How do you judge your system's neutrality?

A fuller extract of ideas from G. Holt’s audio glossary seems relevant in context:

1. Frequency response: Phase shift and distortion can sound like frequency-response aberrations.

(Note 1: Perception of neutral frequency response may be subjectively similar to perception of low distortion and correct phase. One is therefore be tempted to use the term neutrality in all three contexts.)

2. Balance: The subjective relationship between the relative loudness of the upper and lower halves of the audio spectrum; "tonal balance."

3. Coherent: seamless from top to bottom… no audible evidence of different… colorations in different frequency ranges.

4. Continuity: Uniformity of coloration (across the operating range).

5. Discontinuity: A change of timbre or coloration due to the signal's transition (across the operating range through) dissimilar coloration.

6. Seamless: Having no perceptible discontinuities throughout the audio range.

(Note 2-6: He identifies colorations as shifts in tonality across the frequency range. My previous reference to neutrality as “continuousness of musical expression across frequency range” was an attempt to describe this. I question whether there can really be such a thing as continuous coloration, positing instead that inevitable variations in coloration across frequency range indicate unsolved problems in the playback system. To distinguish problems in playback from problems in recording, the trained listener merely needs to listen to a wide variety of recordings on the same playback system.)

7. Pitch resolution: The clarity with which pitch…is perceived. Poor pitch resolution makes all notes sound similar…

(Note 7: Here he touches both on Bryon’s original notion of coloration as a failure to differentiate, and on Cbw723’s remarks on Shannon entropy.)

8. Fast: Reaction time, which allows a reproducing system to "keep up with" the signal fed to it. (A "fast woofer" would seem to be an oxymoron, but this usage refers to a woofer tuning that does not boom, make the music sound "slow," obscure musical phrasing, or lead to "one-note bass.") Similar to "taut," but referring to the entire audio-frequency range instead of just the bass.

9. Smooth: Not necessarily a positive system attribute if accompanied by a slow, uninvolving character.

10. Control: The extent to which a (system) sounds as if it is "tracking" the signal being fed to it. The sound is tight, detailed, and focused.

(Note 8-10: Here he identifies fast & controlled dynamics as essential to an uncolored presentation. Absent precise dynamics, differentiation is lost (e.g. “one-note bass.”))

11. Definition (also Resolution): That quality of sound reproduction which enables the listener to distinguish between, and follow the melodic lines of, the individual voices or instruments comprising a large performing group.

12. Detail: The subtlest, most delicate parts of the original sound, which are usually the first things lost by imperfect components.

13. Veiled, veiling: Pertaining to a deficiency of detail and focus, due to moderate amounts of distortion, treble-range restriction, or attack rounding.

14. Focus: The enhanced ability to hear the brief moments of silence between the musical impulses in reproduced sound.

15. Hangover: A tendency for reproduced sounds to last longer than they should.

16. Ringing: The audible effect of a resonance: coloration, smear, shrillness, or boominess.

(Note 11-16: Higher resolution defined in this sense a good thing, and is connected by Holt to Bryon’s original notions about the distinction of differences. Interestingly, the related quality of Focus entails removal of additive colorations so as to reveal interstitial silences. Also, Holt connects compromised resolution to “attack rounding”—or failed dynamics.)

17. Error of commission: Signal degradation due to the addition of sounds that were not present in the original signal. Distortion and coloration are examples of errors of commission.

18. Error of omission: Signal degradation due to the loss of information that was present in the original signal. Smearing and treble loss are examples of errors of omission.

19. Gestalt response: The evocation of a complete memory recognition by an incomplete set of sensory cues. A gestalt response to the few things an audio system does outstandingly well can make imperfect reproduction seem more realistic than it actually is.

(Note 17-19: Here he gets into interesting territory similar to Bryon’s ideas about coloration as additions, subtractions, or alterations. I think we may conclude that Holt regarded coloration primarily as an error of commission. The idea about Gestalt response is fascinating insofar as he suggests that absent errors of commission, a less than fully resolving playback system may be convincing and perhaps even uncolored in the strictest sense of the word.)

20. Neutral: Free from coloration.

21. Uncolored: Free from audible colorations.

Finally, Holt used the terms Subjectivism and Objectivism in a particular sense. Subjective reviewing is his term of art for critical observation based on controlled listening, psychoacoustics, and a precise vocabulary to evaluate colorations. He considered Objectivists to be the so-called “meter men” test bench-oriented reviewers of Julian Hersh school who tended to ignore the science of listening. Holt’s Subjectivism is not inconsistent with objective analysis. As he says in the glossary, given a precise definition of terms “there is no longer any excuse for an audio reviewer saying, ‘I can hear a difference, but there's no way of describing it.’ Now, there is a way.”

So Holt's view was that audible colorations can be precisely described, and if inaudible, do not exist.

Learsfool wrote:

This has been discussed in a different thread before, the idea that many audiophiles assume that some "colorations" they are hearing are caused by their system, when in fact they are on the recording itself. I have seen more than one situation where two people could not agree on which was the case ("Well in MY system, it doesn't sound like that!" etc). This is yet another reason why I don't think there could ever be much agreement on any two people's sense of "neutrality."

Like many people, I have had this experience both with audio and video. While in video, a proper calibration usually clears up the ambiguity between software colorations and hardware colorations, I agree that, in audio, it's not so simple. But again, I'm more optimistic about the possibility of some agreement some of the time. It's also worth pointing out that my proposed operationalization of neutrality does not require us to be able to differentiate software from hardware colorations as such. It only requires us to make judgments about how much differentiation a system is capable of.

In the case you mention, where variations in the sound of the same software on two different systems leads to ambiguity about software vs. hardware colorations, my method of determining neutrality offers a potential solution to the deadlock. By determining which system has greater differentiation, particularly in the domain where the ambiguous coloration occurs (e.g., instrument timbre), you can conclude which system is the more neutral of the two. If the coloration in question does not occur, or occurs less, on the more neutral system, then it is likely to be a playback coloration added or aggravated by the less neutral system. This method is, of course, fallible. But I believe it is useful for providing clues to distinguishing software colorations from hardware colorations. And if you were willing to move equipment around, the same method could be employed for distinguishing equipment colorations from room colorations.

Cbw - Your suggestion about revising the definition of 'inaccuracy' is a good one. Will work on it.

Dgarretson - I have a question for you about Cbw's entropy theory and your first operationalization of neutrality, namely, the idea that increasing neutrality results in increasing source convergence. Would Cbw's explanation of coloration in terms of decreasing entropy explain the convergence you predicted? And since his explanation extends to the entire playback chain, should we predict that, as whole systems become more neutral, they will sound more and more similar to one another?

Dgarretson, thanks for the clarification. I actually didn't know about the Holt dictionary. The only place I had previously seen common audiophile terms defined was in Robert Harley's book. I'm sure there isn't much difference between those two, anyway. That Holt book would be fascinating reading. I find it amusing that he defines "neutral" as free from coloration, as you guys are, but then feels the need to define "uncolored" as free from AUDIBLE coloration. This is especially funny to me coming from the guy who defined "subjectivist" reviewing! Harley's book also speaks of the objectivist/subjectivist divide in the same sense as Holt does, by the way. It was my understanding, though, that Harry Pearson was the one who defined many of these terms originally, and he was certainly the one who defined the concept of "the absolute sound."

Getting back more on topic, I have one comment on your observation that "To distinguish problems in playback from problems in recording, the trained listener merely needs to listen to a wide variety of recordings on the same playback system." While I do agree with this as far as it goes, this is only part of it. I believe the trained listener must also do the opposite - listen to the same few recordings that one is very familiar with on a wide variety of playback systems. This is much more useful for evaluating the playback end of the equation, while the former helps distinguish problems in recording.

However, no matter how precisely we can define various different terms and types of colorations, etc., (and I am not saying this is not valuable) people will hear many of them differently, for many different reasons. Setting aside personal preferences/reference points, one audiophile may have a much better/more trained ear than another. One could also have a better but more untrained ear than the other, a case which can really confuse the issue for both. Another example I find is all too common in the audiophile world - someone who thinks they have a good and/or trained ear, and knows quite a bit about the science behind audio products, but unfortunately doesn't actually hear very well. I'm sure we can all think of a dealer where this is the case! Others mentioned the effects of aging/hearing loss - everyone's ear, no matter what level it is/was originally, can/does/will change, for better and/or for worse. Unfortunately, as an orchestral musician I am guaranteed to lose at least 20% of my hearing over the course of my career. What sounds better to you today may not tomorrow, and this will change your perception of many types of "colorations." I guess I've said all I really have to say on the subject, though I do find the discussion of terms interesting and will continue to follow the thread.

Speaking of ear training, I would urge all audiophiles to consider taking a formal aural skills course - these are often offered as adult extension courses at music programs in large universities. This sort of ear training is much more valuable for actually listening to your music (as opposed to your system), and always leads to much more enjoyment of your music, no matter what type you listen to. It will also have the benefit of greatly increasing your ability to listen for your system's characteristics, especially the more musically-related ones. And it is much cheaper than a new component for your system, too, LOL! It's all about the music in the end, or should be, anyway. I have greatly enjoyed the discussion - thanks for starting the thread, Bryon!

preference for a particular coloration(s))and the general tendendency for most"audiophiles" to disagree about anything negates many of the aforementioned arguments presented so far.

Learsfool – I appreciate your thanks and I’m grateful for your participation. Though our points of view never converged, I learned a lot from our debate along the way. Your final comments about training the ear, which I think of as training the brain, are well taken. Although we didn’t discuss it much in this thread, the ear/brain is probably the most important “component” in the system.

Cbw – I have given some thought to your suggestions for revising the working definition of ‘inaccuracy.’ Here is what I came up with:

INACCURACY: An alteration to information in a component or system that eliminates, conceals, or corrupts information about the music.

Defining 'inaccuracy' in this way raises the question: How does this new concept of 'inaccuracy' relate to the concepts of 'accuracy,' 'neutrality,' and ‘resolution’? I would like to offer a new proposal about the relations among these concepts. To begin with some definitions:

ACCURACY: 1. The relative amount of information about the music presented by a component or system, comparing output to input. 2. The degree of absence of inaccuracies.

INACCURACY: An alteration to information in a component or system that eliminates, conceals, or corrupts information about the music.

NEUTRALITY: The degree of absence of coloration within a component or system.

COLORATION: An inaccuracy audible as a non-random sonic signature.

RESOLUTION: The absolute limit of information about the music that a component or system can present.

These concepts form some of the basic units of a set of equations that express the relations among accuracy, neutrality, and resolution:

1. CA = (1/L+N+D)
2. CN = (1/DoC)
3. CR = CA + FR
4. SA = SoCA
5. SN = SoCN
6. SR = SA + FR

Where…

CA = Component Accuracy
CN = Component Neutrality
CR = Component Resolution
SA = System Accuracy
SN = System Neutrality
SR = System Resolution
FR = Format Resolution
L = Loss
N = Noise
D = Distortion
So = “Sum of”
Do = “Degree of”

Taking them one at a time…

1. CA = (1/L+N+D). A COMPONENT’S ACCURACY is determined by the amount of loss, noise, and distortion within the component. More specifically, a component's accuracy is INVERSELY PROPORTIONAL to its loss, noise, and distortion. The other way of saying the same thing: A component's inaccuracy is DIRECTLY PROPORTIONAL to its loss (the elimination of information), noise (the concealment of information), and distortion (the corruption of information).

2. CN = (1/DoC). A COMPONENT’S NEUTRALITY is determined by its degree of coloration. More specifically, a component's neutrality is INVERSELY PROPORTIONAL to its degree of coloration. This equation was proposed by Cbw in an earlier post.

3. CR = CA + FR. A COMPONENT’S RESOLUTION is determined by the accuracy of the component and the format resolution of the source. Specifically, a component's resolution is DIRECTLY PROPORTIONAL to its accuracy and the format resolution.

4. SA = SoCA. A SYSTEM’S ACCURACY is determined by the sum of its components’ accuracy. Specifically, they are DIRECTLY PROPORTIONAL.

5. SN = SoCN. A SYSTEM’S NEUTRALITY is determined by the sum of its components’ neutrality. Specifically, they are DIRECTLY PROPORTIONAL.

6. SR = SA + FR. A SYSTEM’S RESOLUTION is determined by the system's accuracy and the format resolution of the source. Specifically, a system's resolution is DIRECTLY PROPORTIONAL to its accuracy and the format resolution.

I'm not really proposing a deep mathematical relationship among these concepts. The equations are more of a mathematical analogy for describing the logical and conceptual relations among these categories.