Wavelets are recognized to be an efficient tool for signal compression. They can be matched to different parts of a signal so that only a few wavelet tree nodes can carry most of the signal energy. Restrictions to a maximum level (number of down samplings and therefore analysis bands) arise because computational complexity is increased or only a finite data record is available. Further, a non-adaptive wavelet transform does not allow optimal matching between signal and wavelet filter bank and for best basis wavelet as adaptive procedure, we cannot merge adjacent terminal nodes when they come from different parents. We reduce these restrictions and combine several potentials to increase the compression rate. By changing the sampling frequency we change the relative location of the signal frequency spectrum with respect to the terminal node filter bands. We show that even though the amount of redundant data increases when the sampling frequency goes above Nyquist's, we nevertheless achieve improved compression rate.
[1]
Deepen Sinha,et al.
Low bit rate transparent audio compression using adapted wavelets
,
1993,
IEEE Trans. Signal Process..
[2]
M. Vetterli,et al.
Wavelets, subband coding, and best bases
,
1996,
Proc. IEEE.
[3]
K Ramchandran,et al.
Best wavelet packet bases in a rate-distortion sense
,
1993,
IEEE Trans. Image Process..
[4]
M. Victor Wickerhauser,et al.
Adapted wavelet analysis from theory to software
,
1994
.
[5]
Ingrid Daubechies,et al.
Ten Lectures on Wavelets
,
1992
.
[6]
Ronald R. Coifman,et al.
Entropy-based algorithms for best basis selection
,
1992,
IEEE Trans. Inf. Theory.
[7]
Adrian S. Lewis,et al.
Image compression using the 2-D wavelet transform
,
1992,
IEEE Trans. Image Process..