Digital BW, The Print

Dana H. Myers wrote:
> David Whistance wrote:
>   
>> I'm not sure where this comes from. 16 bits of data gives 2 to the power of 16 shades of grey, ie 65,536 of them. Should be plenty to divide by 17 stops, particularly as the relationship between a stop and a set of shades of grey is arbitrary - you can map them how you like with either the scanner software or Photoshop. What does surprise me is that Steve gets this from an Epson V700, however I've seen some of his images and he does undoubtedly capture a very large SBR with detail at both ends so I'm sure he's right.
>>     
>
> No one is questioning that David gets a tremendous SBR; it's the
> specific number that sounds too good to be true.
>
> Dynamic range of a sample is limited to log-base-2(#levels).
> This is a theoretical limit; it assumes a perfect sampling
> chain.  Real-world performance is not going to be as good
> in general, there's a noise floor in the electronic chain,
> flare in the optics and limitations in the A/D converter itself.
>
> So, in theory, the most that could be represented in a scan is
> 16 stops; practically speaking, it's probably more like 14 stops.
>
> Cheers,
> Dana
>   

I think y'all are over thinking this. What scanners do is convert the 
density range they see on the film to the digital range they have 
available. That's all that's going on here. The scanner doesn't have any 
idea what the SBR was at the original scene. None. All it knows, ALL it 
knows, is the density range of the film it's scanning.

If I can put 17 stops of SBR into a density range of 3.0 on the film, 
the scanner will convert it correctly to it's digital range of 0-65535 
(assuming it can read through that much density of course). If on 
another film I put that same 17 stop SBR into a density range of 1.8, 
that too will be scanned into a digital range of 0-65535.
--
Bruce Watson