[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Blind Source Separation by Sparse Decomposition

Rich & Michael,

When clipping a signal as suggested you get some of the spectral components
"re-coded" as pulse-width modulation. Although the resulting signal is
spectrally substantially different, the variation in spectral "flux" by time
(especially with speech) is the same, hence we are probably talking about a
Gibsonioan invariant in this case.


Mikael Fernström, M.Sc.
Interaction Design Centre & UL Interactive Media Group,
Foundation Building, University of Limerick, Ireland
Phone: +353-61-202606, Fax: +353-61-202572
Email: mikael.fernstrom@ul.ie Web: www.idc.ul.ie