It's a different nasal tonality than the DX1000 have, or the Grados have. In fact I would call it more "shouty" than "nasal" coloration. The best way I can find to describe it is as the singer giving the impression that is having difficulties to deliver the voice.
My theory is that it's produced for the 5.5-6KHz peak. This enhances the 3rd formant area which changes a bit the pattern of the vowels. This can be also perceived as an "airy" feeling in the voices. It's also noticeable on wind instruments and violins playing high notes, they have a more piercing and airier character than they have on other phones. On violins it's kind of a metallic rasping tone and on trumpets with mute the timbre gets a bit harsher.
I don't find it annoying unless the own recording is "hot" already. I must say this has slightly improved by using the Gilmore Reference instead of the B22, and maybe for some additional playing time on the phones. Mine are about 80 hours so far.
IMHO the O2 are more effortless, relaxed, musical and engaging than the HD800. Their stage, while lacking the depth of the HD800's, is more precise, it changes following recorded differences in size and separation. The HD800 are more "generic", maybe not as much as K701, which seem to have a wide stage on whatever you play, but still always wide and open, showing less spatial differences among recordings than the O2 or even the R10. The O2 may sound as narrow as a pair of Grados or almost as wide as the R10 depending on the discs you play. No other phones I know of can do that with the sheer resolution of the O2.