Oh how I wish Class D amps ...

Simon, - my bad! I tend to attribute everything to DSP forgetting that even simple analog filtering is form of signal processing.

As for class D - I believe that it is very good for the money (important to me). Opinions vary a lot. Some believe that class D is at reference level while others that class D is LO-FI.

Stereophile reviewed Bel Canto REF1000M (Icepower) ending review with this conclusion:

"The Bel Canto Ref1000 Mk.II can be compared with the cream of the other amps I've had in my system: the Classé CA-3200, Mark Levinson No.433, and Ayre Acoustics V-6xe. Each of these distinguishes itself in different ways, and particularly with different speakers. Because of this, I think I must keep the Bel Canto Ref1000 Mk.IIs as a reference amplifier—an easy decision even when based solely on its sound, but also: in my living room, none of the others can be so easily hidden in plain sight."

Class D amps might be very revealing showing shortcomings of the system but also might have better synergy with some speakers than others. Also, long break-in is most likely required. It took about 400 hours for my amp to sound best.

It is often said that Icepower sounds good only with linear power supplies. On the other hand newest linear class AB amps from Jeff Rowland (one costing $46k) use switching power supplies.

Well ... I guess I am wrong for assuming, but I do assume that EVERYTHING is always subject to system synergy. Having said that, the Cirrus speakers are the single best tool I have ever seen for component and wire evaluation. They are an open window, and PERFECTLY show every slightest change made upstream. So I am fortunate to have a system that does allow for very good evaluation.

I have no adgenda to bash digital amps, just a desire to see them improve.

Please see first schematic::http://users.ece.gatech.edu/mleach/ece4435/f01/ClassD2.pdf
L1 / C1 are how the switching frequency is removed.....and should induce a phase shift in the audio band. No?

Simon,
Want a good example of an Analogue computer?
Please see link:

http://en.wikipedia.org/wiki/Analytical_engine

Charles Babbage could never get it together long enough to build it, but within the last decade or so, Smithsonian built one to the original drawings, and it worked fine.
The machine would have been used to calculate log / trig tables for navigators and such.

Well ... I guess I am wrong for assuming, but I do assume that EVERYTHING is always subject to system synergy. Having said that, the Cirrus speakers are the single best tool I have ever seen for component and wire evaluation. They are an open window, and PERFECTLY show every slightest change made upstream. So I am fortunate to have a system that does allow for very good evaluation.

I have no adgenda to bash digital amps, just a desire to see them improve.

I think we all might have a bit of "that" same centricity that our speakers (or other gear) is neutral and that it shows the distractions of all other gear, but I tend to think that is not always so.

I think most might offer that "switching" amps have exceptional highs (in the better ones) and if your speakers are not letting you hear that, then what does that tell us?

In the end, it is finding the components that sum to our preferences.