The Placebo Effect
Aug 28, 2008 at 2:40 PM Post #94 of 101
Quote:

Originally Posted by Know Talent /img/forum/go_quote.gif
what I find interesting is the Jitter snake oil... jitter is measureable but generally found to be inaudible under a threshold of roughly 250ps (or is it ppm???).
So with jitter, you have something that can be measured but manufacturers still claim audible improvements when lowering jitter below already inaudible thresholds.



Depending on whose research (as opposed to web anecdotes) you read the best evidence so far is that jitter is inaudible below either ~ 20ns for signal correlated jitter (Benjamin and Gannon, 1998) or ~ 250ns for random jitter (Ashiar et at, 2004) when there is a musical signal present.

The biggest problem with the manufacturers of jitter solutions is however that they rarely[1] actually measure jitter before and after their solutions are implemented so for most of them you have to take it on trust that jitter has actually been lowered at all let alone know by how much. Also none so far , to the best of my knowledge anyway, have backed their claims up with double blind listening tests.

[1] Paradoxically the manufacturer of a new low jitter clock design did some tests pitting their design against existing models they found that their ultra low jitter clock (they did measure the jitter) was not universally preferred over the higher jitter competitors.

But there is more. If you read Sterephile's reviews of CD players and also the measurements that they take you will find it impossible to make any kind of correlation between measured jitter (if used as a single indicator) and subjective quality.
 
Aug 31, 2008 at 7:04 PM Post #95 of 101
I recently did a compare between the headphone outs of my CDP and integrated amp against my GS-1, the results of which can be found here:

http://www.head-fi.org/forums/f5/onk...9555-a-357405/

While I was able to hear differences in the bass (which the GS-1 handled better) those differences were very minor. What I did note though was that overall, the GS-1 provided a more pleasurable listening experience. Given the fact "pleasurable" can't be measured, I do have to consider that placebo came into play. It's very easy to not want to believe in placebo, especially when money is involved. At the same time though, any number of factors can produce a listening experience that goes against what you might expect from a scientific point of view.
 
Sep 1, 2008 at 5:16 PM Post #97 of 101
---------Read my Signature Quote -------------
 
Sep 3, 2008 at 5:43 PM Post #98 of 101
Quote:

Originally Posted by dvessel /img/forum/go_quote.gif
Some of this may be obvious but here it is anyway..

The term is formally used in Medicine. It's used in studies by having two groups to test new medications. The control group to test for the placebo and the ones getting the real treatment. Each group is lead to believe that they are getting the real thing. This is because simply believing in the treatment can often lead the body react on its own. The two groups are compared to see if the medication is having any real effect.

The way it's used in audio, it's believing something is there when it is really not. This can apply to anything and everybody has experienced this at some point. It's transparent. You will never know since it's not a conscious effort.

How to prevent it? I'm not sure you can unless you have the right mindset. Simply being aware of it can help. I always try to play devils advocate. Weight both sides. Polar thinking to minimize taking of sides prematurely.



Yes, the mind is very powerful..
 
Nov 7, 2009 at 5:59 AM Post #101 of 101
For suggestion and headphones:

If one wants to get at the bottom of this issue, we need something along the lines of a real study, with real controls. Among the considerations:

1. Compare several types of headphones, new and broken in.
2. Use a top-quality, high-resolution system or setup.
3. Try to obtain two types in data or feedback: the reactions of listeners, and hard measurements (if possible).
4. Double-blind measurements would be ideal, as would comparisons using appropriate statistical tests.

There's nothing to fear with a good study done skillfully. The problem is, a satisfactory study is going to be expensive. Moreover, one should expect a response from a similar experimenter who wants to cross-validate the study. In short, anyone expecting instant validation is in for a rude awakening.

I also agree with Hirsch with his comment: "placebo" is a loose, generic term which would best be served by "expectancy." Inaccurate use of terms only complicates issues.

Frankly, I have enough experience with this kind of experimentation to predict the results. To get significant results, one needs results that are accurate to the <.05 level. A lot of listeners may not have the ability to discern fine differences in sound between headphones, or may personally feel they're insignificant by their own personal standards.
 

Users who are viewing this thread

Back
Top