Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!ll-xn!mit-eddie!bloom-beacon!mit-hermes!sundar
From: sundar@mit-hermes.AI.MIT.EDU (Sundar Narasimhan)
Newsgroups: sci.math,comp.sources.wanted
Subject: routines for doing digital filters
Message-ID: <2875@mit-hermes.AI.MIT.EDU>
Date: Wed, 5-Aug-87 08:35:25 EDT
Article-I.D.: mit-herm.2875
Posted: Wed Aug  5 08:35:25 1987
Date-Received: Sat, 8-Aug-87 08:59:47 EDT
Distribution: na
Organization: MIT AI Lab, Cambridge, MA
Lines: 16
Xref: mnetor sci.math:1759 comp.sources.wanted:1841

Hi:
I have run into the following problem and was wondering if anyone 
had a good solution -- I need a fair notch filter at a very low frequency 
(0.4Hz at 12.5ms sampling rate), to be used in the input part of a digital 
controller (which doesnt limit the filter to being causal).
The standard Parks-McClellan algorithm that I use blows up when
the number of coefficients exceeds 1000.

So, does anyone have code (preferably C or Fortran) that does filter design
which doesn't have this limitation? 

-Sundar
p.s. send your responses to sundar@hermes.ai.mit.edu, or
     mit-eddie!mit-hermes!sundar  since I read news only occasionally.