On 8/27/2010 12:02 PM, Orchid XP v8 wrote:
> Mike Raiford wrote:
>> Sooooo ... I wanted to try to make an FFT filter.
>
> You're aware that you can perform linear filtering operations without
> using an FFT at all, right?
>

Well, yeah... of course. There's dozens of polynomials out there to do 
it, too. But try designing an arbitrary filter with that. :D

The kicker is, all of those filter functions can be built using 
capacitors, resistors and/or inductors.

> An FFT convolution is faster when the filter kernel is large, however.

Indeed. But using FFT makes filter design easier.

... And using FFT also makes doing a vocoder rather simple.

> I sat here and thought about posting the 8 lines or so of Haskell it
> would take to do the same task. And then I realised that Warp would
> complain that it's not processing the data in-place, so it has different
> performance characteristics, and then I'd have to build both your
> version and my version and do detailed timing statistics, which would
> require me to install the MS .NET runtime and some sort of C#
> compiler... Nah. I have better things to do with my Friday afternoon. ;-)


Thbbbbbbt.

-- 
~Mike

Post a reply to this message