Continuous wavelet transform of mirrored 1-D signals
[scalo,f,T,a,wt,wavescaled] = contwtgnmir(x,fmin,fmax,N,wave);
signal (in time) to be analyzed
respectively lower and upper frequency bounds of the analysis (in cycles/sec).
number of analyzed voices
specifies the analyzing wavelet An order "wave" derivative of the Gaussian is chosen
scalogram (squared magnitude of WT)
frequency samples (geometrically sampled between FMAX and FMIN).
time samples
scale vector (geometrically sampled between 1 and FMAX/FMIN)
coefficient of the wavelet transform. X-axis corresponds to time (uniformly sampled), Y-axis corresponds to frequency (or scale) samples (geometrically sampled between Fmin (resp. Fmax/Fmin aand Fmax (resp. 1) First row of WT corresponds to the lowest analyzed frequency.
when the analyzing wavelet is Morlet or Mexican hat, wavescaled = wave. For an aritrary band-pass analyzing function, wavescaled contains columnwise the (N) scaled version of it
If x = [a b c e f] is the signal to analyzed, contwtmir runs contwt on the mirrored version XxX = [c b [a b c d e f] e d]. The number of mirrored samples depends on the analyzed scale and the wavelet length. USE AN ORDER "wave" DERIVATIVE OF THE GAUSSIAN