I've thought about that too. And there's some truth to it, if you're predisposed into thinking there's no change, it's more likely you won't hear a change. That's why some scientific method needs to be involved.
On the forum for the music software I use, a really smart guy (this is me saying that.
), was schooling some people on sampling rates. Saying there's no audible difference between 48 and 96 kHz or higher. I agreed that any content above 24 kHz is useless, but made the mistake of thinking that 96 kHz can more accurately describe lower frequency content too.
This wasn't just a gut instinct, I had done the math, and was sure I had even heard a difference in the very software the forum was dedicated to discussing. So I output two wave files, each reproducing the same signal (three slightly detuned saw waves around 10 kHz), and tested them with the ABX Comparator plugin for Foobar.
OK, the point of this story. A lot of people don't understand the purpose of, or how an ABX trial works. First, it's purpose isn't to find out which of two things is better. It's only to find if there is a difference between them. Performing the test is very easy, on the other hand setting up the test to be truly double-blind can be incredible difficult. So in my case I loaded the two files, and told Foobar to test them on me. It randomly selects one of the files, that'll end up being X. Then I'm freely able to play A and B, the two files being tested. I just have to decide which one is X. I did 10 trials, and didn't look at the results until the end. Boom! 10 out of 10. There absolutely was a very different sound between the two, and I could match one to the other every single time.*
I've heard people say ABX trials are stressful, and they can't focus well enough for the test to be accurate. But in the case of amps, these same people say there is a huge difference between two amps. Night and day, right? If the differences are so great, that you can remember after minutes of silence while connections are changed, it should be no problem to identify one vs. the other when you can switch instantly. That's how it was in my test, it usually went this way, "play A, play B, play X, I think X is B, play B, play X, pick X is B, next trial." Toward the end I was so used to the differences, I'd go, "play A, play X, pick X is B." It comes down to, if the differences are great, the trial is easy, but if the two samples are so alike that even is straining and stressing, and wearing yourself down with the time spent on each iteration of the trial, you're basically left to guess which is which...you're guessing, and the samples are the same within your detectable limits.
*How could I tell the difference between a 96 kHz file and a 48 Khz file so easily? That's another story.
