comp.lang.idl-pvwave archive: archive » Re: Savgol

Home » Public Forums » archive » Re: Savgol - Differentiation problems

Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend

Re: Savgol - Differentiation problems [message #37969]

Fri, 13 February 2004 01:01

richard.martin
Messages: 3
Registered: February 2004

Junior Member

Worked it out now - a real IDL gotcha this one;

SAVGOL essentially fits a polynomial around a local group of data points,
which can then be differentiated to give numerical derivatives. The SAVGOL
function returns a set of filter coefficients based on the leading term of
the polynomial.

However, when a higher order derivative is specified, the parameter
instructs the routine to return the nth coefficient of the polynomial NOT
the leading coefficient of the polynomial differentiated n times. Of course
this is the same thing, but different by a factor of n!. Of course IDL
completely fail to mention this in their documentation. Along with the need
to divide by the stepsize^n.

In other words, nth derivative output from the SAVGOL/CONVOL combination
needs to multiplied by n!/((x[1]-x[0])^n), where x is the abscissa. You do
sometimes wonder whether the IDL people have read the text in NR at all.

Dave

Report message to a moderator

[Message index]

Re: Savgol - Differentiation problems

By: richard.martin on Fri, 13 February 2004 01:01

Previous Topic:	Re: Working In 3-Space
Next Topic:	Re: Looking for Mittag-Leffler function

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Fri Oct 10 14:28:48 PDT 2025

Total time taken to generate the page: 0.70349 seconds