On 27 Apr 2007 08:13:37 -0700, Mark wrote:
You can't fix a time issue with a frequency fix. It's like pushing
down a ping pong ball into a bucket of water with your thumb, it just
keeps rolling around it. Someday the minds of dsp will realize this.
Time response and frequency/phase response are just two orthogonal
views of the same thing. For every frequency/phase response, there is
one and only one time response that is associated with it. And for
every time response there is one and only one frequency/phase
response associated with it. They are duals of each other related by
the Fourier transform.
So IN THEORY you CAN equalize almost any frequency/phase / time
response.
The exceptions are those responses that null to zero at some frequency
or those that peak to infinity i.e. oscillate. or those that have a
time response that is longer than the capability of your equalizer.
Other than those exception cases, you CAN IN THEORY equalize any
arbitrary response.
Now IN PRACTICE room equlization does not work because even small
changes of position in a room can make a big change in the response.
Also the equalizer may need to boost the gain by a large amount at
some frequencies leading to dynamic range problems, but these are
PRACTICAL limitations. In THEORY, it can be done.
We have a nomenclature problem. You are absolutely right that you can
neutralize ANY resonance in both frequency and time domain by
inserting its lossy counterpart. But room problems are almost never
anything to do with resonances - they are to do with standing waves
which exist in the spatial domain. So any frequency or time domain
solution you try can only ever work at a single point in that room.
Move away from that spot, and your correction is gone.
d
--
Pearce Consulting
http://www.pearce.uk.com