What is the rationale behind the prohibition of DBT discussion? - Page 24

Explain to me the relationship between information theory and the question of how various distortion mechanisms are perceived.

I never said it directly mapped to how that distortion is perceived.  That is a great topic which deserves further research.  Things like weighting THD to match how the average person would perceive it would be a great step forward.

 

I don't claim that there are currently measurements for everything you'd like to know about a piece of equipment.  Its still important to listen and see how you like it.

 

I do however claim that you should be aware of the fact that it easy to fool yourself and keep that in mind when considering gear.  I do claim that it is in principle possible to measure any attribute of a piece of gear.  Just because there isn't currently a soundstage measurement doesn't mean one couldn't be devised.  The information is there.  Your brain manages to decode it.  There's no reason it should be impossible for a computer to do so.

 

I will also reiterate that in a scientific context, technical terms have a specific meaning.  It is dishonest to try to equivocate them to the lay meaning of the word or to try to redefine them.

I am not aware of any general meaning of the word "accuracy" in science. Not like the word "theory." There may be specific sub-fields that define "accuracy", but no general accepted definition. If you think there is, define "accuracy" for me.

 

Okay, I looked this up, and sure enough there is a use of the word "accuracy" in order to distinguish it from "precision." I don't think anyone here is really using these two terms in order to differentiate their meaning.

 

As far as equivocation, it is everyone on this forum who thinks equipment that measures well presents the recording in a truthfully who is equivocating.

 

First, you use the term "accuracy" divorced from perception.

 

Then, you claim that you want accurate equipment because it, in some way (not sure what words you would use) presents the sound "on the master tape" or whatever.

 

That is equivocation because you are jumping from mathematical accuracy to perceptual accuracy.

Mike, keep in mind that there's also applied science in addition to theoretical science. Applied science often catches the blind spots of those who work on theory. The reverse is true, as well.

We're not talking about theory here. As far as I know, there's zero theoretical support for audio cables. Odd thing is that there is theory for other types of cables, where you're dealing with power transmissions, extremely high frequency applications, and many others. So this has been studied, but nothing we know applies to aftermarket audio cables.

So the best approach is an applied one. Hands on. You have to put all of the possibilities on the table. One possibility is that there are undetectable physical differences. Another possibility is that the differences are entirely imagined by humans. It is intellectually dishonest not to consider psychology. It is very real and has to be tested for.

The sum of evidence so far suggests that there are no physical differences in audio cables and that human psychology explains perceived differences. Further, this squares with everything known about physics and human psychology. I think that psychological testing is the most important route with cable testing. I wish a research institution would take it up with a well-funded study. Far from "hate" or playing "gotcha," it would yield important study into how people think and percieve sound. There are lots of irrational beliefs - maybe this would unlock some understanding.

The difference between the two camps here is that skeptics are willing to put psychology on the table while believers completely rule out the possibility that beliefs and emotions affect what they hear. But you have to consider that, or else you're not looking at every potential answer. I think that cables will be conclusively proven not to have any effect. However, it won't be electrical testing that settles it, but rather testing the listeners.

What do you mean by perceptual accuracy, that seems like a contradiction in terms ?

 

If one person descibes a component as "accurate" and another describes it as "inaccurate" we have a problem - who is right - both perceive it as having some level of accuracy either a lot or nearly none. The only conclusion is that the term in this use does not help us here at all as we have no unequivocable measure from human perceptions. For instance Mike Fremer once described a "cable" as highly accurate, when measured it was anything but accurate - the perception of accuracy was wholly wrong - the cable was highly distorting, noisy and had an unflat FR - by no rational criteria could it be described as accurate, nor could it by any stretch be described as high fidelity. The accuracy was a perceptual illusion and this misleading.

 

However the measurements described in great detail the precise limits of the cable.

 

Of course anybody is free to prefer crap performing kit, that is a given, but being misled to think that something is accurate when it clearly is not is a real failure of perception.

 

Given the choice I'll take the measurements over opinions.

 

Edited by nick_charles - 8/31/10 at 5:13pm

In the context of audio reproduction accuracy refers to the degree to which the output signal resembles the input signal.

 

You appear to be defining accuracy as fidelity to the signal before it was recorded.  That could be defined as a type of accuracy if you added an appropriate qualifier.  I can't think of one off the top of my head.  What you are asking for is impossible in one sense and possible in another.

 

Information is lost in the process of recording.  It can't be retrieved.  It is forever lost.  You seem to be saying (correct me if I'm wrong) that you enjoy particular kinds of distortion which make the signal seem more 'live' to you.  To a first approximation such a thing is impossible.  (It's actually just really really improbable.)  You may like a particular type of distortion and think it makes the signal better, to your ears at least, but it does not make it more accurate.  It makes it less accurate.  There's nothing wrong with liking it.  Just don't confuse euphonics with accuracy.

 

The happenstance which causes a particular type of distortion in a particular piece of equipment has the probability of adding the same information which was was lost during recording and mastering of a piece of music is in the same ballpark as all the atoms in the arm of a marble statue happening to move in just the right way as to physically wave at you.  In theory, a very sophisticated DSP which took into account the physical locations of individual instruments and the peculiarities of a studio's acoustics and the microphones with which a performance was recorded could reproduce the lost information to a degree which would render the differences inaudible to a human.  In reality it would appear that the post production which most music undergoes renders this impossible.  No DAC, amp, cable, output transducer, nor any other kind of tweak is capable of such a feat.

Easy - device A is more accurate, (or more transparent).  Both THD measurements are below the audible threshold of perception, whereas a difference of 2 or 3dB is not.  Measurements such as THD, FQ response, phase distortion etc, are meaningful in many ways, but understanding the limits of human hearing provides us with further data.  I would take device A any day.

 

Precision has to do with how small a measurement you can make and how fine the resolution of your test is.

 

A ruler is a good example.  How fine the divisions are is the precision of the ruler.  One with marks every 1/16th of an inch is more precise than one that is only marked every 1/8th of an inch.  I you are measuring something that is 1 1/8" long then it doesn't matter which one you use and your accuracy of measurement is the same with either ruler.  If you are measuring something that is 1 1/16" long than the more precise ruler will give you a more accurate result.

You're actually wrong. As I suspected, those who clamor so loudly that I am using the "accuracy" wrong don't even know about the sub-field of science called "Measurement Science" where the term appears in a scientific context.

http://en.wikipedia.org/wiki/Accuracy_and_precision

Aha! So you had to refer to a model of perception in order to answer. Which demonstrates my point: accuracy outside the context of perception is useless.

 

 

 

Only if you think "accuracy" is an inherently empirical term. Let me copy my post from the other thread here:

 

 

 

 

 

Yes, perceived accuracy is subjective. That does complicate things. In my opinion what you are doing is retreating from a complex situation. "Oh no! It's complicated! Well let's find something simpler to work on!" You don't want to face the reality that the only useful measure of reproduction in audio is how it is perceived.

 

 

 

 

This is tautological reasoning. Your conclusion that it was an "illusion" follows from your definitions of the terms, not from any truthfulness. You define the cable as inaccurate because it measures poorly, you make an unstated assumption (that measured and perceptual accuracy should correlate), and it follows directly from your definitions that his perception is an "illusion."

 

 

This is odd. Audio exists for the experience of it, not for the measurements. In one sense it is strange, but I also see it as the natural choice for someone who is retreating from the complexity of correlating measurements to perception.

So I'm wrong because someone on wikipedia thought up a better analogy than mine?

 

 

 

Yes, I see audio science as applied science---making use of ideas from science and engineering, and balancing that will subjective evaluations of equipment. The important thing is not to think the measurements are the ultimate arbiter.

 

 

You keep bringing everything back to cables. I'm not talking about cables exclusively.

 

Let's talk about speakers. No one disagrees they have audible differences. So we can get a bunch of measurements, but we still have the problem of correlating those to perception.

 

There are more opinions than that. I believe that perception is unreliable, but I don't make the mistake of retreating to a "safe" empirical position because of that fact.

You're wrong because you defined it wrongly. Your explanation was fundamentally incorrect. Odd that you think that wikipedia page is making "analogies." EDIT: I just noticed the wikipedia page uses the word "analogy", but they really meant to say "example". Their "bullseye" example is an exact measurement situation to which the terms directly apply, so it is an example.

 

Your attempt to define it was not an analogy either. It was a definition and an example.

 

Edited by mike1127 - 8/31/10 at 6:09pm