[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Optimizing Negatives for Scanning
Making the Proper Negative for the Scanner
Disclaimers: My work and efforts are to produce the highest quality
Pt/Pd prints. My interest in digital negatives is to be able to get the
original information from the original negative, manipulate it as
desired, and produce a Pt/Pd print. Any other outcome is not
considered, even though it may be of value to others. The scanner used
for the following was a HP ScanJet 6300C. The noise levels from this
scanner will likely differ from other models which can influence
results.
Just as it is so very important to customize the negative for the
selected printing process, so it is important to produce the proper
negative for scanning. Those who have used a scanner to record the
information from a negative should be familiar with the relationship
that bit depth of the scanner must be increased as the dynamic range of
the negative increases. If the bit depth does not increase, vital
information from the negative will be omitted from the scan. This
information likely first disappears under a level of noise. Next the
information becomes limited because a certain number of bits can only
provide a certain number of tones and the image requires a certain
number of tones be present so that posterization effects are kept
hidden. Posterization is not eliminated in a digital negative, only
minimized to a fine enough level. (The actual desirable number of tones
is not known at this point.)
To determine proper dynamic range and processing requirements of the
negative, several negatives were made at the following developments.
The film used was 4x5 Tri-X exposed at 200ASA. The first number is
development time in minutes with continuous gentile agitation in a tray
(one negative at a time). The second number is the concentration of the
HC-110 developer in ml/litter. A Zone VI temperature compensating timer
was used.
3@32
4@32
5@32
6@32
4@54
5@64
6@64
7@64
A typical full range Pt/Pd print would have the negative processed at
5@64. The dynamic range of these negatives increases as one moves down
the list. This is the usable, practical dynamic range. Any range
adjustment due to base plus fog is considered negligible for this
study. Any density beyond that related to Zone XII is dismissed as not
practical to use. Zone XII is considered because a palladium print can
easily produce all those zones.
Each of the negatives was scanned with 12-bit pixel depth and converted
and stored as 8-bit (the way of my scanner). The closest negative with
the highest dynamic range and without noise in the highlights was found
to be 5@32 (this is half the processing given to a typical Pt/Pd
negative). The noise would appear at the same density, so as each
negative that had more dynamic range, more of the upper end would
consist of noisy data.
Data from two of the negatives can be found at the following link.
http://home.att.net/~jeffrey.d.mathias/info/scan_noise_negative.html
It is true that the noise can be smoothed out with various software
techniques, but regardless the original data covered by the noise is
corrupted and lost.
A similar evaluation can be made by scanning a Stouffer 21-step. Find
the densest step that does not show noise and that density will relate
to the maximum density of a film processed a certain way. Note that the
actual film used is a better test as the characteristics of the Stouffer
21-step may be different, however densities can be evaluated fairly
easily with the Two Hole Method. An example using the 21-step can be
found at the following link. Scans are at 12-bit, 1200 ppi with noise
reduction turned off and no sharpening and exposure control set to full
range.
http://home.att.net/~jeffrey.d.mathias/info/scan_noise_example.html
CONCLUSION:
In order to get optimum results it is recommended to match the
negative's dynamic range with the scanning equipment used. A dynamic
range too high will result in the loss of information, most likely from
the highlights. A dynamic range too low will result in wasted data
capacity. It would be preferable to error too low so as to maintain
image integrity.
NOTES:
One must test their own negatives as many parameters are capable of
influencing the results.
One must keep in mind that the examples here only deal with the noise
level and that actual prints must also be made and evaluated from
digital negatives produced by a consistent procedure.
Just to give something additional to ponder:
It is guesstimated that a typical negative made for Pt/Pd printing would
require 14-bit data without noise. If it is assumed that 2 bits will
contain noise (typical), then a 16-bit scanner should be sought if one
wishes to scan those negatives.
More to ponder:
It is not known if or what differences a Pt/Pd print may exhibit between
using a 16-bit scan from appropriate negative and a 12-bit scan from
appropriate negative. It is suspected that a negative with a higher
dynamic range may enable a higher print quality. It is not known, but
could be possible that this is part of the reason for a Pt/Pd print to
have better tonal separation than some other processes. This comparison
could also be made controlling built analog negatives, if one has the
time. As always, any contributions are welcomed.
--
Jeffrey D. Mathias
http://home.att.net/~jeffrey.d.mathias/