Wavelet bandwith, centre frequency
6 views (last 30 days)
Show older comments
The Wavelet Toolbox provides the fucntion scal2frq which gives a so called pseudo-frequency to a corresponding scale.
As I understand it, each scale represents a daughter wavelet. In contrast to a periodic sine wave doesn´t a wavelet represent a frequency range rather than one frequency?
How can one calculate the bandwidth of a wavelet?
Is there a way of getting the actual function of the scaled wavelets used in cwt in order to perform an fft on them?
0 Comments
Accepted Answer
Wayne King
on 19 Nov 2011
Hi Chris, there are a couple things you can do:
1.) If you want to see the support of wavelets at different scales, you can input a shifted Kronecker delta sequence
x = zeros(1000,1);
x(500) = 1;
cfs = cwt(x,1:2:64,'sym8');
% next part obviously not needed
for nn = 1:32
plot(cfs(nn,:)); pause(0.2);
end
2.) The other thing is that the Wavelet Toolbox now has a DFT-based continuous wavelet transform, implemented in cwtft.m and an inverse CWT in icwtft.m. The wavelets supported in cwtft have their Fourier transforms explicit. See the reference page for cwtft.m
Also see:
for some examples.
0 Comments
More Answers (1)
Wayne King
on 19 Nov 2011
Hi Chris, here is a quick example of the first option above:
x = zeros(1000,1);
x(500) = 1;
cfs = cwt(x,1:2:64,'sym8');
Fs = 1000;
wavdft = abs(fft(cfs(5:10:30,:)'));
wavdft = wavdft(1:length(x)/2+1,:);
freq = 0:(Fs/length(x)):500;
plot(freq,abs(wavdft),'linewidth',2);
xlabel('Hz'); ylabel('Magnitude');
legend('Scale 10','Scale 30','Scale 60');
See Also
Categories
Find more on Continuous Wavelet Transforms in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!