and for everyone else, you can research "intersample peak" and intersample peak overs. It is absolutely a known phenomenon among mastering engineers and discussed regularly, and demonstrated in various testing.
The consensus on audibility is that it is hardware (DAC) dependant. Can be very objectionable on very accurate hardware but generally not audible in lossy or low end repro chain. To avoid requires headroom in the encoding not always allowed due to the loudness wars.
Probably why I prefer DSD. I am unfortunately very annoyed by any high frequency or peak glitches in digital. Makes me want to immediately throw the offending component riight out the window...