Humans, on the whole, are pretty suggestible creatures. A friend won his local science fair by showing that people had difficulty identifying Kool-Aid flavors when they either couldn’t see the color of the liquid or were served a dyed version. A similar test gave samples of white and red wine to experts for review. The testers reported vastly different flavors, despite the fact that the red was just a food-colored version of the same white wine. Blind-folded, they would have reported identical flavors.
There are a number of psychoacoustic effects that can influence the way our brain perceives audio, on top of (and in addition to) our general suggestible nature. A basic example would be the Equal Loudness Contours (or Fletcher-Munson Curves), which show us how our tonal perceptions can shift with simple level changes. In a more extreme case, I read an article by a gentleman who claimed that he could hear a difference between two bit-for-bit identical copies of the same audio file. (No comment.)
When we need to make a call on something, it pays to approach the situation with a dose of humility, as well as awareness that our preconceived notions can conflict with what’s really happening. Rather than form one corner of what I call the “judgment triangle,” I try to base my decisions upon my knowledge and experience, and check that against theoretical predictions (second corner) and empirical evidence (third corner). If all three agree, I’m on the right track, and if there’s disparity, I might be missing something.
For example, if the claim is “96 kHz audio sounds better than 48 kHz audio,” I’d first think of the experiences I’ve had that support or contradict that claim. It’s also important to be specific here: What do we mean by better?
The question from the theoretical corner would be something like, “What does sampling theory predict would be the result of a doubling sample rate?” This is the part of the process where we live up to the titles of “engineer” and “technician.”
The empirical corner might ask, “In a scientifically valid environment, can listeners accurately identify 96 kHz and 48 kHz samples? If so, how consistently? 100 percent of the time? 50 percent?”
I’m not getting into the answer to these questions, because we’re talking about the process here, not this specific example, although I will say the results surprised me. Dr. Floyd E. Toole’s “Sound Reproduction” (Focal Press, 2013) does describe a number of these experiments, but more relevant to our discussion here, Dr. Toole also discusses a number of the powerful psychoacoustic effects that can influence our perception. Since none of us is immune to them, it would be prudent to be aware of their existence.
What’s my point? Our brains are powerful instruments but not infallible. I know I’m not the only one who’s made an EQ adjustment until I heard the change I wanted, then looked down and realized I was operating a different channel. I would have sworn I’d heard a change!
I know now that in similar situations, there’s a chance that I might be mistaken, and that’s OK. That’s how we learn. If I can learn, I’ll be better at my job, which is good for everyone.
So I say, so ahead — dazzle me. I’d like to learn something, just don’t get too upset with me if I don’t believe you right away.