Yahoo Groups archive

Digital BW, The Print

Index last updated: 2026-04-28 22:56 UTC

Message

Re: [Digital BW] When is DMax not Dmax?

2012-11-18 by Tony Sleep

On 18/11/2012 00:59, Paul wrote:
> But scanner specs don't show that, they show Dmax in the negative. Like
> you say, Dmax is Dmax. And so the scanner specs give me information on how
> my highlights might look like in the print - and of course I'd like some
> detail there too, I don't want them blown out. But it looks like typical
> scanner specs can't help me with detail in the print's shadow areas.

  The DMax does, as you say, correspond to the darkest area within which 
the scanner is able to capture tonal detail. This is highlights in 
transmissive mode with negs, shadows in tranparencies, and shadows in 
prints in reflective mode.

This limiting factor is very much complicated by sensor noise. Ability of 
sensors to record detail at minimum signal is invariably compromised by 
this - there will always be a shading of detail into noise which will 
obliterate detail.

Stated DMax that is unaccompanied by a qualifying statement of the noise 
level present (at the measured DMax) is meaningless, a bullsh*t marketing 
claim. What matters here is signal to noise ratio. Manufacturers almost 
never state this anywhere.

The same applies to ODR (optical density range), since it takes DMax as 
its lower bound. Without qualification regarding noise level, it means 
nothing.

There is no valid or useful way to compare scanner abilities from 
published DMax or ODR, you can only tell by using them.

Optical system abilities are also never specified in any useful way, but 
lens systems are as crucial as in enlarging. You just don't get excellent 
lenses in cheap scanners, because a fine macro costs too much to manufacture.

There is also OE software to contend with - eg scanners with 10,12 or 14 
bit ADC's with OE software that works at 8 bit throughout, then merely 
upsamples to produce 16bit output if selected.

For years HP, Epson and Polaroid OE sofware was hobbled like this. This 
has severe bad effects at the dark end where bit values are small and 
rounding errors are very significant. End to end 16 bit working was one of 
the great advantages of Vuescan, and the improvement in noise was 
dramatic, on some models gaining almost a full stop more usable shadow detail.
-- 
Regards

Tony Sleep
http://tonysleep.co.uk

Attachments

Move to quarantaine

This moves the raw source file on disk only. The archive index is not changed automatically, so you still need to run a manual refresh afterward.