Update 9/2012 (warning! Non-scientific, subjective and unsubstatiated assumptions! i.e. I listened to the darn thing): I have been fiddling around with RCA voltage line attenuators for a while to get my headphone amp to play nice with my xda-1. It distorted before.
Well I decided to try it with my upa-2. I noticed a surprising difference. It was subtle. Before without the attenuators, as the volume increased, things felt a little harsh or compressed or strained would be the best way I can think of putting it. However everything felt very clear but at times not as realistic as I would have hoped.
So I used these voltage attenuators and I actually recognized a difference. Basically the sound signature changed. All this has to be kept in mind that the difference is very subtle. The treble was slightly laid back, the mid-range became more prominent and the bass became just slightly more undefined. Warmer would be the best way to put it.
Pros: Depth was easier to define and room ambience came through better. Voices sounded more like they were in the room than before.
Con: There was a slight loss in detail. Very powerful microdynamics were slightly dulled. Movies didn't sound as good. It felt like the upper bass was slightly too much.
Anyway, from what I gather voltage attenuators aren't supposed to change the sound so I was surprised that it did.
But then I came across this:
www.enjoythemusic.com/magazine/equipment/0803/rothwell.htmWhile I don't agree with some of his descriptions especially the degree as I feel it's a very subtle change, I think it echoes a lot of my sentiments.
He brings up the issue with overloading power amps with pre-amps that push out a signal that's too powerful due to amps being driven to full power at low voltages. I have no data to back it up but I do wonder if this is what is happening. If so, that doesn't really make me very happy as I would think that emotiva's products need to work BEST with emotiva's products and if there is some input overload on the upa-2 then it is rather dissapointing.
I do want to mention (again) that I don't have any facts about this or the specs of the inputs and how things mess up if there's too much power or any proof of input overload or poorly matched emotiva parts. However, that's where I am leaning toward simply because of my experience.
My primary purpose was to allow the xda-1 to play at louder volumes to bypass the lossy attenuator's effect a bit better and so I was surprised when I noted the sound signature change and a more pleasant feel to the sound. Another user has the RCA attenuators ordered who has similar electronics to mine. I shall wait and see what his experience is.
Anyway since then I have left the RCA attenuators in there as I find the sound more pleasing to my ears. I may do some more non-blind A/B testing when I get the fancy to.
Anyway short summary with RCA attenuators: overall music reproduction feels improved and more pleasing. Better depth. Dynamics feel more natural with a slight sacrifice on treble "snap" and microdetail. Movies are not as pleasing. I feel (without proper proof) that it may be a possibility that the xda-1 is outputting a voltage that may be overloading the UPA-2 (as well as my headphone amplifier).
Edit: Another possibility I have entertained is that this may simply be the limitation of class A/B amps as they get past pure class A mode limits. My speakes are reasonably efficient and the slightly harsh compressed sound comes on mainly when there are dynamics which aren't necessarily tremendously loud but loud in comparison.