opaqueice wrote:
>  ABX measures whether a difference is audible, or not. 

Technically, ABX gathers data points on whether the humans think they
detect an audible difference. You then apply statistics to compare the
results against random values. So you can say "we found X percent of the
 humans reported a difference, which with the sample size, lets us
reject that its random with Y confidence.

Usually Y has to be 95% or so before its considered science.

There is always a chance, in this case 5% that there was no difference,
or that the results were random.

Doing ABX as opposed to AB testing just allows us to make the same
statement with fewer data points.


-- 
Pat Farrell
http://www.pfarrell.com/

_______________________________________________
audiophiles mailing list
[email protected]
http://lists.slimdevices.com/lists/listinfo/audiophiles

Reply via email to