Is learning how to listen/training our ears important or valid? After all, if differences are real you shouldn't have to learn how to spot them, right? It should be instinctive, shouldn't it.
To address this I'm going to step back 45 years, leave the field of music reproduction, and take a look at hearing from a different viewpoint. Back then I was an Army ROTC cadet in a Ranger training company with two senior NCO Green Berets (Special Forces) with a minimum of 5 combat tours on A teams in Vietnam each. On the first field training exercise, they took us out in the woods, sat us down and told us to close our eyes and listen. They then had us describe what we heard. Then they brought our attention to things they heard but we didn't. Like the episode in the TV series Cung Fu. They then told us the first thing you did when getting to a new area was sit down and diligently listen to the sounds around you. The reasoning was that by learning what the surroundings sounded like in a normal situation, you would be able to notice when something changed. This is important because when the normal sounds change, something new has been introduced into the are and this could mean danger.
Once you learned this skill, you would realize something had changed in the environment around you even if you weren't consciously aware of what it was. Which is a survival mechanism. These two Green Berets were exceptionally good at it. When I ask one of them how he did it, he replied that when your life depended on it, you learned.
The point is that while the ability to hear differences can be improved and refined through practice. It has nothing to do with "golden ears" (used in derision when subjective attributes are discussed). Just attention to detail.
Back to music reproduction and science.
A scientific experiment is very limited in scope and as such, the conclusions drawn from that experiment are not generalizeable beyond the parameters of that experiment. This means no single experiment can provide definitive answers across a wide range of applications. It takes many experiments, each looking at different parts of the larger question, and even more seeing if previous results can be replicated, before widely generalizeable conclusions can be drawn. One double blind experiment conducted under one set of circumstances is not capable of providing a definitive answer. Especially if people thing they will be able to draw accurate conclusions from a recording of the experiment, transmitted via YouTube and played back on their equipment in their room.
Whether a double blind experiment is the most appropriate design to address the question of perception is another topic altogether. Before anybody starts hollering about a gold standard, while a double blind placebo controlled trial is considered the gold standard for determining cause and effect, it is not the most appropriate design for every type of question.