The Objectivist vs. Subjectivist Debate (Long) (1 Viewer)

[John Royster]

And so it detiorates into the validity of DBTs.

Like it always does, like it always has.

This is really the key point. Can DBTs be considered science if science itself cannot trust its own tests?

 

[ManW_TheUncool]

I would think for one to be reasonably sure about the validity of a particular DBT, one would need to prove that the DBT is valid in the first place.

If you're gonna use DBT to verify the existence of subtle sonic differences, you need to prove that such subtle sonic differences can actually be heard consistently in the DBT itself when they really do occur. Obviously, that doesn't mean jitter must be audible in the DBT in order for it to be valid for jitter DBT. But some other known phenomenon w/ the same sonic difference should be audible in the same DBT. Otherwise, the DBT doesn't really prove the negative since it may not be valid. Of course, if it's positive, then the DBT can be used to prove or disprove something else w/ the same claimed sonic difference.

I don't really know, but is that really the case w/ the so called "valid" DBTs? It sounds like no from what I'm hearing (and not hearing) in these debates.

_Man_

 

[Lee Scoggins]

If two guitars are tuned to exactly a440 then why do they sound different? Can we measure that?

John,

Good point. I really meant that one guitar may be naturally recorded than another...and that cannot be easily measured without the use of the greatest test instrument of all - the human ear!

 

[george kaplan]

I'm still not clear on the validity issue of double-blind tests that you guys are talking about. But given the volatility here, perhaps I'd rather not know.

My advisor (who did some pioneering work in psychophysics among other areas) once told me I was wise to become a psychometrician instead of a psychophysicist. I guess he was right.

 

[Rick_Brown]

Lee, I am certainly not beating up on science, and your points on jitter were well said. What I was trying to do with this thread was to show how subjective listening and objective science are intertwined. That when we firstly percieve "something" that we can't explain, we need science to try to analyze and demonstrate the principals at work.

The evolution of digital recording, CD mastering, CD production and CD playback equipment design since the introduction of the CD is a good example. In the 80's critical listeners knew that something was not right. Then-current science could not explain it, but since then more and sophisticated scientific work has come up with many discoveries that have advanced to sound of CD's substantially.

As Ayn Rand said, "Our sensations, of course, do not identify any of these facts, but they do constitute our first form of grasping them and our first lead to their later scientific discovery."

 

[Lee Scoggins]

As Ayn Rand said, "Our sensations, of course, do not identify any of these facts, but they do constitute our first form of grasping them and our first lead to their later scientific discovery."

Yes, and while we are waiting on later discovery, put on your favorite LP or Super Audio CD and listen subjectively for differences.

 

[Lee Scoggins]

I'm still not clear on the validity issue of double-blind tests that you guys are talking about.

I guess what is frustrating is that some "objectivists" believe that if a group of listeners do not conduct a blind test then hearing an audio phenomena is not really valid.

For example, in my jitter test, six of my recording engineer friends clearly here a difference in 20ps and 200ps difference created by a master clock. But we do not need a DBT to validate our own critical listening skills.

Some AES papers explain high end audio phenomena through DBTs but if you search the archives, you will see that these tests are very hard to implement. Indeed, there are papers on how difficult the tests can be as a validation tool!

It should be noted that several high end companies employ double blind listening sessions to help refine new gear. I believe that Revel and Audio Research do this, for instance.

Nevertheless, I am excited about my new jitter test CD that will have the 20 ps and 200 ps music samples. I plan to send these out to several people on chat rooms and ask which track sounds better. Since the tracks will not identify low and high jitter samples, it will be as blind as can be. It will be interesting to see what happens across several playback systems of varying resolution. We also plan to play it back for the Audio Society on one highly resolving system at a later date.

 

[george king]

George,

Who did you work with? Was your advisors mentor Smitty Stevens? The Harvard group was a lot of fun.

 

[Brent Hutto]

Let me ask this as clearly and non-confrontationally as I can manage.. When you say


To expand on this, I would expect the reasonable DBT protocol to specify that each testing session include large differences that should be obviously audible, lesser differences that should be audible with careful listening as well as the tiny differences of which there's a question as to whether they are audible or not. You might expect 20 out of 20 test subjects to correctly perceive the obvious, 18 out of 20 to perceive the lesser and only 14 out of 20 to perceive the effects under investigation.

If the "obvious" differences are not perceived by a significant proportion of the testers, then there's either a problem with the test procedure or you should have used better testers or your differences weren't as obvious as you thought. OTOH, if a large proportion of testers perceive the obvious and lesser differences but the effects under investigation are only identified correctly by chance, then you've demonstrated that your testers are unable to perceive the effects being investigated.

 

[RaulR]

If two guitars are tuned to exactly a440 then why do they sound different?

Of course they sound different. No instrument can ever be tuned exactly. The mere act of strumming or plucking detunes a guitar.

 

[John-Miles]

I never read all of this thread, nor will i. But I do have one point to make.

Yes we may hear differences, but that dosent necessairly make it better.

Take all the changes you can make to a system, such as more expensive cables and such, or air bladders to reduce vibrations. when you spend more money yes you will likely hear a difference, but that does not necessairly mean that difference is a good thing.

ultimately you have to just be happy with what you hear.

 

[alan halvorson]

In audio, the argument is always put Objectivist vs. Subjectivist, the "Subjectivist" favoring pure listening tests, the "Objectivist" championing double-blind testing. That is not the argument at all. The true argument is between believable and unbelievable data. Subjectivists use listening as their test method - but so does double-blind testing. How, then, can double-blind testing be wrong if, as a subjectivist might say, it's listening uber alles? Whatever flaws or advantages subjective listening has, they are also present in DBT to the exact same degree.

All an "objectivist" asks for is a test setup in which listener bias is factored out. It's still a "listening" test. In other words, double-blind testing is a subjective process but the data produced is believable and can be analyzed using statistical methods. The purpose of a DBT is to identify differences, not to evaluate them. It doesn't require that differences be explained; it's only concerned that they exist or not.

You may certainly ask, and science often does, if a particular test was properly conducted. But if the test is properly conducted, the conclusions must be accepted.

The inference from a great many DBT's is that an electronic component's (speakers are a different ballgame) sound can be evaluated using objective data only. All differences are due to frequency response, distortion, noise, and the like - measurable qualities. Mystical qualities need not be called upon.

I'd like someone from the "subjectivist" camp to answer this: if all you need is your hearing to evaluate a component's sound, why do you need to see the component or know which is currently in the test?

 

[Lee Scoggins]

You might expect 20 out of 20 test subjects to correctly perceive the obvious, 18 out of 20 to perceive the lesser and only 14 out of 20 to perceive the effects under investigation.

My understanding is that this graduated approach will not satisfy strict DBT adherents.

 

[Lee Scoggins]

I'd like someone from the "subjectivist" camp to answer this: if all you need is your hearing to evaluate a component's sound, why do you need to see the component or know which is currently in the test?

No one requires to see anything Alan. In fact, I evaluate sound by closing my eyes. This does not logically prove, however, that DBTs are effective.

 

[Ted Lee]

i've always contended that sound is completely subjective. how in the world is someone going to argue with me about what i am (or am not) hearing?

i hear it, therefore it is.

 

[Brent Hutto]

Well, I bailed out of the "learning to hear things that might or might not be there" hobby a couple decades ago so I'm not up on the current politics of the subject. I was describing very generically a way of estimating where the border between "reproducable" and "wishful" results might lie. There's no protocol that could be designed that will satisfy "strict adherents" to both sides of a religious debate.

 

[alan halvorson]

Lee:

No one requires to see anything Alan. In fact, I evaluate sound by closing my eyes. This does not logically prove, however, that DBTs are effective.

Closing your eyes counts for nothing - you should know that. You still know what is under test. It doesn't logically prove that DBTs aren't effective either. My question remains unanswered: why does anyone need to see or know what component is under test to be able to judge it's sound?

As a note to everyone, please think about listening tests. What actually happens during a listening test? What is the process? What are the potential hazards? Describe these things to yourself in detail. Then ask yourself, how is doing the same listening test under a DBT situation any different? Again, whatever flaws listening has, they are present whether you doing a "pure" test or a DBT. You cannot use these flaws as an argument against double-blind testing without criticizing "pure" listening also.

 

[george king]

Alan,

You said
quote
It's just electricity. It's characteristics, especially in respect to mundane audio, are extremely well known. The burden of proof that objective measurements aren't enough is on those who invoke mysterious causes for what they hear, or think they hear.
unquote

Not to be unkind or mean, but that is one of the most patently absurd statements I have ever heard in this debate, and simply highlights how badly this debate is always framed. It is not simply electricity. Bits are not just bits when it comes to perception. Your brain does not know sound waves, or light waves from squat. All of the input is transformed into an electro-chemical signal (action potentials and neurotransmitter release) that your brain can understand. From there, it is not clear how you perceive stimuli - in other words, how do you recognize a signal.

Your perception of any stimulus (auditory, visual, etc) is not some simple transform of the stimulus input. Go to this sight and click on eskimo/indian

http://www.sandlotscience.com/Ambigu...iguous_frm.htm

What do you see? Depending on a shift in your attention you can see 2 different figures. THIS IS IN SPITE OF THE FACT THAT THE STIMULUS INPUT (ALL THAT ELECTRICITY YOU TALK ABOUT) HAS NOT CHANGED ONE BIT.

Folks, this debate is not about engineering and physics as there is not a single engineer or physicist here that can tell me how or why I perceive an event the way I do, even though we understand the underlying physiology reasonably well.

This is not to say that your perceptions are not dependent on the input, nor is it to say that you can't study perception with the scientific method. I am however saying that AT THIS POINT IN TIME physical measures of the operating characteristics of equipment will only go so far in telling one about perceptual phenomena. There are no mysterious phenomena to be invoked. However, saying that we do not know what they are is better than saying I know the truth, when that truth is wrong (think Copernicus).

 

[Jack Briggs]

Again, after four decades of debate (since J. Gordon Holt started voodoo audio with the original Stereophile in 1962), no one has convincingly made the case for electronics imparting sonic "signatures" on the final sound after factoring out the broad and the obvious: similar power and distortion ratings, closely matched output levels, etc.

But, what the hey. It made for a nice living before the high-end audio phenomenon started its rapid decline upon the advent of home theater.

All these years later, and not one self-styled golden-eared type has been able to rise above mere chance when double-blind tests are performed.

James "The Amazing" Randi should extend his longstanding $1 million offer to anyone who can demonstrate or prove that psychic phenomena exist to Harry Pearson and his ilk as well.

Sorry for being slightly inflammatory, but being immersed in home theater lo these many years has enhanced my appreciation for the value-to-dollar ratio of the newer hobby over the fast-fading world of high-end audio.

We've come a long way since the absurd days of Mark Levinson's silver audio cable and the Tice Digital Clock and Magic Bricks.

 

[Lee Scoggins]

Mark Levinson's silver audio cable

Hey, silver cable has wonderful audio properties if done right - look at Valhalla cable which is the best sounding on the planet right now. We should congratulate Mark for being early on this discovery.