Jones_r wrote:
It can fudge a couple of specific problems, but mostly it's a disaster

Now I'm certain you haven't heard a digital room correction device in
your life.

Sadly, I have heard an awful lot of them.

Basically it has a lot to do with the fact that it doesn't work. Back
in the seventies, everybody and his brother were heavily EQing rooms, and
invariably it caused a lot more damage than good.

Without AT LEAST an Athlon 800mhz CPU, digital room correction, at the
advance state that I'm talking about, can not be done (at least not in
real time). So, did they have Athlon CPUs back in the 70's ?. Somehow
I don't recall.

Nope, although there were some attempts to deal with impulse response
differences in rooms back then. There was a demo at some AES show using
a GenRad signal processor that... well, it didn't sound good. This was
a dedicated piece of hardware that did convolutions, rather than trying
to do it with a general purpose computer.

What they were doing is to correct the frequency domain (and even in
that case, they didn't have the necessary accuracy or number of
filters. If you need to attenuate a frequency, let's say 435.575, by
-1.75db, then a fixed filter tap of 300hz, isn't going to help at all.

Right, this is the most common way of handling these problems, and
as I pointed out, frequency domain correction doesn't solve time domain
problems.

Also, what happens if you need to correct 15,000 individual
frequencies, in order to get to a flatter frequency response time ?,
did they have eq's back then with such a big number of filter taps ?)
and they didn't touch the time domain, at all. Now you tell me, how
can this work ?, right, it can't.

You can use as many filters as you want, and the response will still be
different a short distance away. You can flatten the response in one
part of the room out as much as you want, and the response a foot away
will still be totally different. Sorry about that.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."