Discrete wavelet decomposition practical approaches/algorithms/complexity

Started by Alex Z August 26, 2006
Is there anyone who has an experience of discrete wavelet transform
implementation to achieve signal subband separation ?
I'm working on certain signal processing that is handled in subbands,
the signal is separated into reasonably narrow frequency subbands and
then each subband is processed separately. The subband separation is to
be attained by means of discrete wavelet transform (dwt() in MATLAB)
decomposition. Following subband processing, the processed subbands are
re-composed back which yields the processed signal.
In MATLAB the decomposition is performed by dwt() while recomposition
is inverse dwt, i.e. idwt().
I'm approaching the step of practical implementation assessment in
which I have to gain an understanding of practical wavelet
decomposition (and recomposition) algorithms used in DSP and their
computational complexity. I have to get clear understanding of whether
actual DWT algorithms will work on particular DSP hardware at
particular constraints.
I'mm be grateful if someone could share his experience in related field
or at least to point me to the appropriate resources that cover (at
least partially) the issue.

Thanks in advance, Alex