How would it work? It's like a digital zoom on a cheap digital camera. The CD word length is 16 bits. If you put that into a 32 bit word, you have to guess what goes in the less significant bits.
Try the Muse Model 10. Not since I first got into Pass Labs Aleph electronics (and later, X1preamp) have I been so impressed with what a brilliant engineer can achieve. How he does it I don't know, but I am guessing that software propgramming has something to do with it. This is the first, and so far the only, Redbook CD player I have heard (and own) that makes CDs sound like a genius format. I don't think extra bits is the answer. The proof is in trhe listening, in any case. I spent a lot of money on this unit, knowing I would never get it back--Muse does not have the upscale image among audiophiles that some brands do. I bought it anyway, becuase of what it did for music in my room. I have not looked back.
Which computations ? A 20 bit DAC gives 120dB dynamic range .. greater than ANY analogue, and greater than any amplifier is capable of reproducing. Bob Stuart of Meridian gave a paper saying that any more than 20 bits is a waste, and DVD-audio (which meridian authored) is only 24 bits because most commercial DSPs use 24 bit word length, and it makes mixing easy (no overflows).
Sampling rate is probably more important one you get above 16 bits.
But ultimately implementation and design matters more than miscellaneous specs.
I'm always skeptical of the "latest" technology. I remember when CD players were new - my first player was a McIntosh MCD7000 and it had 14 bit DACs because, although 16 bit convertors were available, the technology was so new that none were linear enough for McIntosh to use.
24 bit convertors are now very mature - but 64 bits? Me thinks we're playing a specsmanship game here. I would submit that Sean is right - at this point in the technology curve implementation is far more important than any particular technology.