wavoman
Headphoneus Supremus
- Joined
- Jan 19, 2008
- Posts
- 1,873
- Likes
- 45
Smelly -- I agree with your point that many of the papers published in AES do not use statistics properly, but worse: the correct statistical papers (and the ones by statisticians) are focused on the A/B/X testing protocol, which has a lot of flaws in its own right.
You know my credentials: PhD in Statistics from Yale, A.B. in Statistics summa cum Laude from Princeton, winner of The American Statistical Association Theory and Methods Prize, co-author of the now classic nonparametric multivariate two-sample test (with thousands of citations). But this is all well-worn ground here ... many posts and threads. What's the point? Audio testing is very hard; very few good studies have been done. We all know that.
But you also know the basic answer -- it is obvious that there are NOT large, audible differences in cables, DACs, or loudspeaker AMPS (above the total junk level) that can be perceived by the population at large. If there were, it would have been revealed by now, and just the opposite is true.
This is settled. What is not settled is the more subtle question: can a well-trained, golden-ear'd audiophile reliably and blindly perceive these differences, in real-world, relaxed home listening situations, and are the differences large enough to merit spending money (some subjectivity in this last part).
Tough question.
You know my credentials: PhD in Statistics from Yale, A.B. in Statistics summa cum Laude from Princeton, winner of The American Statistical Association Theory and Methods Prize, co-author of the now classic nonparametric multivariate two-sample test (with thousands of citations). But this is all well-worn ground here ... many posts and threads. What's the point? Audio testing is very hard; very few good studies have been done. We all know that.
But you also know the basic answer -- it is obvious that there are NOT large, audible differences in cables, DACs, or loudspeaker AMPS (above the total junk level) that can be perceived by the population at large. If there were, it would have been revealed by now, and just the opposite is true.
This is settled. What is not settled is the more subtle question: can a well-trained, golden-ear'd audiophile reliably and blindly perceive these differences, in real-world, relaxed home listening situations, and are the differences large enough to merit spending money (some subjectivity in this last part).
Tough question.