alexis-mignon/pydeconv

blind deconvolution algorithm

Opened this issue · 4 comments

I'm curious what algorithm is your blind deconvolution implementing?

It's an adaptation from the algorithm presented in :

High-quality Motion Deblurring from a Single Image
Q. Shan, J Jia, A Agarwala, SIGGRAPH 2008

The differences with the original algo is that it doesn't use fourrier transforms and the global prior
is a single lorentzian distribution.

It's a very early developpment and for me it doesn't work so well

Is there any chance that I could use this for blind deconvolution of audio signal mixtures?

I don't think so. What kind of decomposition are you actually looking for ?

I have samples of audio recording from two microphones separated by some distance. There are two simultaneous speakers. Basically a convolutive mixture of two sound sources.

Another example of such a mixture is available here (the second example)