Digital BW, The Print

Keith writes:

> I think that in truth the jury is still out
> on this when looking at the latest high pixel
> count sensors.

I've heard similar claims since the Kodak DC50 came out six years ago.
Every new camera is "better than film," as long as you don't look too
closely.

> from what I gather, B&W film, properly exposed
> and processed can have about 8 stops ...

According to the manufacturer data sheets:

Fuji Neopan Acros 100 . . . . . . . .  8-10 stops
Agfapan APX . . . . . . . . . . . . .    12 stops
Ilford FP4  . . . . . . . . . . . . .     8 stops
Ilford Pan F  . . . . . . . . . . . .     9 stops
Kodak T-Max 100 . . . . . . . . . . .  8-10 stops
Kodak Tri-X Pan . . . . . . . . . . .    10 stops
Kodak Technical Pan (pictorial) . . .     8 stops*

* The curves imply that Tech Pan could possible go to 16 stops or beyond,
under the right conditions.

For comparison, here are some color film ranges:

Fujichrome Velvia . . . . . . . . . .     7 stops
Fujichrome Provia 100F  . . . . . . .     9 stops
Kodak Portra 800  . . . . . . . . . .    12 stops
Kodak Kodachrome  . . . . . . . . . .     5 stops

These are the ranges between the start and end (toe and shoulder) of the
characteristic density curves, i.e., these are the ranges in which the film
density will vary enough to hold detail.  As you can see, the ranges are far
greater than one is ordinarily led to believe.  In particular, it looks like
the commonly-held ranges for slides date from the Kodachrome days, since
todays E-6 slide films manifestly hold a lot more.

I can verify that slides hold a great deal of detail, a fair amount of which
is not visible on the light table.  It can be recovered with a deep scan and
some Photoshop work, though.

The range of Tech Pan seems almost open-ended, depending on how you expose
and develop it.

The actual density ranges on the film itself (between clear film and maximum
silver or dye) are typically over 100:1, and sometimes as high as 3000:1
(Kodachrome and some slide films).  This represent a log density of as much
as 3.5, which can give scanners quite a bit of hard work.

> ... wheras colour and digital has about 5 at best.

Even color film does well.

As for digital, I don't have any data on that.  I do know that, in theory, a
CCD is limited mainly by thermal noise and the capacity of the photosites;
it is possible to achieve 16-17 stops, under ideal conditions.  Ordinary
digicams, though (even pro digicams) never come close to this ideal,
however, and I don't know how far they get.

> In terms of colour prints, I think the latest*
> inkjet systems gives do give prints that are as
> good as any - and a whole lot easier.

I find it cheaper, easier, and faster to get Fuji Frontier prints instead of
trying to print my own.  And the quality of the Frontier prints blows away
my own ink-jet prints.  No contest.

The situation isn't so cut and dried for black and white, however.

> It MAY be true that a really good colour darkroom
> can do better, but it's a lot or work.

Enlargers are history.  Even pro labs just do high-resolution scans of the
film, then use laser printers to expose the photographic paper.  It gives
much sharper and better results than an enlarger would.

> It may well be different, but it would be difficult
> to measure objectivly.

I agree.  The difference, when it is there at all, is usually in the
"quality" of each pixel, that is, the degree to which each pixel accurate
represents a sample of the original scene.  MF gives more pixels to start
with, so the quality of the resulting pixel is higher, for a given final
resolution.  It isn't always obvious, but it seems to make a difference on
scenes with a lot of random detail.