Quote from: knutinh on 2010-05-25 07:32:43

For the general case of reproducing a soundfield from a true (or virtual) acoustic event, I would claim that we are not even close.
*Stereo recording techniques and playback systems has enormous uncertainties.
*Surround is basically a refinement of stereo, but still many of the same drawbacks
*Binaural recordings are probably the closest, but still limited by lack of individual HRTFs, non-standardized headphones, lack of head-movement.

Wavefield synthesis seems like a promising thing, but it is my impression that it is immature, prohibitly expensive, and we need to see major theoretical advances (as well as production advances) before it is ready for primetime.

The problem you are talking about is not a problem with the transparency  of audio components but rather inherent problems with  doing an adequate job of capturing and reproducing sound fields.  We don't yet need a new recording format because we don't know what to put on it. The unique audible character of live music is lost by the time the audio signal gets to the output terminals of the microphone.

The movie industry still had places to go (the jump from PAL/NTSC to 1080p is amazing) so it's easy to convince people to buy new technology. The audio industry has already reached high end in the 70s with the AudioCD format: once you reach transparency, what more can you do?

Another open question is if a more accurate (physical) reproduction of the acoustical event will increase our subjective impression of realism.

There seems to be some evidence that this isn't necessarily the case.

I'm always impressed by the effectiveness of foley artists who can simulate noises for feature films that IMO are far from reality, but are nevertheless very convincing. Fashion might play a role as well.

Quote from: Arnold B. Krueger on 2010-05-26 12:09:27

The problem you are talking about is not a problem with the transparency  of audio components but rather inherent problems with  doing an adequate job of capturing and reproducing sound fields.  We don't yet need a new recording format because we don't know what to put on it. The unique audible character of live music is lost by the time the audio signal gets to the output terminals of the microphone.

I agree with what you are saying, but I was responding to a claim that sound-reproduction was "transparent". My gripe was that it is probably transparent in the sense of terminal of the microphone to terminal of the loudspeaker. In terms of reproducing a musical performance to a listener, it is usually not. In other words, there is still room for audio reproduction to make large subjective improvements, just like the video industry. How to achieve those improvements (and if the customers really want them enough to pay for it) is an open question.

What are those?

Most relevant to the general notion of sound quality is the observed gap between ‘‘objective’’ physical accuracy and ‘‘subjective’’ perceived naturalness. Indeed for most auditory scenes, the 2-D configurations were judged by the participants as more natural and realistic than the 3-D configurations although spatially incomplete, thus indicating potentially negative effects linked to providing ‘‘too much’’ information. These findings underline the difference between illusion and accuracy pointed out by Rumsey (2002): the illusion of ‘‘being there’’ is not necessarily related to true spatial fidelity. This counter-intuitive observation, from a physical point of view, indicates the importance of considering subjective psychological attributes in the evaluation of perceived sound quality. Furthermore, the lack of preference for 3-D configurations could be explained by the unfamiliarity with 3-D audio reproduction, although the natural world is always present in 3D. As surround sound systems become more common, 2-D audio reproduction systems may sound more familiar and thus more ‘‘natural’’ than 3-D configurations, which are not widely used. The results reported here suggest a shift from physical descriptions to cognitive ones in exploratory studies, to identify relevant perceptual features and better understand how acoustic phenomena are perceived and cognitively processed before addressing physical parameters in more controlled experiments.

If a listener can be seated in a concert hall, blindfolded, and be served A)A real orchestra playing, B)Playback of a recording of the same, and not reliably be able to distinguish them, it would seem that we are on to a perceptually transparent sound reproduction system (recreate all perceptually relevant information in an acoustic event, remove all perceptually relevant coloration in a playback location).

Quote from: knutinh on 2010-05-27 09:00:59

If a listener can be seated in a concert hall, blindfolded, and be served A)A real orchestra playing, B)Playback of a recording of the same, and not reliably be able to distinguish them, it would seem that we are on to a perceptually transparent sound reproduction system (recreate all perceptually relevant information in an acoustic event, remove all perceptually relevant coloration in a playback location).

Recalling that for most recordings that people buy, the "concert reality" you are trying to emulate is the band's stack of speakers on the stage, even for acoustic instruments.

If a listener can be seated in a concert hall, blindfolded, and be served A)A real orchestra playing, B)Playback of a recording of the same, and not reliably be able to distinguish them, it would seem that we are on to a perceptually transparent sound reproduction system (recreate all perceptually relevant information in an acoustic event, remove all perceptually relevant coloration in a playback location).

Quote from: knutinh on 2010-05-27 09:00:59

What are those?

I expected that question but didn't have time to look up the sources right away
Here is a quote from Perceptual evaluation of multi-dimensional spatial audio reproduction by Guastavino and Katz, JASA 116(2) 2004:

Quote

Most relevant to the general notion of sound quality is the observed gap between ‘‘objective’’ physical accuracy and ‘‘subjective’’ perceived naturalness. Indeed for most auditory scenes, the 2-D configurations were judged by the participants as more natural and realistic than the 3-D configurations although spatially incomplete, thus indicating potentially negative effects linked to providing ‘‘too much’’ information. These findings underline the difference between illusion and accuracy pointed out by Rumsey (2002): the illusion of ‘‘being there’’ is not necessarily related to true spatial fidelity. This counter-intuitive observation, from a physical point of view, indicates the importance of considering subjective psychological attributes in the evaluation of perceived sound quality.

Here's a real mouthful:

"..the importance of considering subjective psychological attributes in the evaluation of perceived sound quality."

Here is a passage from Perlman (2004):

"I ask Jack if he has a CD player as well as the turntable, and he says no.
There's a debate whether tubes are better than solid-state, he tells me, but
there's no debate whether analog is better than digital: it is. A cheap CD
will beat a cheap turntable, but even a moderate turntable will blow away
any CD. CDs are strident and flat; Jack can't listen to one for more than
15 minutes at a time. On a CD you can hear the pluck of a double bass,
but on an LP [long-playing record] you can hear the body of the
instrument, the wood. "

The audiophile's rich sentimental memories of listening to LPs in former (perhaps better) days conditions form many of the "...psychological attributes in the evaluation of perceived sound quality."

Quote

Furthermore, the lack of preference for 3-D configurations could be explained by the unfamiliarity with 3-D audio reproduction, although the natural world is always present in 3D. As surround sound systems become more common, 2-D audio reproduction systems may sound more familiar and thus more ‘‘natural’’ than 3-D configurations, which are not widely used. The results reported here suggest a shift from physical descriptions to cognitive ones in exploratory studies, to identify relevant perceptual features and better understand how acoustic phenomena are perceived and cognitively processed before addressing physical parameters in more controlled experiments.[/size]

IME there was a lot more bad-mouthing then there was actually bad equipment. Case in point would be the CDP 101. It wasn't precisely sonically perfect, but it had spectacular performance compared to the best day of vinyl and analog tape's now-obsolete lives.

Quote from: knutinh on 2010-05-27 09:00:59

If a listener can be seated in a concert hall, blindfolded, and be served A)A real orchestra playing, B)Playback of a recording of the same, and not reliably be able to distinguish them, it would seem that we are on to a perceptually transparent sound reproduction system (recreate all perceptually relevant information in an acoustic event, remove all perceptually relevant coloration in a playback location).

I suspect that it may well be possible to approach this using current technology. The overwhelming influence on the soundfield in this scenario is the hall itself. So if you record the orchestra "dry" and play it back in the hall, the influence of the hall will still be felt.

But that's not really the goal, is it? The actual goal is for the listener to be seated *in his own living room* and experience a sound that is like sitting in that concert hall with a real orchestra playing. And as far as I can tell, we are a long way from that, aren't we?

One of the problems with comparing 2-D sound to 3-D sound is that it is immediately clear to the listener which he is listening to. This is just like listening to LPs, the audible artifacts are unmistakable. There is no such thing as truely blind comparison between 2-D and 3-D since they identify themselves by means of their sound quality, aside from any issues of realism.

Quote from: Arnold B. Krueger on 2010-05-26 19:58:01

IME there was a lot more bad-mouthing then there was actually bad equipment. Case in point would be the CDP 101. It wasn't precisely sonically perfect, but it had spectacular performance compared to the best day of vinyl and analog tape's now-obsolete lives.

On what do you base your spectacular performance claims?

There are, of course, people who believe that analog equipment is what offers spectacular performance.

In Mobile Age, Sound Quality Steps Back

My position is that early digital audio was imperfect. Analog audio is imperfect.

How do you decide which sounds better?

I thought you were in agreement that the specs and measurements don't always accurately predict how something will sound.

Not all distortions are created equal.

Most listeners would take a bucket full of 2nd harmonic distortion over a thimble full of aliasing.

I think the situation changes when you reach transparency. With ABX you can demonstrate that a digital reproduction is indistinguishable from the original and, in theory, quickly deflect any bad-mouthing.

I think digital audio turned people off early in its history because it sounded different than analog,

because it was a less forgiving recording medium

and because analog distortions had become euphonic

whereas digital distortions were unfamiliar

It looks like you're sticking to your story that digital audio was transparent from its commercial inception.

Since I did not have opportunity to work will all systems back then, I will accept the possibility that the best systems of the day were transparent.

I really can't accept the proposition that ALL digital audio back in the 1980's was transparent.