You may want to hold off on getting excited about upsampling. I just finished an audition of the Perp P1-A in my system, using alternatively an EVS Millenium DAC, the Muse 296, and Lexicon DC-1 processor. In all cases, I found the sound was BETTER when driving the DAC directly from a transport, than when inserting the P1-A in-between and doing resolution enhancement to 24/96 (for the EVS and Muse), and 20/48 (for the Lexicon). The upsampling seems to rob the music of its dynamics, and the bass gets DRAMATICALLY softer and flabbier. I found this shortcoming to be most pronounced, perhaps surprisingly, by the best DAC; the Muse 296. Using the P1-A, I strongly preferred the Sony SCD-1 alone to the P1-A/Muse combo. When taking out the P1-A and driving the Muse directly with a transport, OTOH, I preferred the Muse over the Sony alone perhaps in a 60/40 ratio. In short, my one person's impression is that upsampling is definitely not a panacea, it is not lightning in a bottle, but in some systems to some tastes, it may offer a slightly different presentation that is marginally preferable.
24/96 Upsampling
I've got a SFD2 Mark II, a very good class A type DAC. SF tells me that they will have a 24/96 upgrade for this DAC out this summer. They also have a D2D-1 stand alone box for $600. Anyone experienced the difference in a good DACs performance using one of these new upsampling digital to digital boxes?