There’s a misconception here about the nature of digital signals:
In (linear PCM-) digital audio “Over” doesn’t exist the same way like in the analog domain, as there’s no digital value for it.
The “Over” in ADI-2 indicates that full scale (the highest possible level) is reached for at least one sample.
From the manual, 15.1 Global Level Meter:
For digital I/Os, Over appears when the signal is at the maximum possible level. As usual in digital audio, Over is already displayed even if there is no real overload.
Therefore reducing the digital level cannot “heal” the very likely already clipped and distorted signals.
The only (negative) result is, you no longer have an indication any more why the music sounds like it does if you feel it sounds distorted.
Leave your player at 0 dB gain to get bit perfect playback.
You will see “Over” with contemporary popular music most of the time.
Why?
We are in the middle of “Loudness War II”.
Producers and record companies again think, that they have to sacrifice sound quality by squeezing out the last bit of loudness with compressing, limiting and clipping their productions, or the audience will think ‘it doesn’t sound good” because it’s a tiny bit quieter than other tracks.
Perverted, but true.
And that’s while we have to face the fact that more and more distribution channels like Spotify, Apple Music, TV broadcasts and many others are working with “loudness normalization”, which brings down the level of all productions to the exact same perceived loudness, no matter how loud they are “produced”.
Super-perverted, but again true.