That’s so true. I know some people initially worry about not having a hot enough audio file that they’re not fully exploiting the fidelity of the available bit depth they’ve got, but most of their audio lacks the dynamic range for it to really make much difference if you bother analyzing it, anyway. Bigger issue they come back with is later being surprised they’re so much quieter than everyone else, which, god forbid, gets into that best practice stuff meeting demands where they want to be loud enough. Once streaming or hosting sites get through with their files, though, the audio has been subjected to worse than what basic normalization is ever going to do to their files… assuming the user didn’t normalize right up to the 0dBFS brick wall.
Normalizing the volume should not reduce the dynamic range unless you’ve chosen to do compression during the processing to a target based on something like RMS rather than just automatically adjusting by the difference between the current maximum peak of the audio file vs peak target. Most software normalizing defaults to peak. You could actually take an audio file with maximum peak at -15dBFS, double its volume by +6dB twice, save it in the same format it was already in, then lower its volume by 12dB, and it’d still be exactly the same original audio data.
Perhaps if all you’re judging it by is dynamic range, but I’d otherwise take umbrage with that considering Engine OS treble roll-off and having 100X the intermodulation distortion of the original 16/44.1 audio you pass through it. I suppose that’s largely separate of the topic of how Engine OS records or normalization works, though.
I don’t own any of the Prime all-in-ones, but doesn’t it just record in situ the exact digital audio at the same digital full scale levels as what you’d see on the computer over the USB from it? If that’s the case, and assuming the master volume control is at unity so you have metering accurately reflecting what the channels and record out are doing, then the top meter LEDs should represent -1dBFS below the brick wall.
So, if that’s all the case, then really lazy or daft users could just run their levels very hot while staying out of the top meter LEDs. That’s what Pete Tong usually does when streaming from home while keeping his inf:1 peak limiter turned on.
Personally, I usually have the streaming software add dither, then boost by 6dB, then have a 1st stage soft limiter, then the hard peak limiter, and instead of running the mixer hotter than I would usually, I just continue to stay out of the second-to-top meter LEDs like I always do for the hottest, least dynamic tracks. That gives me a couple dB of headroom for slip ups even prior to the peak limiter… and I don’t look like a noob with the way I’m running the mixer levels. 
