Re: Using Haar wavelet with wv_cwt [message #94789 is a reply to message #94000] |
Fri, 13 October 2017 10:12   |
benjamin.castellani
Messages: 7 Registered: October 2017
|
Junior Member |
|
|
On Sunday, December 18, 2016 at 12:32:21 PM UTC-7, owentr...@gmail.com wrote:
> Hi all,
> I've been trying to use wv_cwt to perform the wavelet transform on a 1d time series. It works fine with the morlet and paul wavelets but I cannot use it with the Haar wavelet. All I get in the console is "Keyword parameters not allowed in call."
>
> Any thoughts?
>
> Many thanks,
> Owen
I apologize that our wavelet documentation is a little "light" and confusing. The issue is that the Haar wavelet family is of type discrete. WV_CWT is for continuous wavelets only. You will want to use WV_DWT (for discrete).
You need to run WV_FN_HAAR separately first to create your coefficients based on the order you define.
Then you input all that information into WV_DWT to develop create wavelet transform.
============
Example:
coefficents_parameters = wv_fn_haar(1,scaling,wavelet,ioff,joff)
data = randomu(1,64)
part_wv = wv_dwt(data, wavelet,scaling,ioff,joff, N_LEVELS=3)
Does this make sense? For continuous families, you can use WV_CWT directly. For discrete families, you must wave the WV_FN_FAMILYNAME function first, then WV_DWT.
============
Hope this helps.
P.S. I will look into making this more clear in the IDL Documentation Center.
Ben Castellani
IDL Team
Harris Geospatial Solutions
|
|
|