I didn’t actually listen to it for those 20 hours; I just left it running, with the bookshelf speakers tucked under a blanket.
The difference is so obvious that I believe it could easily be measured.
When I first took the amplifier out of its packaging and ran it for 2–3 hours last night, the high frequencies were clearly distorted, and the sound felt more constrained.
I have quite a bit of experience with "breaking in" amplifiers, though in only a small percentage of cases has the process actually changed the sound.
For example, the GZLOZONE 3255 and the "blue board" 3255 (the one with two large capacitors)—I bought both brand new from AliExpress, and out of the box, they sounded rather lackluster. In those cases, 20 hours was plenty of time for that to go away and for the sound to become normal.
Perhaps 5 hours would have been enough; I’m not sure.
On the other hand, amplifiers like the Ampapa D1 showed no change whatsoever, even after 250 hours of operation.
I did that purely out of curiosity.
The SMSL MHA1 also didn’t change its sound at all—though, to be fair, it wasn’t brand new when I got it; it might have already been "broken in" when FanatNsk performed measurements and maximum power tests on it.
As for the SMSL AO10 (based on the 3116 chip), I bought it new, and it didn’t sound great.
After 150 hours, nothing had changed, but by 200 hours, it began to sound noticeably more pleasant—though, in this instance, it could just be a trick of the mind.
Then there’s the D5 Pro, a clone of the NHB-109; when it’s brand new, the high frequencies are truly awful. After 150 hours of break-in, the highs improved, and by 200 hours, they got even better—though they likely won’t improve any further. Regardless, the high-frequency range remains the weak point of this amplifier.