Charlie is right on that analog can be very good - as a matter of fact, there are mastering tape machines that run at 30 inches per second and use 1" wide tape for mastering two track stereo. Such machines give even 24 bit digital a good run for the money in some aspects. In other aspects, digital has it all over analog. One of these is modulation noise in any tape recording system, another is the somewhat lacking bass response of tape, especially as the tape speed gets as high as 30 i.p.s. Both have trade-offs that must be weighed when working at such levels of precision.

12 bits is actually pretty darn good - that CD that I circulated (and got me put away in a Mexican jail for copyright infringement ) proves this fact.

I have actually been questioning however if analog noise floors and digital ones can be compared, even somewhat directly. It can be proven that with the addition of dither to a digital word, it is possible to hear below the theoritical noise floor limit for a particular word length. Gonk has a CD that I made that proves this fact (it has test signals I made myself, so there is no danger of me ending up in the slammer again ) But there has to be a limit to how much can be done when adding dither to the least significant bit of a digital word. There has to be a downward limit, below which nothing can be perceived.

Analog however does not have any theoritical lower limit of resolution. The signal just continues to be submerged further and further into the noise floor.

I guess a test could be devised to determine just how much difference there actually is, but personally, I'd rather go out for pizza.

Anyway, I'm open for discussion on this subject.