[news]

"KH" wrote in message
...

Yes, and the OP made no claims whatsoever that current recording
techniques *do* capture sufficient information for a perfect playback. To
the contrary, I read the question as how many channels, recording and
playback, would be required to make it perfect.

How do you construe such a question as a claim that current recording
technology is perfect?

Keith

Keith,

I think that once again there is a communication gap here. I take the OP's
question as how many channels do we need to fix it? Maybe he realizes that
we are not going to get there with two channel because that just addresses
the front sound, so the next logical step is to increase the channels until
it is just about there. Scott and I answered no we can't get there from here
for similar enough reasons. The way I state it is that to do stereo, a
loudspeaker based system, you have to run the sound through two different
rooms before you hear it. Still some agreement among all of us, generally,
right?

OK now, where we begin to diverge is that Scott, and maybe you, and a lot of
others want to solve this problem by decreasing the room sound as much as
feasible without turning it into an anechoic chamber. I disagree because we
need the room to anchor and localize and externalize the sound in a real,
physical space. I go even further to say that this is the whole idea of the
stereophonic system, to reconstruct the sound that was recorded in another
room - your room.

THAT is the concept that we will violently disagree on. You try to tell me
that you can't do that - it will make everything sound like your room, and
that isn't accurate, isn't the idea. I point out that if you do it right,
the speakers will disappear as the apparent source of the music, and the
room will disappear as a "nuisance variable" as Floyd Toole puts it, that is
even noticeable on playback. The room simply does not have much of a
reverberant field, and so what you hear is what is contained in the
recording - if you get the spatial part right.

By this time your head is exploding, so I attempt to erase all knowledge
(and misconceptions) about stereo and such and start from scratch.

The recording does not, and cannot, contain an exact picture of the original
performance, for the basic reason stated above, nor is it intended to, due
to the nature of the production of the recording. This recording is a new
work of art, based on the live event, that we will use to reconstruct a
realistic representation of the original within our room. This new work may
be an attempt at the realistic reproduction of the original event, or it may
be a totally new work concocted in the studio and/or recorded with multiple
tracks and built up to taste. Doesn't matter that much, what we need to do
is not try to place the sound directly into our ears but rather use it to
reconstruct the sound at home. Kind of like rolling in that player piano
and placing it properly so that it will have good sound in its new
performance.

"But that's not accurate, wah wah wah." We are not doing accuracy, we are
doing realism. Which leads me to the most controversial part.

I have been lectured that this is a two dimensional process, that we simply
cannot record all of the little sound patterns that existed in the live
performance, with the early reflections off the walls of the concert hall,
the ceiling and floor and all of the decorations and diffusers and on and
on, so all that remains is some "illusion" of what might have been. Let's
characterize these myriad reflections as first you get the first arrival
sound and then afterward comes "zing, ping, pow" the myriad reflections from
all directions. But we cannot record zing ping pow with our crummy system,
so they are gone - vanished. Right?

Here we go on final approach.

For simplicity sake, we record a piano trio - piano, bass, and drums. We
close mike them to get a crisp, tight recording. Back at studio, mix the
piano into center, bass at left, drums at right (or anywhere within the
stereo field due to the happy factor of summing localization being able to
place sounds anywhere between). We then play back on speakers placed in
positions that are geometrically similar to the instruments and microphones,
somewhere away from all walls in positions that you might place them if they
were there live.

On playback, first arrival sounds come from appropriate locations - AND -
because of this similar positioning and radiation patterns of the speakers -
what do you know - zing ping pow happening all over again right in your
room! The drum sound will reflect off the front and right side walls, the
bass off the left walls and the piano will remain mainly centered.

With larger orchestras and wetter recordings you can think of the process as
"close miking the soundstage" in a way that records the instruments AND
their early reflected and a tad of the reverberant sound in the live space.

SUMMARY: We couldn't record each and every zing ping pow during the session,
but we really did capture the essence of it enough to get it spatially
correct on playback and have it arrive at the audience from incident angles
that are very similar to the live situation and for the same reasons. So no,
they are not right in the two channels if you just kill all room sound and
attempt to listen near field or something, but if we reconstruct the main
spatial positioning of the speakers and let the room work for us rather than
against us, it will sound very realistic and even bring most of the original
ambience back alive - if you understand how to decode it with the physics of
the playback situation.

In other words, the path to glory lies not in increasing channels and
smothering the room sound, but rather understanding the system and shaping
the playback sound fields just right to mimic most of the characteristics of
the typical live sound fields.

Sorry once again to be so wordy and lectury, but this justs scratches the
surface of how to do stereo right, which I call Image Model Theory.

Gary Eickmeier