Originally Posted by elmoe
The reports I've read for every single ABX test so far have always been significantly less than that.
That is because the testers quickly learn that they get better scores with faster switching. Auditory memory fades quickly, and longer switching times have been proven to make the results less reliable. In a proper ABX/DBT, it makes a negative result more likely, while in casual sighted testing, it is easier to get a false positive (i.e. imaginary difference) if there is a longer time to forget what the other source sounded like. Which is why audiophiles believe longer switching time is better, because it "improves" their usual (flawed) testing method. It is also why "modding" audio equipment (like op amp, cable, etc. upgrades which inherently take some time to perform) is popular, and large differences are often heard with no objective evidence.
Note that fast switching ability does not mean that the sample length or the overall time spent on testing also has to be short; with a thorough comparison of a longer sample that has more different sounding parts, there may be a better chance of finding an audible artifact.
Edited by stv014 - 3/26/14 at 3:38am