The Objectivist Audio Forum: Post #2 Definitions - Page 4

I understand your point and I agree with you completely, but I do not understand how discussing our personal opinion in the objective forum will be any different then discussing it in a different part of this site.

My point is, if the objective forum is only going to perform tests using human ears, then we are really back to where we started...with no actual proof.

We can all discuss and debate until we are blue in the face if a cable really changes the way music sounds to us, and none of it will actually prove scientifically what the truth really is.

Doing a DBT on 2 cables may yield 7 out of 10 people can not tell the difference between the 2, but to me that is not actually scientific proof.

Huh.

From my point of view, this is exactly what it is about.

Well said, I hope it actually turns out that way.

This is a good point. What does "7 out of 10 people can not tell the difference" mean? It means, during this single test (or series of tests) 70% of those selected couldn't tell the difference. It also means 3 out of 10 could or think they could.

For those that could not, we can dismiss them from future studies for the moment. Now, how do we deal with this 30%? Are we going to assume that they didn't hear anything but lied? Or perceived but were deceived? OR! might we assume that in fact they did hear a difference or perceive a difference and try to figure out how and why? Some will simply drop the argument after seeing the 70% score: "The others must be living on a placebo effect"

Yes, this could be true. No doubt it may very well be true! But it might not be so. If not so, let's adapt, propose new studies or methods of researching.

Cables might be the fun stuff to pick on (or those joy stones, quantum chips etc) but when it comes to amps or front-ends etc., there are more people who feel/believe/know that they hear a difference. How about starting here, and then working down to the level of shiny stickers?

If folks wanted to focuss on cables (since it is the hottest topic), cables being used in tests would need to be tested before hand for conductance/resistance, inductance, RFI, etc. If no differences were measured, then two or more cables could be used in a test. Merely throwing in a few brands of cables and concluding something isn't very scientific since differences, easily measured among the products, are unavailable yet necessary in a proper results discussion.

Some subjectivists certainly won't be happy with any reasonable explanation of how the world works. Others will be open minded and want to learn, particularly if the proper literature is cited. However, at the same time, some objectivists won't be happy with any reasonable explanation as to how a person might be able to hear a difference if they too walk into the experiment biased.

Will a given experiment cover all variables? No. Isolating which variables to test and then determining proper means of standardizing and testing is paramount to any meaningful outcome. If that means using blood tests, EEGs and a T-Stat oximeter to get proper basal levels for those undergoing the DBT then so be it! This data shouldn't be excluded because it is too difficult to gather or analyze. All sides should be open to discovering the truth behind their gear and how and why we enjoy music. One shouldn't necessarily have to go to insane lengths to prove a point, occam's razor and all that. However, it also means that new data being discovered ought to be considered and implemented in simpler tests before dismissing outright, someone's claims.

I really have no interest in performing tests. I leave that to people who make a living at that. My primary interest here is in discussing how people can efficiently and economically achieve better sound from their equipment. For myself, I do that in a logical manner by making an observation, coming up with a hypothesis and supporting it or disproving it with evidence. That's a rational and objective approach to problem solving.

I'm not interested in tests for tests sake. I'm only interested in tests as they relate to helping people get better sound out of their systems. Objective charts and numbers and test tones are only useful to me when they are related to how music subjectively sounds to human beings. Ultimately, we listen to systems with our ears. That's what really matters.

If the topic of the new forum is going to be limited to just reports of individuals' tests, or if people aren't allowed to make subjective observations as a beginning point for discussion, I'm not interested in participating in that particular forum. Perhaps I'm not one of the folks this new group is designed to contain. I'll just continue to participate as I have been in the other forums. Nothing wrong with that.

All I care about is that discussion is free and open, based on logic and supported with evidence, and free of ad hominem attacks. To me, that is "reason", "objective", "rational"... whatever you choose to call it. I don't have to wear a white lab coat and don goggles to do that.

See ya
Steve

The next step is to increase the sample size for those three and see if the reason they scored well was due to statistical random chance, or whether they actually could hear a difference. If they are able to hear a difference, you determine what percentage of the public they represent. That will establish your thresholds to be able to say, just how important the difference really is to most people.

One answer leads to more questions. That's what the scientific method is all about.

See ya
Steve

Actually, it doesn't mean any of these things, nor does it "prove" anything one way or another. If properly constructed, it simply shows whether the number of people identifying a difference correctly is greater or less than their chances of identifying a difference correctly by chance/mere guessing.

With a large enough sample, what we can than induce from that is that the two cables (or whatever) being tested have no audible difference. This is only an induction, not a proof, but a strong one that should at least give one pause before emptying the bank account on a new piece of wire.

We do the same thing everyday when we take medicine, with the help of the FDA. If a large enough sample of people correctly identify a difference between two treatments it doesn't prove that one treatment is better than the other, it only helps us (via our doctors and pharmacists) decide which course of treatment to take.

--Chris

This is not exactly true. Quite a few studies indicate a placebo induced some form of comfort, either in terms of pscyhosomatic pain relief, overall feeling of well-being etc. Meanwhile the cancer spread and overall vitals decreased. The best course of action in these cases was clear - give the cancer meds. The benefit of the placebo in rarer cases was merely this - if the cancer meds aren't working and are actually increasing degeration of other organ functionality, providing the placebo will at least provide some comfort while not aiding in hastening death.

I'm not sure what part of what I said you don't think is true. I didn't really make a truth statement... in fact, I think your example is a good illustration of what I mean when I say "it helps us to decide a course of treatment."

Many factors help us decide, which is the core of our pursuit here at Head-Fi. But notice that the studies of subjective benefits of placebo were carried out in an objective manner. Likewise, subjective/perceived benefits in audio equipment can be examined in an objective manner. Maybe the feeling of ownership of a beautiful, expensive cable subjectively improves one's enjoyment of the music, but this can still be objectively investigated to help determine whether the perceived difference is auditory or psychological.

--Chris

Agreed.

How do you see that applying to audio?

See ya
Steve

I don't, this was in response to hempcamp's comparison of testing by the FDA for medicine.

The experimental designs need to be more sophisticated. The 15-person panel (or 15 people at different times, need not be all at once) is/are given a series of A/B comparisons, repeated over and over, at different (A/B matched each time) volume levels, and with different musical passages. Individual conditions are repeated multiple times.

After each A/B test, the listener writes on a specially prepared pad (no one knows anyone else's answers):

Time of day, plus whether s/he heard a difference, and if so, which was preferred.

False comparisons are thrown in: the same cable is used but called out as A/B.

The order of what is called A and what is called B is changed.

Additional cables (ringers not under test), especially damaged ones known to be audibly poor, are also switched in without telling anyone.

After all this we can analyze and find biases, sort the golden ears from the tin ears, see if someone who notes a difference can do it reliably etc. People who can't hear any difference but say they do at random are smoked out in a heartbeat.

It is not a 7 out of 10 thing, and we do not test further the people who found differences (there were many replications in the base test, so this is not needed). You see, we can tell if the differences are reliable, because we have replication and tricks thrown in.

We get more people and run another panel to validate the conclusions.

This is all based on food industry testing protocols. At dinner with a bunch of other statisticians some years ago I claimed to be able to tell South American coffee from African coffee ... they had both on the menu. Well they sent me to the men's room and prepared 10 2-cup challenges. The bastards even blended some of the cups! I failed totally. We also spent like $80 on coffee!

I am so eager to do this -- we must plan this for Can Jam '09 ... or before.

You will probably enjoy this:
Malcolm Gladwell on spaghetti sauce | Video on TED.com