Planar Speaker Asylum: digital technologies, power circuits & jitter by Arbelos

While i respect your technical skills, and agree that the TriPath board is a very interesting option, i think your commentary is quite undifferentiated.
Digital amps maybe share *some* sonic characteristics, but they have different technologies. Spectron uses a proprietary technique, only the BelCanto is based on the TriPath module.
Some digital amps (Sharp) had widely differing reviews: Bad (Colloms) - Rave (Mike Fremer). There are probably widely differing technical characteristics (speak: weaknesses).
The Spectron Musician II sounds outstanding in my usually tubed ears. I feel that the Spectron technology might be sonically superior, and having read comments by Ric Shultz about a modded Spectron vs. a slightly modded BelCanto give hints into that direction. Problem with the Spectron seems to be reliability, but i think it's mainly fuses that are sometimes 'eaten' at switch-on.
The Spectron is in a different class power-wise BTW. It is a very subtle and extremely expressive amp, bordering on SE territory. A real gentle giant.
I heard the multi-k$$$ Sharp digital integrated amp on the Coincident Technologies Total Eclipse and, conversely to my (Colloms-induced) luke-warm expectations, it sounded outstanding, better than the Berning ZH 270 on these speakers, and way better than a very renowned italian 845 SE amp, the New Audio Frontiers. Just almost audio dreamland, for me one of the less than a handful of *really* satisfying audio systems i ever heard.

My feeling is that one of the main problem of digital amps is keeping the clock jitter-free within an extremely 'magnetized' ie. high-current environment. The earlier TacT amp had pretty mediocre jitter measurements and i suspect this is true for at least the Sharp amp, if not all others too in differing degrees.
Now this seems to be less of a problem when an analog source is used, as in the Sharp, Spectron and Bel Canto. Probably because, if a common master clock is used, the majority of jitter-induced distortions are cancelled out by the reciprocal converting processes.
The wished-for jump in quality for digital input signals might be a premature assumption therefore.