View Single Post
  #60   Report Post  
 
Posts: n/a
Default THE ENDLESS DBT DEBATE

Mkuller wrote:

While DBTs are effectively utilized in psychoacoustic research, there is little
or no evidence that they are appropriate or useful for audiophiles to utilize
in comparing audio components with music as a bias control method. In fact,
the few reported published studies show that when used in this way, DBTs do not
show subtle audible differences between components, but only gross frequency
response and loudness differences and then only when pink noise is used as a
source.


A correlation between audible differences and measured differences when both
are under controlled conditions is very powerful evidence indeed. Such
correlations are lacking in the subjectivist evidence. That there is a
difference between the results of controlled tests and sighted tests is
something that is expected and in fact should be. It would be sad indeed if
people had no auditory imagination --- indeed it would be impossible to
produce music and build great instruments if that was the case.


There are at least two important elements missing from these amateur DBTs:
1. Pretesting that the the actual subtle differences can be identified with
the program material utilized, i.e. is the music selected actually a sensitive
enough source to identify, say a difference in midrange dynamic contrasts in a
DBT. That a DBT is sensitive to the limits of audibility is meaningless if
that does not apply to THIS DBT.
2. Pre-training the subjects to listen for the specific differences
(midrange dynamic contrasts) prior to conducting the DBT.


I believe these two protocols are standard in psychoacoustic research, but have
not been applied to any amateur DBT I have seen. Certainly there are other
problems with amateur pseudo-scientific use of DBTs in audio as a method of
bias control. The bottom line is this - they have never been proven as
effective for use in comparing audio components with music in the way they are
blindly advocated on RAHE. Sighted listening, i.e. no bias controls, seems
more effective in identifying subtle differences.


I'm not going to hash over all the usual stuff. I'm weary of arguing about
it and nothing will likely change if I did. But you do make your point clearly
and succinctly, even if it no longer makes any sense to me. (I used to be a
subjectivist) Mr. Mirabel could hopefully learn something from your writing style.