Head-Fi.org › Forums › Equipment Forums › Sound Science › Links to Double Blind Tests
New Posts  All Forums:Forum Nav:

Links to Double Blind Tests - Page 2

post #16 of 20
Quote:
Originally Posted by Pio2001 View Post
Only the one who got 100% correct made a significant score. The others are compatible with chance, given that they are picked among 145 different results.
Was precisely is compatible with chance? The number who could detect out of the large number of participants? Would it be compatible with chance that they all happened to be the ones using headphones?

I'm not trying to be argumentative; I'm truly interested. I'm really not very knowledgeable about statistics.
post #17 of 20
Quote:
Originally Posted by Pio2001 View Post
Only the one who got 100% correct made a significant score. The others are compatible with chance, given that they are picked among 145 different results.
Strictly speaking the headphones and surround sound settings should be regarded as different conditions, so really it is 4/30 not 4/145, the p values were < 0.05 for all 4
post #18 of 20
Quote:
Originally Posted by PhilS View Post
Was precisely is compatible with chance? The number who could detect out of the large number of participants?
The fact that three listening tests among 145 got a good score can be pure chance.

However, the score of the fourth one, who got 20/20, is so unlikely (one chance out of one million) that even after 145 tests, it remains a strong success.

Quote:
Originally Posted by PhilS View Post
Would it be compatible with chance that they all happened to be the ones using headphones?
Difficult to tell. There were other factors than headphones, like sex, age, occupation (related to music), or musical sample chosen. Paradoxically, the more we have information, the more likely we are to find one factor where scores are better. This is because our analysis is done after the raw results are published, and we are considering situations that were not supposed to be the subject of the statistical analysis to begin with.
To get proper results, a statistical model or simulation would have needed to be done before the test in order to include all these factors, and clearly define minimal scores for each of them.

Now, a good way to answer the question would be to perform audibility tests with the same speakers and headphones, and see if people hear higher frequencies with these headphones. If so, it would be a good indication that the use of headphones were a relevant factor in the success of those 4 listeners.
post #19 of 20
Quote:
Originally Posted by 883dave View Post
In this thread please provide links to "documented" double blind tests for audio equipment...transducers, amps, sources, cables and tweaks.



2. Matrix-Hifi: Pruebas ciegas Could someone translate?
I'll translate part of the Behringer vs Benchmark (DAC1) test, since that's what I think most would be interested in.

basically, they say that 10 out of 14 testers couldn't perceive any difference, and that the other 4 decided that they couldn't do the test but couldn't tell the difference between the two.

The conclusions section says that there is really no audible difference in any area between the two on the same source.
post #20 of 20
Quote:
Originally Posted by 883dave View Post
In this thread please provide links to "documented" double blind tests for audio equipment...transducers, amps, sources, cables and tweaks.
883dave, I think this thread is a great idea. May I suggest that you make this a "annotated bibiliography" - as in, each link would have a one-line description (or a brief summary even ebtter). This will allow people to read the references that look appealing or relevant to them, and also sort out which references are more authoritatve (i.e. peer-reviewed journal vs. some dude's test that he posted on an internet forum).
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › Links to Double Blind Tests