[aimlink]

Quote:

There are many published studies utilizing within-subject designs where a patient can receive several treatments (drug and placebo).  Within-subject designs also have dedicated statistical analyses that (by eliminating between-subject variance) are quite powerful and widely utilized.

 
See Maverickronin, as I say, I'd never pretend to be an authority on any of this...even my own profession which is vast in its own right.
 
It's just a matter of what you wish to prove or which side you're on.  It never ends really.....

 

[mike1127]

 
 
Quote:

Math is nebulous?  ...  
An input signal can be measured to a degree of accuracy far beyond what our ears and brains can manage.  The same for the output signal.  Any differences between the 2 signals is distortion.  By definition. 
 

Okay so I do the differential measurement with two different devices. I plot the differentials, which are waveforms of about one minute duration, or, if we are sampling at 44.1 KHz, about 2.7 million numbers each. I hand two lists of 2.7 million numbers to you and ask which one has more distortion. How do you answer?
 
Quote:

You can't have a conversation with someone who insists on redefining random words in you vocabularly.

There's nothing inherently empirical (numerical) about the word "accuracy". See my post on the thread "a misconception about measurements"
http://www.head-fi.org/forum/thread/509814/a-misconception-about-measurements/15
 
 
 
Quote:

You're trying to define 'accurate' as 'pleasant to my ears.' 

Not sure where you get this idea. I'm defining an accurate system as one that accurately reproduces a reference experience of music. If the reference experience is unpleasant, than an accurate reproduction will be as well.

 

[Shike]

Quote:

 
 
Okay so I do the differential measurement with two different devices. I plot the differentials, which are waveforms of about one minute duration, or, if we are sampling at 44.1 KHz, about 2.7 million numbers each. I hand two lists of 2.7 million numbers to you and ask which one has more distortion. How do you answer?

I'd say you're doing the test wrong so to speak.  You need a total of three devices, one of which will be the reference (amplifier producing a specific signal if you will).  You then measure the other two devices (transducers for example), then do a a waveform analysis against a transducer and the amp (not transducer and transducer).  Whichever deviates least from the signal produces has lower distortion.
 
 
Quote:

There's nothing inherently empirical (numerical) about the word "accuracy". See my post on the thread "a misconception about measurements"

 
Just because it can used subjectively in a conversation does not mean it doesn't have another context in regards to the scientific community.
 
 
Quote:

Not sure where you get this idea. I'm defining an accurate system as one that accurately reproduces a reference experience of music. If the reference experience is unpleasant, than an accurate reproduction will be as well.

 
You're never going to get the "reference experience" from a recording that doesn't have all the information anyway.  What you're doing is taking a final mastered product and adding and changing things to make it sound what you think it should sound like, but without a true point of reference you can't claim any level of numeric accuracy which is what's being discussed in sound science.  I'd argue you can't even claim live accuracy unless you've listened to all music you play on your system live at the day, time, and place it was recorded.

 

[wavoman]

Quote:

There are many published studies utilizing within-subject designs where a patient can receive several treatments (drug and placebo).  Within-subject designs also have dedicated statistical analyses that (by eliminating between-subject variance) are quite powerful and widely utilized.

Right on!  Also called "SS1" or "single-subject designs".  Sometimes the order in which the treatments are given matters, and then the analysis is "repeated measures design" or "individual as block" designs.  Any modern text on experimental design covers these.
 
But there's more.  A drug treatment may very well have a real and (somewhat) lasting effect on the subject.  That's a problem in repeated measures design.  Not the case with audio testing, although there are audio memory issues.  There are also ethical issues in giving sick people "placebo" instead of "drug", so some designs try to minimize this.  Not the case with audio testing.
 
The measurement in drug trials is not as subjective on the part of the subject generally (sometimes it is, as in "are you feeling less suicidal now?"), while in audio it is 100% subjective and subject to wicked bias.  That's why false comparisons (swindles) of A/A and B/B are so powerful and important in audio ... they smoke out some types of response bias:
 
-- I love to show off, I want to help the tester, I tend to find differences even when they don't exist
-- I always love the second thing I hear ... it's better, right?
 
Having a known clunker C is also very powerful, smokes out:
 
-- I hate this nonsense, and the test leader is an ugly jerk, I ain't reporting no difference no how no matter what I hear
 
These are powerful biases, typically unrealized by the subject himself!
 
Wait, there's more!
 
The point of a drug trial generally is to establish a base level of efficacy in the population so the drug can be released.  In audio, unless you are a manufacturer or a magazine doing research, the population effect is less interesting.  Pooling the results of SS1 is not important to me.  I only want to know what I can hear and what I can't.
 
Borrow gear, use audition and return, buy-and-resell, go to meets ...  do your own blind testing, typically single not double blind with an unbiased friend doing the gear switch, throw in false A/A and B/B comparisons, rotate in a known piece of junk C if you can ... use everything else (physical measurements, reviews, group blind tests) only as a rough guide.  Be aware that non-blind testing will ALWAYS have the placebo effect on you no matter how hard you try. 
 
Enjoy the music and don't upgrade so much.

 

[sohels]

Quote:

 
 
Okay so I do the differential measurement with two different devices. I plot the differentials, which are waveforms of about one minute duration, or, if we are sampling at 44.1 KHz, about 2.7 million numbers each. I hand two lists of 2.7 million numbers to you and ask which one has more distortion. How do you answer?

I'm sorry, but you're just being obtuse here. Reading this may help: http://en.wikipedia.org/wiki/Total_harmonic_distortion

 

[wavoman]

Quote:

With respect to advertising and marketing claims, it doesn't matter much if the marketer "believes" in their product; rather, the basic measure is whether they have a reasonable basis for their claims. A subjective belief, without more, generally is not sufficient, but this also is heavily dependent on the claim being made..... The honest belief probably gets one out of the fraud issue.

It's also important to remember that advertisers are responsible for all reasonable interpretations of their advertising claims. So if an advertiser implies something is proven, that advertiser would be well advised to have a reasonable basis upon which to make that claim, not just a belief.

We agree, no fraud.  As to basis of belief, these folks will cite all types of "experiments" they have done, lots of (bogus to you and me) "evidence" -- way more than subjective opinion.  I'm think they would pass the "reasonable" test (indeed, these people are reasonable!).
 
Manufacturers less so -- many seem to fall over the fraud line.  But individual dealers, even those that deep freeze their own cables, and trainers, especially trainers-to-the-stars ... true believers, accurate researchers and "experiment leaders" (in their own minds) ... really believe what they say and think they have a firm empirical basis ... not criminals.  I won't mention names, but there are some respected electronics designers here who are just as misguided about what they measure and build, making junk science claims a-plenty.  But they believe! 
 

 

[Prog Rock Man]

Rationalising the evidence of blind testing and cable measurements with the very credible reports that cables do sound different is IMHO at the heart of the debate. Aimlink has made some cracking points

 
Pro-cablers at present have a strong case. So many report hearing differences, audiophile review magazines mainly support them as does a good part of the audio industry. Then blind testing is an easy target for criticism and finally there are not enough published to be totally convincing.
 
Anti-cablers also have a strong case, as measurements do not support or correlate to audible differences, there are explainable reasons why people report differences due to placebo and psychoacoustics and the blind testing conducted so far does support the claim there is no difference.
 
I have rationalised the two positions as from my own reading and experience blind testing shows no audible difference and placebo and psychoacoustics explain why I can then hear differences when I am sighted again. (Things get more complicated again as my mood in particular, but also the time of day throws in other variables as to how much I enjoy my music. That causes 'night and day'

differences to the sound!)
 

 

[Uncle Erik]

I don't give much weight to the audiophile rags. There's the well-known argument about advertising revenue. But I also find it curious that "Stereophile" runs extensive tests and measurements for sources, amps and transducers. They have the capability to run performance measurements on cables, but never do so. Instead, they only run completely subjective cable reviews. Why? I bought my SACD player in part because of the strong measurements it produced. Likewise, I look at their measurements for other gear and find them very useful. Why won't they quantify cables?

That some people believe strongly in something is evidence of nothing. Look at all the medical quack "cures." Some believe that crystals heal. Others base their actions on tarot cards, astrology and palm readings. Or run a Google search on radon health mines - people inhale radon as a cure, even today. Then there are those who perform rituals and use "lucky" tokens when gambling. Gambling is pretty easy to nail down with mathematics: the house always wins in the long run. You cannot use mass belief as evidence. The best evidence against this is that there is mass belief in conflicting belief systems. I won't touch on one of the verboten topics here, but one of them involves very strong beliefs in entirely contradictory belief systems. You can argue the merits of one belief system against nonbelief, but when you compare belief system X against belief system Y, you find them in partial or complete conflict. Therefore, if X is true then Y must be false and vice versa. The often unconsidered point is that both belief system X and Y might be wrong.

This also directly applies to cables. Cables here are discussed in the abstract, whether they work or not. But when you get down to it, there's a wide variety of belief systems in cables, each manufacturer has its own philosophy (for lack of a better term) on design. If you look at these various philosophies, they are very much in conflict. If manufacturer X is right, then manufacturer Y's cables are totally bogus, and vice versa. Now, all of the manufacturers cannot be correct. If the claims are taken at face value, they obviously conflict. Weirdly, even contradictory designs are claimed to have the same benefits. How can that be? Assuming that one of them is correct, the other would be wrong. But people report the same improvements from conflicting designs. The only conclusion I can reach is that the "differences" are entirely psychological. Otherwise, one design would produce results while the conflicting design wouldn't.

 

[Catharsis]

To the OP:
 
That first post (and other posts) of yours was incredibly well written and conveys your thoughts - and mine - with precision. 
 

 

[Catharsis]

 
Quote:

 
But on the other hand when one brings it up the threads tend to end up being a circular argument about the validity of the reported tests or even the validity of DBT itself, which gets really boring very fast, so to avoid that I'm kind of glad it is not permitted, and if I'm not mistaken that is the main reason it is not allowed.

That's why people who are against DBT should be banned from discussing their problems with a sound scientific method. 

 

[Prog Rock Man]

 
Quote:

 
 
........
 
Again. I agree wholeheartedly. Amps and sources are much much easier to test in blind conditions and doing so could put a nice ceiling on the extent of the effects being described. 'Under sighted conditions I notice the bass being tighter going from a Compass to a Phoenix but I could not pick the two apart under blind conditions.' What an extremely valuable piece of information to have. It does not disprove the sighted observation but it shows that the difference is not so strong as to survive the change in listening conditions.

Superb point.

 

[sokolov91]

So,
 
instead of fussing over cables and such, we should all just dim the lights, so as not to see our system and use it for what it is meant for... to listen to music... and not our gear
 
We are now saving boatloads of money on cables, and some on electricity too!
 
Talk about a win win situation!

 

[maverickronin]

Quote:

See Maverickronin, as I say, I'd never pretend to be an authority on any of this...even my own profession which is vast in its own right.
 
It's just a matter of what you wish to prove or which side you're on.  It never ends really.....

Didn't I say it would be done if it could be done?  I was unaware that there were circumstances in which this was possible.

 
Quote:

 
There's nothing inherently empirical (numerical) about the word "accuracy". See my post on the thread "a misconception about measurements"
http://www.head-fi.org/forum/thread/509814/a-misconception-about-measurements/15

Yes there is.  Whether you like it or not, information can be quantified.  The differences between two sets of information can be quantified.

 

[mike1127]

Quote:

Quote:

 
 
Okay so I do the differential measurement with two different devices. I plot the differentials, which are waveforms of about one minute duration, or, if we are sampling at 44.1 KHz, about 2.7 million numbers each. I hand two lists of 2.7 million numbers to you and ask which one has more distortion. How do you answer?

I'd say you're doing the test wrong so to speak.  You need a total of three devices, one of which will be the reference (amplifier producing a specific signal if you will).  You then measure the other two devices (transducers for example), then do a a waveform analysis against a transducer and the amp (not transducer and transducer).  Whichever deviates least from the signal produces has lower distortion.

 
What I wrote was fine. I was talking about comparing input to output of each device. You want to compare output to output. Fine. The point is the same. You entirely skipped over the interesting part by writing "do the waveform analysis" and "deviates least from the signal". What waveform analysis? How do you calculate how much each waveform deviates from the signal?

 

[mike1127]

Quote:

I'm sorry, but you're just being obtuse here. Reading this may help: http://en.wikipedia.org/wiki/Total_harmonic_distortion

 
No, I'm making a specific point.
 
So you have two lists of 2.7 million numbers. I ask you: "Which one has more distortion?"

Total harmonic distortion is a way of collapsing a continuous waveform into a single number.
 
There are an infinity of ways of collapsing a list into a single number. To give just another example, do a fourier transform and calculate the maximum deviation of the frequency response over the range 20 to 20,000 Hz.
 
So let's say I tell you:
 
For device A, the THD is 0.1% and the deviation is +/- 2 dB
 
For device B, the THD is 0.05% and the deviation is +/- 3 dB.
 
Which is more accurate? Justify your answer by referring only to numerical models.