differentiating evently sampled data
- To: mathgroup at smc.vnet.net
- Subject: [mg41108] differentiating evently sampled data
- From: Stepan Yakovenko <yakovenko at ngs.ru>
- Date: Sat, 3 May 2003 03:28:21 -0400 (EDT)
- Reply-to: Stepan Yakovenko <yakovenko at ngs.ru>
- Sender: owner-wri-mathgroup at wolfram.com
HI! It's possible to differentiate evently spaced data using DFT (Discrete Fourier Transform). First of all, we have to make data periodical. The simpliest way is to work with g=f+Ax, where A is chosen so that g[x_min] = g[x_max]. And g'=f'+A. Then \partial_x g = InverseFourierTransform[-i*k*InverseFourierTransform[g]]. I'm going to use this for my own needs and I can tell you about the results when it'll be ready (if some one is interested, of course :)). One of the merits of this method is that one can filter out high frequences on the way while working with a noisy data. -- Best regards, Stepan mailto:yakovenko at ngs.ru