[music-dsp] Newbie question about frequency analysis

Darrell Burgan darrell at palancar.net
Tue Nov 20 23:44:42 EST 2001

Sorry if this is a dumb question, but I will be the first to admit I'm not
exactly educated in this subject.  :-)   Anyway, I'm considering writing a
VST plugin for which the first stage is to split the incoming signal into
three frequency bands, roughly the low, midrange and high frequency portions
of the signal.  I'll then process each of these three bands separately,
doing different things to each, then mix them back together again in the
final stage.

Now, I understand the difference between time and frequency domains, and the
need to do FFT of some sort to translate between them.  But is FFT really
needed for this specific application?

What I'm considering doing instead is this (in real time):

step 1:  make copies of the input signal such that I have three exact input
streams, one sample at a time.  call them streams L, M and H

step 2: put stream L through a 6-pole lowpass filter that cuts off
everything above frequency X;  put stream H through a 6-pole highpass filter
that chops off everything below frequency Y; and put stream M through a
6-pole bandpass filter that chops everything above Y and below X.  At this
point, I should have a pretty well separate trio of streams...??

step 3:  do my processing on each stream as needed

step 4: mix them back again

This intuitively seems to make sense, although I'm a bit concerned about the
fact that even with 6-pole filters there will be a small amount of overlap
between the three streams.  What's the fly in the ointment with this
approach??  Am I just completely out in left field on this???

Thanks very much for any advice!!!

Best regards,
darrell at palancar.net

dupswapdrop -- the music-dsp mailing list and website: subscription info,
FAQ, source code archive, list archive, book reviews, dsp links

More information about the music-dsp mailing list