[ARSCLIST] A fundamental Flaw: Was Sampling Theory (was Fred Layn's post on the Studer list re: Quantegy)

[James Lindner <jim@xxxxxxxxxxxxxxxxx>]

A couple of months ago I was at a series of live concerts and I started
wondering if there has been a fundamental flaw in the way we capture sound -
either analog or digital. When I say fundamental I mean fundamental. I am
inspired by the comment a few emails ago about the sound of a live band or
double-reed. Like all of you I have spent many years with sound - both as a
professional and as an individual - and I have to say - that looking from
50,000 feet - that it feels to me that there are fundamental errors in the
way we record and reproduce sound. I too have heard many, many different
speakers and microphones - but really do any of them really reproduce
ACCURATELY what we are hearing? I really don't think so. Why? I am not sure
- but I have some ideas, or really some questions - and I am not sure that
the fundamental questions have really been asked in a VERY long time.

For instance - is it only about sound pressure? Is it only about frequency
response? Is it only about phase cancellation? If so - lets be honest - we
should have had it licked a LONG time ago. But we haven't. We are not
capturing presence - or at least much of it. Multi channel systems come
closer - but still there are fundamental issues. I was thinking about a
system that used a net of microphones in an area and a signaling system in
the network that somehow took into account temporal delay at different
frequencies reaching different "microphones" (I think they would be
fundamentally different then any mic. made today)at different times. The
reproduction system would have to be similarly different. I envision
actually 2 different types of signals being captured. The audio pressure
information that we now capture - but also the temporal delay and bounce
across the spectrum - information that we don't capture at all other then
capturing the RESULT as sound pressure, after it is all mooshed together -
when it is too late. The reproduction would be changed based on the
information contained in the temporal data stream. There would need to be a
similar net of speakers for reproduction. I am not sure that this is it
entirely either - because no acoustic environment is identical - so you
would have to take into account the acoustic environment on capture as well
as reproduction. Reproduce the result in 3-d space of the acoustic objects
in the original space. Kind of like an acoustic virtual reality recreation.

It occurs to me that sharing these musings in a public list may not be the
best place - perhaps over a beer or two would be better in a more private
forum, these ideas may seem like craziness to some of you. Perhaps I am on
the wrong road entirely - nevertheless I have to ask myself and ask all of
you - an exceptional collection of knowledgeable people - the fundamental
question - have we somewhere made a BIG mistake and missed something?

I totally agree with many of the comments - while we can debate Analog and
Digital issues - they are relatively subtle as compared to the BIG issue
which is - why does sound recording (analog or digital) sound so bad as
compared to the live performance? This is not subtle - it is obvious. It
just isn't close. For me ribbon speakers come the closest in terms of
reproducing that sense - but they fail too. What have we missed? Where did
we go wrong? It seems to me that all our engineering went in one direction
and no one ever looked back. Well I am looking back - and wondering. Where
did we go wrong? If the goal was really to reproduce the sound of live
performance - then clearly we have not succeeded. Something is clearly
missing. What?? And how do we fix it?

jim

*
        Jim Lindner
*
        Media Matters, LLC
*
        Email: jim@xxxxxxxxxxxxxxxxx
*
        Address: 500 West 37th Street, 1st FL
        New York, N.Y. 10018
*
        eFax (646) 349-4475
*
        Mobile: (917) 945-2662
*
        www.media-matters.net



-----Original Message-----
From: Association for Recorded Sound Discussion List
[mailto:ARSCLIST@xxxxxxx] On Behalf Of Mike Richter
Sent: Tuesday, January 18, 2005 1:58 AM
To: ARSCLIST@xxxxxxxxxxxx
Subject: Re: [ARSCLIST] Sampling Theory (was Fred Layn's post on the Studer
list re: Quantegy)

At 09:58 PM 1/17/2005 -0800, Rod Stephens wrote:
>I think the basic problem is that we have no "perfect" listener, no
>matter how "golden" his or her ears are.  Everybody hears things from
>their subjective standpoint (or sitpoint).

I'm sorry, but I believe you are overconstraining the problem. I suggest
that determining which is better is a secondary issue. What we need to
determine first is whether they are perceived as different. If a system
with enough frequency response can be invoked, then any listener can be
asked to determine same / different. (Well, I suppose it takes someone at
least up to Sesame Street level - they have a program or two on the
subject.) Some may be insensitive to the differences (if any); others may
be too discriminating; but my guess is that there will be a substantial
consensus.

If we find that they cannot be made to sound the same, then the question of
which is more accurate becomes relevant. For that, 'golden ears' or
objective measurement will be needed. But why not try taking an easy first
step first? It may turn out that we don't need a second.


Mike
--
mrichter@xxxxxxx
http://www.mrichter.com/