Failing Application Test

I think I have figured out why hardly anyone can pass the test. The instructions call for minimally processed, basically "raw" audio. There are no technical requirements given. Only later after being rejected, on this forum, I learned the standard is -90 db. In my opinion, this is not possible without using any software effects, so assuming your performance is good and there are no other critical flaws, if you take the instructions seriously, you're still going to fail 100% unless you're in a studio, maybe, or using very high-end equipment and soundproofing. Even then, I'm not sure it can be done.

There's also some confusion in terms, for me as an audiobook narrator. What you call hiss or white noise, we call room tone, up to -60 db, and it's not a bad thing. It's not noticeable until it's gone and then the difference can be distracting. If it's consistent, it's fine, maybe necessary, for long-form narration. Total silence between words is hard to pull off for hours of audio, so we don't try, because any inconsistency can make a bad product.

Breath sounds are a similar issue. It's time-consuming to take out every single breath sound in hours of audio and some argue it sounds unnatural. I've heard some listeners are bothered by it and won't listen, though they might not know exactly what's wrong.

Feedback said my breaths were distracting. I've been told my breathing technique is very good - quiet and controlled. I'd just take them all out at this point. It's easy enough in such a short piece.

Knowing now what I have learned, passing should be no problem. People are using noise gating and other tools in order to pass, probably some noise reduction though it says not to. Conservative use of some software effects should get me to -90 db without affecting the sound of my voice.

Give us just a little more information and I bet your success rate would go way up, but then you couldn't say anymore that only 2% pass, so maybe that's not what you want.