I remember reading that Paul Klipsch hired people he called "golden ears" to test his speakers on. These people were typically not audiophiles or musicians or recording engineers, just people who demonstrated an unusual ability to hear things that others couldn’t. Of course they scored great on standardized hearing tests, but they also could do things like properly equalize an intentionally imbalanced signal very quickly with a high degree of accuracy by ear, or detect distortion at unusually low levels.
I like your "open to bias" interpretation of "subjective." It’s hard to be biased about certain facts that are plain to almost everyone, like the number of drivers on a speaker. One of the useful things about science is that it can take something that is open to bias and make it plain through alternate methods of observation. When a person claims they can hear the difference between two devices but only when they know what they are listening to from visual and other sensory input, it’s difficult to prove them wrong. You can trick them by switching components when they don’t know but for whatever reason even when you let them know they’ve been tricked it fails to be compelling evidence to them because when they’ve checked and are sure about what they’re listening to the perception comes back to them. It can be very hard to overcome perceptions, which is why it takes some considerable training to learn to safely fly a small airplane on gyroscopic instruments when there’s zero visibility out the windows. Proprioception doesn’t match what the little gages and dials are showing my eyes.