|
From: | Eric Chassande-Mottin |
Subject: | Re: [Tftb-help] Problem with tfrscalo command |
Date: | Mon, 19 Nov 2012 17:01:20 +0100 (CET) |
User-agent: | Alpine 2.02 (DEB 1266 2009-07-14) |
The algorithm It's really good for analizing data and the time-frequency plot It's really interesting, but when I try to analize a signal with 262144 samples (2^18), don't work the algortihm because need much memory. Is there an alternative for up the speed and work propertly the algorithm for this size of samples ?
Yes, you can try tfrrmsc() which computes the Morlet Scalogram (alongwith its reassigned version) using a different computing scheme than tfrscalo.m that may work better with long duration signals. If you are not interested in reassignment, you may edit the m-file a bit to remove its computation from tfrrmsc.m. Note that the frequency/scale grid used in tfrrmsc() is different and that you have to use Morlet wavelet. HTH, Eric
[Prev in Thread] | Current Thread | [Next in Thread] |