From: Jim Peters (jim_at_uazu.net)
Date: 2002-01-24 10:02:07
Doug Sutherland wrote:
> > A question: would integer filters be useful for embedded stuff
> > (e.g. Doug's wearable computing), or it is okay to use floating
> > point on these processors too ?
>
> I'm using pentium-mmx embedded boards, they are no different than
> desktop PCs, just smaller, so floating point shouldn't be much of
> a problem. But clock speed is 200Mhz, so I guess I will have to
> try and see how well it does. If integers are faster it may be
> desireable.
Thanks for the feedback on this.
Bearing in mind I want to get the entire filterbank and pitch-shifting
code working on my P-133, your 200MHz boards should be no problem. It
looks like I should be able to make it go faster with floating point
maths by using less filters and doing more with each one.
I also have some working code (in SDL) that displays the magnitude/
phase as 60 bands of colours going across the screen a bit like an
oscilloscope, using hue to represent phase and intensity to represent
magnitude.
See the attached JPG for a rough idea. This shows a slowly sweeping
square wave. This is with the integer filters, and I know there are
some problems with this set of filters at the moment. The new FP ones
should give better separation between adjacent bands and clear up some
of the other weirdness.
Once I get these new filters tested and working I'll release something
for people to play with.
Jim
-- Jim Peters (_)/=\~/_(_) Uazú (_) /=\ ~/_ (_) jim@ (_) /=\ ~/_ (_) www. uazu.net (_) ____ /=\ ____ ~/_ ____ (_) uazu.net
This archive was generated by hypermail 2.1.4 : 2002-07-27 12:28:37 BST