"j." wrote in message
ps.com...
I've looked around a bit for this, and it seems it must have been
mentioned before - but I think there's a pretty compelling argument as
to why double blind tests aren't always the be all end all. (yeah,
this could get bad - I'm no expert, this is just a thought so be easy
on me)

Using DBTs we can determine what the smallest difference a human is
able to detect is. For example, slight changes in pitch or volume.
There is a level that is small enough that the person is unable to
detect the difference, but at twice the change the person can tell the
difference.

So lets say that no one is able to reliably tell the difference between
speaker cable A and speaker cable B in a DBT. Lets also say that no
one is able to tell the difference between amp A and amp B in a DBT.

...but what if enough of these things added together does produce a
perceptible difference? (Obviously this doesn't really blow away the
DBT argument - its more that it shows that the way we do the tests is
not adequate). I mean, I doubt they are out there doing double blind
tests with hundreds of permutations of high end audio gear all the time
- the cost would be amazing.

What you say is quite true. It's definitely possible to conceive of a test
where a single change is inaudible, but a combination of 2 or more changes
at once IS audible. But I fail to see how that burdens DBT specifically.
It would seem to apply to any test protocol.

When you're musing over objections to DBT, always ask yourself "compared to
what?"

Norm Strong