The paper does not contradict your listening experience. The authors classify what you do as "trained listening", which is one of the categories they suggested, depending on the goals of a subjective listening test. Other categories include "easy listening" and "slow listening".
Excerpt from the paper:
With regard to subjective testing, pre-qualification must ensure subjects are entirely familiar with artifacts to be detected and reported when testing is based on immediate conscious responding; but it is indicated to generally be highly aware of:
- Listener experience,
- Listener attention,
- Listening duration.
Experience includes intimate familiarity with the features tested and the listening environment. Attention means allocating perceptual bandwidth fully to hearing and focused listening, or less.
yes, yes, but the real impetus behind this is encapsulated in a true (under)statement that appears a bit earlier in the paper:
"Using traditional subjective testing, it has proven difficult to argue clearly in favor of higher data-rates than 48 kHz/24 bit linear PCM per channel [12]. "
Certain people (including Thomas Lund) has been pushing the idea that we can 'hear' hi rez, for years. Since the experimental data in support of that have been so very sparse, hard-won, and disputable at best, now we get this review to prop the window open.
But this tack, as always, ignores the elephant in the room:
people who tend to say they 'hear' the benefits of hi rez, tend to say they 'hear' it without *ANY* of the elaborate training protocols this review proposes, including extended listening. So, are they 'really' hearing it, or not?Should we believe what such self-professed 'golden ears' (including hi rez cheerleaders in industry, in media...let's call them 'influencers') say about their (sighted, always sighted) perceptions?
Here's a radical idea: The way to test their claim is not to 'train' them (or worse, train someone other then them) to hear artifacts, etc. That's unnecessary. They already hear it, right? The thing to do is to get *them* to take a DBT using
their own usual listening conditions and musical choices...you know, the ones they claim are good enough to let them hear the difference 'just like that', and 'so do my wife/kids/neighbors who drop by'. The only difference being, now they don't know in advance whether A or B is hi rez.
Testing whether it's possible to *ever* hear 'hi rez' -- basically, finding conditions that reveal the true outer limits of human hearing -- is a *different thing* and it is the thing that psychoacoustics is actually interested in. But that's not quite the same as testing golden ear claims. Which is why even if Lund, Stuart, et al. finally, at long last, eked out solid positive results for such as test as theirs, it *still* wouldn't demonstrate that 'golden ears' are to be believed by default. Because the data *will* have to be 'eaked out'. Because from all the data we have, and are likely to ever have, hi rez, if it can be heard at all, is at best *hard to hear*.