My two cents. I should state I am one of those middle people. I believe in both objectiveness and subjectiveness. I have been a regular on other audio forums going back to the usenet days, but don't have the intense distaste for Amir that some seem to have (and in fact have given a donation to this site to help support his continued testing). I am one of those people that believe that answers are not always what we think we are and am happy to support pretty much any research that helps develop further information.
All that said, responding to the original question. IMHO the answer is that both can be true. You can have a device measure objectively poorly, but sound great. Tube triode amps are an excellent example, they measure like crap, but have a wonderful lush sound. There is no doubt they are distorting the signal, but it is a pleasurable distortion for many.
In my main system I have a Schiit Gungnir MB Dac. It replaced a Bifrost Multibit, which in turn replaced a 15 year old Bel Canto DAC 1. There was a significant sound quality difference between the Bifrost and the Bel Canto. What I can't tell you is why. Was it because Schiit is that much better, or had components degraded in the Bel Canto over the past 15 years causing it to not sound as good as new. Objective measurements would be able to tell us instantly if the Bel Canto was working as designed. Oh, the reason I went from the Bifrost to Gungnir was the ability to run true balanced from the dac through my preamp to amp, all of which are true balanced designs, as I do believe the improvements in sound to noise ratios and balanced design emf rejection are well founded and worth using.
Objective measurements have a very important role in audio. They can describe a large amount of what we hear, but not everything. The more measurements we do, the more we learn, and the more we are able to correlate what we hear to the measurements. Just one example, seventh order distortion in amplifiers seems to have a disproportionate effect on what we hear as grain and hardness (per John Curl). There is no way that would have been discovered without careful measurement and correlation to what was being heard. Likewise, many audio engineers will tell you that they can change a component in a design, replacing it with one that is electrically identical, but yet the sound changes. This doesn't mean that the components are in fact identical, it just means we have identified or figured out what electrically is different between the two items and therefore don't know what to measure to identify and predict differences.
There is only one absolute truth in audio, if the user happy with what they have. Whether is it the teenager with a 128kbs phone stream and $2 headphones, or the person with bladders under their components, tiptoes for their cables, and holy water in the corner to dampen bad vibrations, if they are happy with hearing, that is what is important.