neutrality vs. realism

Ernie (Subaruguru) and Muralman, thnak you, but you should give credit to Asa for the catalyzation idea as well as the notion of sufficency.

Ernie, you've really opened a can of worms with the first arrival/second arrival problem. From a designer acquaintance (a scientific objectivist, Asa): "we don't have complete accuracy available as of yet. We may have accuracy on one dimension, such as axial frequency response, but not have accuracy in another dimension, such as sound field arrival vector/intensity accuracy." Moreover, "angle of arrival has much to do with perceived tonal balance. While the data can be gathered there is no established perceptual index for percentage of program intensity per program angle of arrival, even though it plays a very significant role in perceived accuracy. This of course includes boundary effects, timing/phase cues as well as angle of incidence. The perceived summation is quite complex." (These are not my ideas, so don't give me credit for them.)

So, Asa, I think that frequency response is all important in accurately reproducing the recorded event. The caveats are that (1) a perfectly accurate on-axis frequency response may not result in accuracy at your ears in your listening room, and (2) you may find a dollop of sweetness or a rounded edge or a little tweaking with presence for soundstaging effects more pleasing to your ears. (I might too.)

There are lots of other things to measure besides on-axis response; dispersion, off-axis response of speakers, cabinet resonance in speakers, jitter in cdps, amplifier performance into real loads, etc. I think everything that can be heard can be measured. The trick is using the measurements to predict what the listener will experience in his chair in his room.

Asa, the important issues you bring up deserve more thought and discussion, especially considering the secondary (and tertiary, etc.) arrival problems Ernie raises. Is it possible that there is a level of sufficiency that is preferable to complete accuracy in the real world, in order to minimize some of the problems with secondary arrivals? Is it possible that too much auditory information may be a detriment?

Of course, I think I have said, or at least implied, that I really do prefer a little built in loudness compensation on the bottom for warmth and a little (not much, just a tad) of roll-off at the top. My favorite speakers also employ a BBC or Grundee dip centered around 3khz to move centered vocalists back just a bit in the soundstage. I think these deviations from a flat frequency response yield greater realism, that is they conform to what I hear in real life. But, those are just my preferences.

Paul

Paulwp hits a long one again. Although I do not rely on technical statistics, I must admit I peruse them with some amusement. I mainly trust my ears as I hone in on my dream system. I share in the premise there are wave interractions out there that remain unmeasurable. I do not rely on specs for that reason. I would never buy a component based on it's distortion figures. I doubt there is any experienced audiophile that does.

I understand there are live performances and there are live performances. One that I attended this week consisted of electronically reengineered violin music played through an array of outdoor speakers. It was fun, especially the visuals. Never could I recapture that event. Last night I listened to a group playing piano, cello, and violin doing some really thought provoking Chinese scores. I had a good seat, with no early reverberation arrival problems. It was a good live event to judge by.

I remain pleased with how close my evolving system approaches recreating the live experience. Each incremental step takes my enjoyment closer to the heart of musician.

If I may weigh-in here, I have the notion, that a good part of the reason for the measurement aspect of stereo gear is the fact that all components are designed to be interchangeable. Mix and match, if you will. Since the other gear that may be used with any one component is not known my the mfr., an attempt is made to have a "benchmark". This "benchmark" has resulted in a relatively arbitrary, but well meaning, set of measured standards that are meant to assure the user of conformity and usefulness with other units measured by the same standards. So, standards of measurement were adopted for various aspects of performance such as frequency response, total harmonic distortion, power output, etc.

The problem with this is that these processes assume a given idea that flat frequency response or low THD will allow accurate pass-thru to the next component with flat frequency response or low THD. While this may superficially seem to be the case, it is not. Little or no regard is given to the additive or interactive effects of the other components in the chain because the designer is unaware of what the other components will be. So this leads to the designing of equipment in a "vacuum", so to speak. Only the most basic "standards" like output level and impedance levels under static conditions are even considered. All of the other issues are left to the consumer to determine, regarding which items may work well together. And the consumer is ill-equipped to make these decisions, because there is little, if any, data provided for this, and most consumers would not know how to use such data anyway.

This is what leads to the mysterious "synergy" discussions, and the apparent "disconnect" between measurements and sound quality. It is not that the measurements are bad, it is the basic idea of what should be measured, and how to measure it,and provide useful data, that is at the root of the problem.

If a consumer knew that his amp exhibited a THD profile of primarily even-ordered harmonic distortion with a major part at the 2nd harmonic, he could choose a speaker that also had a major part of its THD in the 2nd harmonic,and wire them 180 degrees out of phase, thereby cancelling a significant portion of the distortions of the amp/speaker combination. But, nobody seems to be aware of this type of "system integration design" and none of the measurements really are geared to help anyone do this. Everybody just wants flat response and low distortion, but all components have some dips or rises, and all have some components of distortion. Failure to correctly match these, and other, characteristics will result in additive distortions or frequency anomalies. The consumer is frequently unaware that he is even making this mistake. Also, there is a big difference in having two-tenths of a percent of distortion at the 2nd harmonic, and having one-tenth of a percent at every harmonic point up the scale. In the first case there is a relatively small amount of distortion at one point on the curve. In the second case, there is a smaller maximum distortion rating, but the distortion is all over the place.

So, to sum up, I do believe, as Paulwp seems to, that things that are heard can be measured. But the things measured, and the ways they are measured, and the applications of those measurements leave much to be desired. Now, add that to the things that we haven't learned how to measure, or even realize that they need to be measured, and we end up with a reliance on a faulty set of measurements, and misunderstanding of same, that cannot accurately be used to select our equipment. Thus, many of us rely on our ears to measure what we do know, the sound.

Ultimately the ear is the judge. But I do not discount that certain measurements can lead one to an informed platform from which to begin auditioning, if one can adequately interpret the data that can be found, and apply it in a meaningful way, resulting in a happy combination of components(synergy). And although I personally am a proponent of listening as the final arbiter, I do use design data and measurement data when I can find ways to apply it.

Regarding the brain's interpretation of the sounds generated by the system, I am not in an informed position to comment on that. But I do find it interesting.

Recording music is a highly skilled craft. A skilled recording engineer will take into account the performer, the instruments, the hall, the recording equipment and make assumptions about the playback equipment in order to make what she considers an accurate representation of the musical performance.

For instance, let's say an unaccompanied vocalist has trouble controlling his dynamics and is prone to a slight sibilance. The engineer might pick a microphone with a recessed upper midrange to combat the sibilance and "ride the fader" during the recordings to keep the dynamic range within the optimum area of the storage medium. Alternatively, the engineer could use a compressor/limiter (essentially an automated volume control) and dump the recording into a computer based editor (think of it as a word processing software for music) and repair individual instances where sibilance is an issue. Neither approach is inherently superior and either method can result in a natural sounding recording. The determining element is not the equipment, but the skill of the engineer.

As audiophiles improve their playback systems they may start to reach a point where their systems are capable of readily revealing the artifacts (edits, aggresive EQ, mismatched reverbs, sibilance, low frequency garbage, tape hiss, instrument bleed-thru, air conditioner noise, etc.) of the recording process. These artifacts are not part of the musical performance and as such can only distract from it. I suspect that a large element of how people react to specific pieces of high end equipment revolves around how the equipment deals with these artifacts. I'm over generalizing, but for unknown reasons some equipment heightens and draws attention to these artifacts, while others expose them, but at the same time don't seem to emphasize them. There's so much that we don't know about reproducing music.

BTW, have you ever noticed how in recording orchestras or other large ensembles that the microphones are never positioned where a listner would normally sit?

Agreed, Onhwy61. Also getting back to Rives' room acoustic post.
Once a stereo has reached "perfection" it's the recording which will make or break the reality. Until recordings are perfect, a perfect stereo will show off the recording's faults. Maybe a more musical stereo is better on bad recordings.
Also the perfect stereo needs the correct room to recreate the acoustics of the original event.
The most realistic playback is when a song is new to me and the ear is fooled - at first, for a little while.
Although once got the volume, reverb, and room acoustics just right and felt like Jimmy Hendrix was right in the room. Never could duplicate that again.