We did. At least definitively tried.
Audyssey defeated our original test, when it turned itself on after a power cycle, unknown to the setup person.
We repeated the test the next day on a smaller scale and came to realize with several amps tested, including a t-amp that in fast blind switching no one could tell the difference. Even in full range use. (Double checked -voltage matched with a multimeter and spl level matched with omnimics spl meter)
East coast guys were going to perform the test too to confirm results, but lost interest when they (typical hosts) came to the same conclusion in their own experimentation.
Looses appeal to listen to the exact same thing over and over with no discenable difference between entries.
I'd like to repeat the test on a large scale again some time to confirm the results with a lot of people to put the issue to bed. But its a lot of setup work and the fast switching mechanism can fry an amp with user error. Which happened twice at the g2g (not at my command either time - but it smoked my little t-amp, which was the only permanent casualty.)
Not sure what outside matters as compared to inside. Unless you are saying play it louder at which point no one would disagree that the t-amp will run out of steam first followed by the next least powerful amp and so on and you can tell an amp appart if one is clipping and another isn't, but our test intention was meant to keep the amps within their means and not test output, but sound quality.