Mort54, you seem to be implying that all ADCs are adding 2 bits of noise as a rule of thumb but this is not true.
I certainly didn't mean to imply that, and thought I was quite clear that noise figures vary considerably across different A/Ds. I specifically pointed out that not all 14-bit or 16-bit A/Ds were created equal. In fact, that was essentially my main point.
I was simply trying to point out, without getting too verbose and technical, that there were different grades of A/Ds, and that it was certainly possible to get good quality A/Ds with relatively low noise figures, and poorer quality A/Ds with higher noise figures, all with the same bit depth. I wasn't making any absolute statements about the number of bits of noise present, since my whole point was that noise was a figure that varies greatly with different A/Ds. I was trying to make the point that just quoting bit depth was, by itself, somewhat meaningless, without also discussing the quality of the A/Ds involved.
The 2 to 4 bits that I mentioned several times was offered as an example of a relatively good, low noise A/D that still had the relatively high sample rate capabilities needed for imaging applications. You can certainly get lower noise A/Ds, but they also tend to be available only in lower sampling rates. My example was also based on a quick survey of a few suppliers I usually use in my designs (my day job). It wasn't an exhaustive survey. I ended up using a Linear Tech model as my example of a good, low noise, high sample rate A/D. I didn't look at TI models, and it's certainly possible they have something even better than the parts I looked at.
I'm sure I could have written my original post more clearly, but by the end of my second paragraph, I was already boring myself to tears :-)
In fact it is even conceivable that a poor quality 16 bit SDC will have a worse SNR than a 14 bit ADC.
Absoulutely true.
So at least we can agree that not all ADCs are created equal. It is quite possible for a stunning 14bit circuit to outperform a poor 16bit competitor.
Again, absolutely true. One other point I tried to make in my earlier post, however, is that given the higher costs of MFDBs, it is likely that the MFDB designers probably have a higher parts budget to work with than lower cost DSLR designers. And that, therefore, it was also likely that MFDBs use better quality A/Ds than DSLRs do. Which again tried to make my overall point that simply comparing bit depths was meaningless, without also knowing what kind of A/Ds were used. There are obviously a lot of conjectures in this last point, however, but hopefully I made that clear in my original comments.
Regards,
Mort.