What uncontroversial audible differences cannot be measured?

[wgscott]

I said something in another thread that raised a lot of objections, and this caught me by surprise.  (What I said was that I knew of many examples of measurable differences that cannot be heard, but did not know of any audible differences that cannot be measured.) I could not think of any such examples.

 

Anyway, if there are well-known, established examples of readily audible but unmeasurable differences (that stand up to various reality checks like blind testing, etc), please list them.

 

Many thanks in advance.  I've started this as a separate thread so as not to hijack others, and also because it seems that this is a fundamental point that comes up in many discussions here.

 

This thread has been hijacked and hopelessly derailed.  If I could lock and delete it, I would.  For anyone interested, the question remains unanswered.

[The Computer Audiophile]

Very good topic. 

 

Wish I had something to offer. 

[Daudio]

 

[wgscott]

Just to clarify, I fully understand that there are differences that cannot be reduced to measurement, like an emotionally-charged vs. bland rendition -- you can measure these differences, presumably, but there is little if any indication how to interpret such measurements in terms of how emotionally compelling a rendition might be.  (The differences are objectively measurable, even if the interpretation of them in this manner is not possible.)

[The Computer Audiophile]

This could be very educational, if people have information about this topic. Perhaps another topic that's closely related would be correlating measurements to sonic attributes. I won't derail this topic though. 

[Skip Pack]

Here's and analogy that might have equivalents in audio: Reciprocating engines (others too) are typically described as generating a certain torque value at each measured rpm and horsepower falls out when the two numbers are combined. However the time it takes to accelerate from one rpm to a higher rpm under a near full load can only be calculated if you know rotating mass (as well as other things, I imagine). The analogy here might be the change of each order of harmonic distortion at a near-step change in overall amplitude. I'm not knowledgeable enough to know if this is captured fully in another measurement, but I suspect changes in normally constant test characteristics when other factors are temporally variable might correlate to our perceptions.

 

Skip 

[Fitzcaraldo215]

1 hour ago, wgscott said:

I said something in another thread that raised a lot of objections, and this caught me by surprise.  (What I said was that I knew of many examples of measurable differences that cannot be heard, but did not know of any audible differences that cannot be measured.) I could not think of any such examples.

 

Anyway, if there are well-known, established examples of readily audible but unmeasurable differences (that stand up to various reality checks like blind testing, etc), please list them.

 

Many thanks in advance.  I've started this as a separate thread so as not to hijack others, and also because it seems that this is a fundamental point that comes up in many discussions here.

Excellent concept.  I am undoubtedly and unashamedly biased, but the premise that we (meaning most of us humans, not one individual) can readily hear things that are not measurable is a myth, IMHO.  

 

It is also quite different from the frequent audiophile complaint that there are sonic differences not shown by the measurements actually taken in many cases. That may be true sometimes, but it does not disprove the potential measurability of any truly audible difference.

 

We all know how these positions or misapplications of them get dragged into debates, and they probably always will.

 

But, I will be totally stunned if anyone can provide an example that answers the original question.  

 

 

[Norton]

It was my understanding that , in the UK at least, HIFi journalism went through  and then abandoned a measurements-dominated approach  precisely because so many amplifiers (in particular) measured the same, yet sounded different.

 

Of course, it's possible that pressure from manufacturers of expensive amplifiers that measured  the same as cheap ones may have played as much a part in this, as the genuine shortcomings of relying solely on measurements.  Most probably  didn't make interesting reading either.

[sphinxsix]

First of all I'm afraid for some almost all sonic differences will be 'controversial'   Whatever 'uncontroversial' means for you..

The very first questions that came to mind.

1. Can soundstaging (eg soundstage depth) be measured?

2. Can something that I call timbral richness (eg timbral difference between 2 different saxophones in the same recording) be measured?

I've read so many times reviews in which reviewers with lots of listening experience and lots of technical (measurements) knowledge admitted that gear sounded different from what measurements suggested it 'should' sound like... Never measured gear I listened to though.

BTW do objectivists really buy gear just because it measures well.? Without checking reviews or listening to it? Just be frank, guys

 

[ted_b]

Mood and taste are two that pop to mind.    Not trying to derail, but add that emotional and spiritual biases cannot be measured but clearly have an effect on sq..especially if they arrive from the recording (i.e if all parties in a double blind are moved by the same music).

 

Regardless, this is a great topic, as I firmly believe that all we measure is not all important and that what IS important cannot all be measured ...and the corollary...we should not stop trying to measure things, as most important things can probably be measured at some point, we are just doing it wrong currently.  The quest for discovery is a wonderful thing. 

[wgscott]

16 minutes ago, sphinxsix said:

BTW do objectivists really buy gear just because it measures well.?

 

Sometimes if you want to understand something better, or even learn something new, it helps to dispense with the sophomoric and dismissive caricatures of those who have an opinion that happens to be at variance with your own. 

[wgscott]

11 minutes ago, ted_b said:

Mood and taste are two that pop to mind. 

 

I'm going to suggest (for purposes of the OP of the thread) that this is an example of

 

3 hours ago, wgscott said:

Just to clarify, I fully understand that there are differences that cannot be reduced to measurement, like an emotionally-charged vs. bland rendition -- you can measure these differences, presumably, but there is little if any indication how to interpret such measurements in terms of how emotionally compelling a rendition might be.  (The differences are objectively measurable, even if the interpretation of them in this manner is not possible.)

 

[jabbr]

To paraphrase the saying: not everything measurable has been measured.

[Fitzcaraldo215]

3 minutes ago, sphinxsix said:

First of all I'm afraid for some almost all sonic differences will be 'controversial' Whatever 'uncontroversial' means for you..

The very first questions that came to mind.

1. Can soundstaging (eg soundstage depth) be measured?

2. Can something that I call timbral richness (eg timbral difference between 2 different saxophones in the same recording) be measured?

I've read so many times reviews in which reviewers with lots of listening experience and lots of technical (measurements) knowledge admitted that gear sounded different from what measurements suggested it 'should' sound like... Never measured gear I listened to though.

BTW do objectivists really buy gear just because it measures well.? Without checking reviews or listening to it? Just be frank, guys

 

Yeah, well, we can't exactly measure "controversial", so we will just have to handle that issue subjectively and decide that by consensus. 

 

As to your suggested questions, first soundstaging.  Good one!  I do not think anyone denies soundstaging exists in our listening perception with stereo or Mch sources.  We rarely do see it measured, but I have some test CDs that move the recorded sound sources - jangling keys, spoken voices, etc. - around in controlled ways such that listener responses via laser pointer, etc. of changes in apparent position of the sound source could be measured.   Soundstage depth might be much tougher, but still doable.

 

Part of the measurement problem here is the illusion of soundstage is entirely dependent on speakers, dispersion patterns, room reflections, etc.  So, there is no absolute scale by which to judge soundstage reproduction.  For example, the same speakers would likely give a different perceived soundstage in a different room or if speaker position in the same room was changed, let alone if the listener changed position, turned his head, etc.  Different speakers in the same room might also have different "ideal" positions or maximum soundstage.

 

The point is what is the point of such measurements when there is no one absolute correct amount of soundstage width, depth or height for a given recording?  It all depends, but comparative differences could be measured for a given setup.  One setup could be determined to give a "bigger" soundstage with that recording by objective measures, except that is hardly ever done.  And, is a "bigger" soundstage "better", "more accurate", etc. for that recording?

 

Incidentally, one set of wacko experiments by Zelig and Clawson concluded that apparent soundstage height indicated by a listener with a laser pointer, was, in fact, a proxy for "system quality".  They made many measurements of that varying the electronics or WAV vs. Flac using the same music, trying to keep all else equal, like system and listener position.  I think they proved nothing and were in la la land, but the concept of measuring perceived dimensions (at least height) of the soundstage was definitely used in their misguided experiments.

 

But, Dr. Scott is the boss.  Does this truly qualify as a successful answer?

 

On comparative timbers like richness, that is easy.  I believe that is measurable by comparing traditional measurements, primarily just of frequency response,  of the two saxophones in your example. It is just like they do in spy movies when comparing voice signatures.

[semente]

1 hour ago, sphinxsix said:

First of all I'm afraid for some almost all sonic differences will be 'controversial'   Whatever 'uncontroversial' means for you..

The very first questions that came to mind.

1. Can soundstaging (eg soundstage depth) be measured?

2. Can something that I call timbral richness (eg timbral difference between 2 different saxophones in the same recording) be measured?

I've read so many times reviews in which reviewers with lots of listening experience and lots of technical (measurements) knowledge admitted that gear sounded different from what measurements suggested it 'should' sound like... Never measured gear I listened to though.

BTW do objectivists really buy gear just because it measures well.? Without checking reviews or listening to it? Just be frank, guys

 

 

As far as I know, both "timbral richness" and "soundstage" are properties of the recording (the latter achieved by capturing the ambience cues of a venue or by "creating" space through EQ, panning and reverb).

The most "transparent" system will reproduce the recording in the most accurate manner BUT it is possible to "enhance" some aspects of the recordings (namely "3D-ness") through the addition of effects/artifacts such as low, even-order harmonics, phase manipulations or even through room reflections (in some cases at the expense of "timbral richness").

Which raises an interesting issue: are we trying to (capture and) reproduce space or music?

 

 

As for reviewers dismissing or downplaying the importance of shortcomings, there's an interesting text by Arthur Salvatore named "AUDIO RELATIVISM"-THE NEW DISEASE AND EXCUSE":

 

What is Audio Relativism?

It is the belief system that virtually every component has strong merit and can produce "great sound", for someone's "tastes", if it is matched correctly with other components.

What are (some of) the ultimate implications of this theory?

1. Every component has some validity in the audio marketplace.

2. No component is inherently superior to another.

3. All sound reproduction differences are just a simple matter of taste.

4. There is no objective standard to aim for.

5. (High) Fidelity to the source, or in general, is irrelevant.

6. Audio is an "Art Form", like poetry and sculpting, rather than a science.

This "belief system" is very convenient for its creators; the audio magazines and their 'reviewers'. It provides them with the ability (and the excuse) to find some "good", or some "justification" to purchase, within every single component that is reviewed by them. How?

Because, if their belief is true, each and every component, under the right conditions, can equally satisfy listeners as much as any other component, for either the money or in the absolute sense. It's just a matter of time, or luck, before you find the right "match".

In effect: this theory means that all component performance is "relative" and with no "absolutes"; only "possibilities" exist.

 

http://www.high-endaudio.com/philos.html#Rel

 

My guess?

Audio magazines live of advertising, wholesale prices and long term loans...you don't bite the hand that feeds you.

[AJ Soundfield]

4 hours ago, wgscott said:

Just to clarify, I fully understand that there are differences that cannot be reduced to measurement, like an emotionally-charged vs. bland rendition -- you can measure these differences, presumably, but there is little if any indication how to interpret such measurements in terms of how emotionally compelling a rendition might be.  (The differences are objectively measurable, even if the interpretation of them in this manner is not possible.)

I think you would have clarify if you are talking about electro-acoustic measurements...or measurements of people.

[AJ Soundfield]

Btw, here is example of a controlled study attempting to correlate sound metrics of loudspeakers

Open access, so free for now...http://www.aes.org/journal/online/JAES_V65/5/#paper1

Download the hi rez version if you're an audiophile 

[Fitzcaraldo215]

55 minutes ago, ted_b said:

Mood and taste are two that pop to mind.    Not trying to derail, but add that emotional and spiritual biases cannot be measured but clearly have an effect on sq..especially if they arrive from the recording (i.e if all parties in a double blind are moved by the same music).

 

Regardless, this is a great topic, as I firmly believe that all we measure is not all important and that what IS important cannot all be measured ...and the corollary...we should not stop trying to measure things, as most important things can probably be measured at some point, we are just doing it wrong currently.  The quest for discovery is a wonderful thing. 

We could start a real quarrel here.  But, one of the issues we trip over all the time is the difference between "hearing" and "listening".  I think I heard Conductor Marin Alsop refer to that key distinction a few months ago as it affects an orchestra conductor.

 

I am just as guilty as anyone of carelessly misusing the terms hearing and listening interchangeably.  Audiophiles do it all the time when they say something like, "I trust my ears"  but then talk about the emotional involvement with the music, which clearly involves the brain and the intellect, and not just the ears.  Does anyone doubt that the human hearing/listening mechanism is very complex, involving different layers of sensory  (the hearing) plus unconscious and conscious responses (the listening) of the mind to sonic stimuli?  

 

Confusion and lack of clarity between hearing and listening explains many debates about audio perception, even about gear, measurements, etc.

 

I do not know about Dr. Scott, but I am willing to totally concede that measurements are useless beyond a certain point in our hearing/listening sensory system, particularly when when we are dealing with the emotional responses of our feeble brains.  So, those issues are automatically "controversial", in my view.  

 

Why?  I suspect most all our ears actually respond to aural stimuli in fairly similar ways.  But, when our much more diverse (or perverse) brains capture and process those similar stimuli, there is a huge difference in our individual response and what we think we perceive. Ergo, Chinese music does not sound good to me as music and I lack emotional connection with it. But, as just sounds, Chinese ears are just as good as my Western ones in hearing the sounds.

 

Audio and audio engineering cannot deal with all this, and they try to focus on the common denominator of sound via hearing.  That is all that audio can provide.  The hearing of great emotional things from that sound is entirely up to us in our individual way.  Even outside audio, there are no good measures of emotional responses to sensory stimuli.

 

 

[semente]

1 hour ago, ted_b said:

Mood and taste are two that pop to mind.    Not trying to derail, but add that emotional and spiritual biases cannot be measured but clearly have an effect on sq..especially if they arrive from the recording (i.e if all parties in a double blind are moved by the same music).

 

This is quite interesting for I find that listening to music and evaluating performance are different tasks.

 

[semente]

When I am assessing the performance of a particular piece of equipment I focus on the sound, on certain aspects of the reproduction, not on the music.

I choose passages that I know will emphasize certain aspects and help me determine possible shortcomings or improvements when compared to my references; I believe that good references of both live and reproduced sound are paramount.

And I know that if for example the well recorded bassoon on Webber's Op.75 (BIS) sounds natural and realistic then I will enjoy the way that equipment/system reproduces music, that it will provide listening pleasure.

[watercourse]

A characteristic that I've heard a lot of people use, and that I believe people can distinguish between two systems under comparison is: "slam".

 

Sure, it's the dynamics and transient handling abilities of a system, and damping factor in the amp has something to do with it, as well as speaker driver characteristics, and interactions between the amp and speakers, but I'm not sure how the differences in "slam" would be measured, as it is usually spread across multiple frequencies and is more than just a measure of spls at a given point in time. Waterfall plots may show differences between systems, but how do these differences translate to what the listener perceives as more slam?

 

Also agree with "soundstage depth," "more realistic timbre," ... and I'll raise you "image density" and "imaging specificity." I'm having a hard time visualizing how these sonic qualities can be measured with repeatability.

 

I think that these additional qualities are not controversial concepts, but who knows? A lot depends on listener interest, acuity, and sensitivity. I, for one, have a hard time understanding what PRaT means... "pace" "rhythm" and "timing" seem redundant to me, or at least circle around too similar concepts to really be distinct in my mind.

[mansr]

2 minutes ago, watercourse said:

A characteristic that I've heard a lot of people use, and that I believe people can distinguish between two systems under comparison is: "slam".

My subwoofer (REL) has switch to select normal or "slam" mode. It alters the frequency response. I don't use the slam setting.

6 minutes ago, watercourse said:

I, for one, have a hard time understanding what PRaT means... "pace" "rhythm" and "timing" seem redundant to me, or at least circle around too similar concepts to really be distinct in my mind.

I can see those terms meaning something if assessing a performance. For electronic reproduction, not so much. Even a cheap FM radio conveys those aspects perfectly.

[jabbr]

1 hour ago, AJ Soundfield said:

Btw, here is example of a controlled study attempting to correlate sound metrics of loudspeakers

Open access, so free for now...http://www.aes.org/journal/online/JAES_V65/5/#paper1

Download the hi rez version if you're an audiophile 

I particularly liked "Real-Time Emulation of the Acoustic Violin Using Convolution and Arbitrary Equalization"

 

[watercourse]

2 minutes ago, mansr said:

My subwoofer (REL) has switch to select normal or "slam" mode. It alters the frequency response. I don't use the slam setting.

So is it simply emphasizing certain frequencies in the case of the REL? Would you say this is a universal way to define slam?

If so, when comparing two systems, the one whose frequency curves are more similar to REL's "slam" curve has more "slam"?

[watercourse]

4 minutes ago, jabbr said:

I particularly liked "Real-Time Emulation of the Acoustic Violin Using Convolution and Arbitrary Equalization"

 

It seems like the first article ("Modeling Perceptual Characteristics of Loudspeaker Reproduction in a Stereo Setup") may get at "slam" measurements (termed "punch"), and in the study, is highly correlated with spectral characteristics as suggested by @mansr... thanks for that @jabbr