[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Calibration of Digital Negatives
The question in your original post is something that I have
coincidentally been addressing myself for my digital platinum negatives.
In fact I did the same thing you did and read off the reflection
densities of Burkholder's Ole' images (from both books). The conclusion
I came to is that the reflection density that you would want for your
50% tone depends on how high a Dmax your coating can provide. I found
it interesting that the Dmax of the image in Burkholder's first book was
about 1.5, closer to what one could hope to attain with platinum. The
50% tone in that Ole' image had a reflection density of about .55,
significantly lower than the .74 you would expect from an 18% reflection
gray card. Yet for a Dmax of 1.5, that .55 density for the 50% tone
looks right. To my eyes, a reflection density of .74 is a bit dark for
50%, but nothing I print these days has a Dmax anywhere near 2.2. (In
my densitometer readings of Burkholder's second Ole' image, I get .66
for the 50% tone using a well calibrated densitometer; I imagine they
vary somewhat.) I think you are correct in inferring that you don't
want your 50% tone to necessarily be 50% of your Dmax. What if you had
a Dmax of 4? Here are some numbers that seem to work. For a Dmax of
1.5, I shoot for a 50% reflection density between .5 and .55. For a
Dmax of 2, you can afford to go higher, say .65 to .70 or so, because
the added separation and depth in the darker tones will make those mid
tones look somewhat lighter. In general I have found that correction
curves for digital negatives that linearize the reflection densities as
a function of percentage give me platinum prints that look too dark.
Garet Denise wrote:
> I assume from your reply that a medium gray from a photoshop file (50%,
> level 128) should print at exactly 50% dot area . Yes, I know this sounds
> silly, but I don't know that much about the printing industry. The reason
> that I ask is that the test print which I gave densities for "looks" right.
> If I took 30% out of the middle values I fear the print would look quite
> I intended this to go to the list on Friday, not just to David. (Have to
> watch that 'Reply To' feature).
> Garet Denise
> >Example using your data.
> >Ds = 2.28 (100% dot)
> >Dp = .06 (0% dot)
> >Dt = .74 (your 50% patch)
> >1-(10^-(.74-.06)) = .791
> >divided by
> >1-(10^-(2.28-.06)) = .994
> >times 100
> >= 79.6% dot area
> >In the graphic arts world, we would refer to this output as
> >having 29.6% dot gain.