2. misc-audio
  3. 1257469878

How do you judge your system's neutrality?


Here’s an answer I’ve been kicking around: Your system is becoming more neutral whenever you change a system element (component, cable, room treatment, etc.) and you get the following results:

(1) Individual pieces of music sound more unique.
(2) Your music collection sounds more diverse.

This theory occurred to me one day when I changed amps and noticed that the timbres of instruments were suddenly more distinct from one another. With the old amp, all instruments seemed to have a common harmonic element (the signature of the amp?!). With the new amp, individual instrument timbres sounded more unique and the range of instrument timbres sounded more diverse. I went on to notice that whole songs (and even whole albums) sounded more unique, and that my music collection, taken as a whole, sounded more diverse.

That led me to the following idea: If, after changing a system element, (1) individual pieces of music sound more unique, and (2) your music collection sounds more diverse, then your system is contributing less of its own signature to the music. And less signature means more neutral.

Thoughts?

P.S. This is only a way of judging the relative neutrality of a system. Judging the absolute neutrality of a system is a philosophical question for another day.

P.P.S. I don’t believe a system’s signature can be reduced to zero. But it doesn’t follow from that that differences in neutrality do not exist.

P.P.P.S. I’m not suggesting that neutrality is the most important goal in building an audio system, but in my experience, the changes that have resulted in greater neutrality (using the standard above) have also been the changes that resulted in more musical enjoyment.
bryoncunningham
bryoncunningham OP
893 posts
 
Tonywinsc - No, you didn't say anything to spoil the party. You have made several valuable observations, though some of your objections to neutrality have been discussed at length earlier in this thread. For example, you wrote:

Music is coloration- ie. a blend of sounds and overtones and the nuances of tonal responses. Every instrument, violin, piano, ect. has its own unique coloration as it were.

There has been much discussion on this thread about the distinction between coloration in music and recording, on the one hand, and coloration in playback systems, on the other. The consensus among Objectivists and Subjectivists alike is that coloration is an essential part of music and recording. The disagreement lies in whether coloration can be reduced in a playback system, whether it is desirable to do so, and if so, how. You may find reading the rest of the thread of use, and hopefully fun as well.

I, for one, welcome your contributions.

Bryon
tonywinsc
583 posts
 
I saw that in the posts- as independent researchers many of us come to similar conclusions. Perhaps we are on to something.
cbw723
89 posts
 
If one were to wear yellow glasses while skiing during an overcast day, visual improvement in the snow's light and dark shadow detail would be apparent. Those same glasses on a bright day would not be beneficial...
The improvements in your system may have actually increased the level of contrast above and beyond the original instruments of the musician.


Here, Hamburg is challenging the idea that items (1) and (2) are indices of neutrality. I thought this to be one of the more effective and relevant challenges to my original post, but no one seemed to run with it.

I think he has a point, and that point raises an existential question.

Let's look at the continuum of neutrality as defined in this thread:

At one end of the spectrum, you have a system that plays back, say, a 1kHz tone, no matter what the source. This is the anti-neutral system: everything sounds exactly the same.

At the other end of the spectrum, consider a hypothetical system that processes the source, and through the use of pattern recognition, seeded pseudo-random number generators, and large variety of sampled sounds, effectively replaces the source with something else. One violin might sound like a subway train, another, slightly different violin might sound like a jackhammer, a cello sounds like a babbling brook shifted one octave up and slowed down by 20%. So we satisfy criterion #1: different instruments sound more different. (They just sound nothing like what they really are.)

Similarly, using the same system, we look at the first n bits of any recording (where n is large enough to insure uniqueness over the body of recorded music) and use those bits to seed our random numbers to insure that each recording sounds completely different from all of the others. So now we've satisfied criterion #2 of neutrality: any music collection sounds more diverse.

This absurd system would be, by our operation of the term, more neutral than anything any of us currently has. But I don't think it would lead to improved musical enjoyment. So clearly, within the idea of neutrality we are making assumptions about truthfulness to the source, and consistency of playback.

Which brings me back to Hamburg's point.

If we consider a system that smooths out recording artifacts, we also risk smoothing out sounds that are real features of the music, making the system less truthful (i.e., it suppresses real contrast). In Hamburg's example, the divergence from the truth is the unwarranted exaggeration of contrast. While neutrality, as operationalized here, resists the suppression of contrast, it doesn't appear to resist its exaggeration.

So somewhere in the operation of "neutrality" there is a necessary condition that the playback system maintain truthfulness to some reference point or points. How, exactly, one codes that constraint, I don't know. But if it were coded, would it, in and of itself, be a sufficient condition for neutrality?
almarg's avatar
almarg
9,670 posts
 
Cbw, I think that the scenarios you have cited pinpoint some important points that underly some of the disagreements which have been rampant in this thread.

In my two posts dated 11/20, one of the things that I tried to express, but perhaps didn't as explicitly as I should have, is that if throughout this thread the word "accuracy" had been substituted for the word "neutrality," the amount of controversy and disagreement might have been significantly less.

To me those two terms, in the context of an audio system, mean essentially the same thing. But I'm not sure that they have been interpreted in the same sense by some of the others.

To a first approximation, what goes into the system and what comes out of the system should resemble each other as accurately as possible. Of course, there are then the obvious issues, that have been gone over multiple times in this thread, about not being able to know exactly what is going into the system, about euphonic inaccuracies resulting in sound that is subjectively more pleasing, etc., etc.

But as I see it, those issues, while often being highly significant, are second order effects. And if the word "neutrality" were understood to mean the same thing as "accuracy," which is how I and I believe Bryon and some others have been using the term, I think we would have seen a lesser degree of divergence in the viewpoints of many of the protagonists in this thread.

And by a similar substitution of terms, I believe that the conundrums which you have cleverly posed in your previous thread would be largely reconciled:
Neutrality, as operationalized here, resists the suppression of contrast, it doesn't appear to resist its exaggeration.... So somewhere in the operation of "neutrality" there is a necessary condition that the playback system maintain truthfulness to some reference point or points. How, exactly, one codes that constraint, I don't know. But if it were coded, would it, in and of itself, be a sufficient condition for neutrality?
Substitute the word "accuracy" for "neutrality" in this paragraph, and it seems to that, to the extent that it is practicable to judge accuracy, we resist both the suppression and the exaggeration of contrast.

And if we impose the constraint of truthfulness to some reference point, presumably the listener's prior exposure to live music, while we by no means obtain any certainty of optimal results (either objectively or subjectively), if we interpret "neutrality" in the sense of "accuracy," then I submit that typically there will be a considerable (and useful) degree of correlation, albeit a partial correlation, between the ability of a system to make different records sound different, and the likelihood of obtaining those optimal results.

On another note, happy Thanksgiving to all!

-- Al
learsfool
1,215 posts
 
Cbw, loved your post! Almarg, I have always enjoyed your interesting and informative posts as well. In this case, however, although I see what you are driving at, I am not sure that the substitution of "accuracy" for "neutrality" really changes anything. Those who buy into the concept of neutrality will naturally make this equation of terms; however, those who do not would just as naturally not equate accuracy with it.

A designer of a piece of audio equipment I think certainly would attempt to be truthful to his reference point of live music (and actually, I would not consider that a "constraint" - that's an interesting choice of word - I prefer to think of it as the designer being "free" to try to create the sound he wants, and I am pretty sure the vast majority of designers would feel that way - many of them I think correctly consider themselves artists), but this reference point will still be different for every designer and for every listener. What sounds accurate to one will not to another. I usually ignore the term "accuracy" when encountering it in reviews, etc., unless it refers specifically to pitch accuracy, for example in reference to a turntable's speed accuracy.
More to discover