Re: The Who Mailing List Digest V3 #176

> From: mleaman@sccoast.net (Mark Leaman)
> Date: Mon, 17 Jun 1996 16:03:10 -0400
> Subject: Re: Quick One Stereo vs. Mono Debate
> 
> >Actually audio technology is not to the point where one can take a completed
> >stereo mix and 'unmix' this into a 'new' multitrack.  We're just not
> >there...yet.
> 
> I read an article about this where it was claimed that such a thing could be
> done...but I bow to your superior knowledge. Could you not separate the
> frequencies, which would I know be a long and delicate task, yet would yield
> the same result?

Apologies for the non-Who content.   I'd love to read that article if you
can find it.  Theoretically, I suppose it could be done, however, you'd need
a whole lot of DSP (Digital Signal Processor) power to do this, and it will
probably require some sort of artificial intelligence model to separate
those frequencies.  Right now it's easier to separate noise from a given
signal, than two deliberate signals.  I'd guess that we're at lease a half a
decade from this, based on what I know about DSP, and what I can guess about
AI.

To throw a Who--> spin on all this, someday we may be able to take not only
our favorite performances by the band as a whole but put together something
like this:

Song: Won't Get Taken for a Ride
Vocal: Roger Aug. 12, 1969
Guitar: Pete Feb. 17, 1978
Bass: John May 10 1963
Drums: Keith Dec 14 1945
Venue: Joe's Bar and Grill summer 1967 (always a lively crowd)

Morph all these favorite individual performances into a "perfect"
performance, even though it never "happened".

Some of that is possible now, but it's not as easy as it sounds.  Not easy
at all.

Take care,
Christopher