Preference vs. audibility - please keep them separate.

[Orb]

The problem, of course, is that subjective results are not transitive. This means that if you make something .3dB more intense, most people are going to think it sounds better. And, comparing .3 and .6dB more intense, they will vote for the .6dB. HOWEVER, if you compare the .6 (or .9 or 1.2 or something farther along, depends on source and person) with the original, they will dislike the higher level.

In other words if A>B is true, B>C is true, and C>D is true, that does not mean that A>C or A>D is true.

As to the A-flat remover, it wasn't a particularly good idea in the first place. When we heard about it at Bell Labs in the Acoustics Research Department, the reaction was a mix of amusement and exasperation. The argument that it would take a contrived signal to detect it was at least moderately wrong, and the idea that it would work at all was also pretty strange, given noise modulation in cassettes, distortions in LP's, FM noise, and so on. It was before any perceptual coder (it was an analog system), and perceptual coding would have driven it completely bonkers, too.

No disagreement there JJ, but this just emphasises the point that preference and audibility/quality performance do go hand in hand when talking scientfically.
Although maybe part of the problem is the initial post needs to be expanded because we can all take a slightly different interpretation as it stands and its context; because Harman group is a classic example where their scientific studies do include preference/performance/cognitive sighted-blind as part of a study research paper.
So I accept some of us may be misunderstanding the context of that 1st post.

Cheers
Orb

 

[Phelonious Ponk]

Because it's mainly about design. The testing that everyone seems to think is the key to steering development is really just confirmation.

Or... putting it another way, given the sources, devices, components that have all been developed by objective design methods, it's hard even for a boutique wannabe 'high end' company to get it spectacularly wrong when following the data sheets and bolting them together. A 300kg solid gold chassis won't actually do any harm, even if it does no good either.

What Groucho said. The question that can be answered by a combination of measurement and blind listening, even informal, flawed, blind listening, is not "can a high-end manufacturer put together a great-sounding component using established design standards and the most expensive components, the questions are things like:

Does the $20k component in three boxes actually sound better than a $1000 one in a single box?

Does the $1000 component (say a DAC) sound better than the one built in as just a component of a component that doesn't cost much more than the free-standing DAC alone?

Do two components that measure the same actually sound different?

Is there some undiscovered, unmeasured quality in high-end components that makes them sonically superior?

Answereing these questions, even if just a fairly small percentage of the answers don't go their way, is not in the economic interests of high-end manufacturers and their supply chains. But these answers are very much in the economic interests of audio consumers. Why so many of them so violently resist the gathering of information that is in their own self-interest is a mystery. They should be demanding it.

Tim

 

[Robh3606]

Hello Micro

It can be an interesting debate - how does high-end achieve consistently a fantastic sounding quality using mainly sighted testing as a development tool?

I would argue that it doesn't. I have heard some systems that sounded just plain wrong to me. It's not like they all sound alike, actually some quite different. So which ones right??

Rob

 

[microstrip]

Hello Micro



I would argue that it doesn't. I have heard some systems that sounded just plain wrong to me. It's not like they all sound alike, actually some quite different. So which ones right??

Rob

The answer is easy - you should just pick the top sounding high-end system in your opinion. Ignore those that sounded wrong to you. If you really just find that the good ones sound different from non-high-end, but not better, there is no reason for debate in your opinion. But my opinion is different.

 

[microstrip]

Because it's mainly about design. The testing that everyone seems to think is the key to steering development is really just confirmation.

Or... putting it another way, given the sources, devices, components that have all been developed by objective design methods, it's hard even for a boutique wannabe 'high end' company to get it spectacularly wrong when following the data sheets and bolting them together. A 300kg solid gold chassis won't actually do any harm, even if it does no good either.

Confirmation and fine tuning is the essence of high-end. And the knowledge that is incorporated in the phase of high-end design was mostly obtained through long term listening experience, that was not blind.

Curious that you can not separate "high-end" from "boutique" or "300Kg solid chassis" - it is not good for a clear and fair debate.

 

[rbbert]

The answer is easy - you should just pick the top sounding high-end system in your opinion. Ignore those that sounded wrong to you. If you really just find that the good ones sound different from non-high-end, but not better, there is no reason for debate in your opinion. But my opinion is different.

The problem then, of course, is that it's really hard to know how good your perceptual abilities are at any given time. Too many different things affect them, especially when one is talking about tiny differences, real or not.

 

[Phelonious Ponk]

The answer is easy - you should just pick the top sounding high-end system in your opinion. Ignore those that sounded wrong to you.

This is not an answer to "how does high-end achieve consistently a fantastic sounding quality using mainly sighted testing as a development tool?" This is an answer to "can I find something I like the sound of in the world of high-end audio."

Tim

 

[microstrip]

This is not an answer to "how does high-end achieve consistently a fantastic sounding quality using mainly sighted testing as a development tool?" This is an answer to "can I find something I like the sound of in the world of high-end audio."

Tim

Tim,

Great. Now you pick a secondary sentence from the whole post, ignoring the context and imagine possible questions. Will you next tell us about your nightmares with audiophiles?

 

[microstrip]

The problem then, of course, is that it's really hard to know how good your perceptual abilities are at any given time. Too many different things affect them, especially when one is talking about tiny differences, real or not.

Yes, this is part of life. Here some self training is needed. Use your known recording references, be analytical and cold blooded.

 

[Phelonious Ponk]

Tim,

Great. Now you pick a secondary sentence from the whole post, ignoring the context and imagine possible questions. Will you next tell us about your nightmares with audiophiles?

Alright then. Here's the whole post:

The answer is easy - you should just pick the top sounding high-end system in your opinion. Ignore those that sounded wrong to you. If you really just find that the good ones sound different from non-high-end, but not better, there is no reason for debate in your opinion. But my opinion is different.

I didn't include the part about non high end because...well, because it's irrelevant, and now that it's back in, the post is still about preference and does nothing to support your question --

how does high-end achieve consistently a fantastic sounding quality using mainly sighted testing as a development tool?

Edited for brevity. And the fact that I can't page back to your original post without losing what I've entered here. And because the statement directly above, clipped from Rob's quote a few posts up, still means what it says...even in context.

Tim

 

[Groucho]

Use your known recording references

I see potential for this to give circular, self-reinforcing, and flawed, results. You may have arrived at your choice of reference recordings purely as a result of listening over and over to the same type of flawed system.

 

[KlausR.]

I see potential for this to give circular, self-reinforcing, and flawed, results. You may have arrived at your choice of reference recordings purely as a result of listening over and over to the same type of flawed system.

This Floyd Toole's vicious circle: you judge recordings using speakers which you judge using recordings, which you judge using speakers...

Klaus

 

[NorthStar]

I see potential for this to give circular, self-reinforcing, and flawed, results.
You may have arrived at your choice of reference recordings purely as a result of listening over and over to the same type of flawed system.

True, but it's still a reference by which to evaluate/audition everything else around it, and by doing so you'll eventually discover the faults of your reference and you'll switch to the next (better) reference.

The biggest advantage is with whoever has the most access to all the references. ...And there will be some that some will never be able to access.

 

[microstrip]

I see potential for this to give circular, self-reinforcing, and flawed, results. You may have arrived at your choice of reference recordings purely as a result of listening over and over to the same type of flawed system.

I choose recordings from different labels, known for the quality of their work, and that are significant for the music I like. A few of my reference recordings are from performances I have been live. IMHO it is highly improbable that they were recorded in systems that were systematically flawed in the same way, although they will favor the type of music I enjoy.

And , yes, I have read Siegfried Linkwitz!

 

[j_j]

Once again, for determining your own preference, use what you like. It's your preference, you get to have it, like everyone else gets to have theirs. There is no "flawed preference" when a single individual is concerned unless it passes the end of their nose.

For determination of strict audibility, ABX or signal detection testing (depending on the signal) is necessary. But this does not have to be used in the high-end unless someone PREFERS to use it.

As to "group preferences", one can also call them "mob effects". Still, unless the mob does harm, there's no problem.

 

[Atmasphere]

I don't see people talking about human hearing/perceptual rules here. They mostly seem to be talking preference.

Hearing/perceptual rules are what its all about. Most of us understand the 20-20KHz thing. But there is a lot more. For example, our ears are tuned to birdsong, that is to say that is the frequency range to which we are most sensitive. This is true of all humans.

WRT DBT; (if I can help Myles out for a moment, otherwise I plan to avoid that bullet) the problem is that if you conduct a DBT between two components that are designed to look good on paper rather than obey human hearing rules, even though the specs on paper may appear different its very likely that the DBT between the two will be inconclusive. This is because very little of the paper specs that we've admired for decades actually have all that much to do with how humans hear sound. So DBTs have a bad rap, because the people conducting them often fail to understand this simple fact.

Bottom line, if you want to understand the difference between 'preference' (IOW 'taste', no accounting for it) and 'audibility', the place to start is an understanding of how our ear/brain system actually works. FWIW, that is in short supply in the audio world, which is not great if you think about it....

 

[NorthStar]

I don't see people talking about human hearing/perceptual rules here. They mostly seem to be talking preference.

Hearing/perceptual rules are what its all about. Most of us understand the 20-20KHz thing. But there is a lot more. For example, our ears are tuned to birdsong, that is to say that is the frequency range to which we are most sensitive. This is true of all humans.

WRT DBT; (if I can help Myles out for a moment, otherwise I plan to avoid that bullet) the problem is that if you conduct a DBT between two components that are designed to look good on paper rather than obey human hearing rules, even though the specs on paper may appear different its very likely that the DBT between the two will be inconclusive. This is because very little of the paper specs that we've admired for decades actually have all that much to do with how humans hear sound. So DBTs have a bad rap, because the people conducting them often fail to understand this simple fact.

Bottom line, if you want to understand the difference between 'preference' (IOW 'taste', no accounting for it) and 'audibility', the place to start is an understanding of how our ear/brain system actually works. FWIW, that is in short supply in the audio world, which is not great if you think about it....

Should we take some training courses (sessions) in the art of listening? ...And in its understandings, sources? ...That sounds pretty good to me.
- Harman Kardon is training their listeners for better accuracy in audio blind-testing. ...Makes very good sense to me.

 

[Atmasphere]

I'm sure training would not hurt, but that really does not address what the perceptual rules are all about.

You may be interested to know though, that the brain has a whole bunch of tipping points. One of them has to do with how the brain processes music. Normally music is processed in the limbic system. This is why music can be toe-tapping, to say the least. But if the brain detects too many anomalies, for example if too many hearing/perceptual rules are somehow being violated, the processing will reach a tipping point wherein the processing is moved to the cerebral cortex. When that happens, emotional content tends to be lost. Now this is a a very subjective effect, obviously, the interesting thing is that we have done enough studies that there are now some objective concrete numbers on when this happens in certain (test) cases.

When you are auditioning cables, you are engaging the cerebral cortex rather than the limbic system. So- you can see that *what* you are doing with the stereo will affect the results that you get! This is one of the reasons why DBTs are not good science- because they do not deal with how our brains work and force it to use the cerebral cortex for processing. Yet the two bits of gear under test might yield very different (emotional) results in the home where someone is simply playing music, if one piece is a bit better suited to obeying human hearing rules than the other.

 

[cjfrbw]

That's why audio can become such a Tower of Babel (babble?).

Different contributors talking about different things, but few actually addressing the main processing computer (ear brain system) that does the heavy lifting, or whether the goals are actually cerebral or subliminal/limbic.

I personally like the system to go rather quickly through the cerebral phase and get into immersion in the subliminal/limbic. I have not heard many systems that manage that. Most have a tendency to "shock the monkey" back into the cerebral too often.

 

[microstrip]

Once again, for determining your own preference, use what you like. It's your preference, you get to have it, like everyone else gets to have theirs. There is no "flawed preference" when a single individual is concerned unless it passes the end of their nose.

For determination of strict audibility, ABX or signal detection testing (depending on the signal) is necessary. But this does not have to be used in the high-end unless someone PREFERS to use it.

As to "group preferences", one can also call them "mob effects". Still, unless the mob does harm, there's no problem.

Again, you should separate "group preferences" that you appropriately call "mob effects" from the preference of a group of people who independently made their choices without any mutual interference.