|
| From: | Eric Chassande-Mottin |
| Subject: | Re: [Tftb-help] tfrscalo help |
| Date: | Mon, 30 Mar 2009 18:48:02 +0200 (CEST) |
| User-agent: | Alpine 2.00 (DEB 1167 2008-08-23) |
I think that this way nha says when |ha|=1e-2 :
ha = exp(-(2*log(10)/nha^2)*tha.^2)
= 10^(-2(tha/nha)^2
So when tha=nha, |ha|=1e-2.
This makes the interpretation of nha as a window width more clear.
so, I gave the wrong argument! log(10) is not directly related to the wavelet transform invertibility.
e.
| [Prev in Thread] | Current Thread | [Next in Thread] |