive models with frames or overcomplete representations might be
interesting. Then you can decompose the signal with a broader class of basis
functions, such as symmetrically windowed sinusoids, chirps, exponentially
decaying sinusoids and whatever. The point is to have a sparse
Several useful ideas that are worth trying. My goal is to classify
Indian music into notated streams from audio samples. As I progress down
this path, I will publish my results to further research in this field.
From:Risto Holopainen
On Thu, 10 Jul 2014, Vadim Zavalishin wrote:
From the BLEP discussion we know, that so
far this signal is just a generalized version of the DC offset, thus
containing only a zero frequency partial.
The ``no new partials'' rule comes from
integral exp(kx) = exp(kx)/k
so integration just
