SS amps with little or NO feedback

Magfan - Global feedback is OK as long as it is applied properly. The key is to get improvements in THD, IMD and DF but not introduce TIM. It can be done by using shallow feedback just to reduce THD to about 0.1% and then limit bandwidth at the input to one that amp had without feedback. It is always compromise but amp has to be linear and have wide bandwidth to start with. So, I would accept small amount of feedback but would run away from amps with great specifications like THD=0.01%

What "theoretical" problems you have with your class D?

Al - Fully balanced design has very high CMRR but I wouldn't care about small amount of sweet sounding second harmonic. I would be more concerned with unpleasant odd harmonics created by negative feedback.

Atmasphere - that's what I said. I don't care if sweet sounding second harmonic is canceled (by true balanced amp) while it still might produce unpleasant odd harmonics. Main benefit of true balanced design for me is high common mode rejection - perhaps not worth increased cost.