negative film scanning data

[Dale Kelly]

I am writing a generic scanner driver for linux, I want to develop ICC
profiles for negative films also

BUT, the ICC definition of a negative film is the intent of the film
when it is printed ideally on negative paper. To accomplish this I
need access to technical specifications from film and equipment
manufacturers that are not properly specified in the tech publications

for instance, a color negative film would require

1) the spectral sensitivity of the scanner
2) after scanning you have some resemblance of the dye density of the film
3) you must convert that dye density to print colorimetry
4) to do this you must know the aim optical printer and aim optical
printer setup (filtration) used in aims of the film
5) you must know the spectral sensitivity of the print media
6) you must know the channel independent contrast of the print media from
single layer coatings of the film, typically used in making aims for the
film
7) you must know the process interimage matrix that accounts for sye
wandering either non-intentional or intentional due to DIR or DIAR dye
couplers, and also such effect from dye scavengers, this information is
also typically used in setting aims for the film
8) you must know the spectrophotometry of the print media's dyes

it is not adequate to show or a chart or show sparse (10nm etc. data), you
need as rigorous of data as used in the aim setting of the devices and the
media, in fact you need the EXACT data

why is it that such information necessary to generate negative film ICC
profiles is not readily provided by equipment and film manufacturers? The
same type of data is required for B&W negative scanning

another definition of negative film could be scene colorimetry, instead of
processing the dye density of the negative film to its print aims, you
could unbuild the dye density back to scene colorimtry by reversing the
interimage matrix, reversing the channel independent contrast, and take a
shot at undoing unwanted absorptions of the spectral sensitivity of the
capture negative film

all requires aim data for both the scanner and the film

 

[Dale Kelly]

all requires aim data for both the scanner and the film

this type of information is also necessary to model complex film and paper
recording systems, such as used in the motion picture industry that do not
lend themselves easy to a target/fingerprinting method of characterization

 

[Dale Kelly]

all requires aim data for both the scanner and the film

and if you wanted to impart the look of one film onto the other, or one
film/print media system onto another film print media system, you would
need the same data

 

[Gerhard Fuernkranz]

BUT, the ICC definition of a negative film is the intent of the film
when it is printed ideally on negative paper. To accomplish this I
need access to technical specifications from film and equipment
manufacturers that are not properly specified in the tech publications

for instance, a color negative film would require

1) the spectral sensitivity of the scanner
2) after scanning you have some resemblance of the dye density of the film
3) you must convert that dye density to print colorimetry
4) to do this you must know the aim optical printer and aim optical
printer setup (filtration) used in aims of the film
5) you must know the spectral sensitivity of the print media
6) you must know the channel independent contrast of the print media from
single layer coatings of the film, typically used in making aims for the
film
7) you must know the process interimage matrix that accounts for sye
wandering either non-intentional or intentional due to DIR or DIAR dye
couplers, and also such effect from dye scavengers, this information is
also typically used in setting aims for the film
8) you must know the spectrophotometry of the print media's dyes

IMO the general problem is, that printing of a negative is not a
statically determinate color reproduction process. For instance, my
understanding is, that each photo is individually color balanced and
individually exposed when it gets printed. Therefore, depending on the
different scenes captured on different frames of the negative film, the
_same_ color on different negative frames will eventually map to
_different_ colors on the corresponding prints of these frames. So there
is no static mapping between colors on the negative and the
corresponding colors on the resulting print. But only color reproduction
processes with a static, invariant mapping between source and
destination colors can be characterized/described with an ICC profile.
An ICC profile cannot describe a color mapping which varies depending on
the contents of the image being transformed.

And if we would assume that making prints from a negative were a
statically determinate process, then IMO you would not need all the
above information in order to establish a mapping between scanner RGB
(for the scanned negative) and the colorimetry of the resulting print:

* You would need a (negative) target, with patches covering the
whole gamut of the negative film
* Make a print from this negative target and measure the colorimetry
of the patches on the print
* Scan the negative to obtain the scanner RGB values for each patch
on the negative
* Establish a mapping from scanner RGB to print colorimetry (-> i.e.
the profile)

(however, due to the reasons mentioned above, this is a rather academic
consideration)

Regards,
Gerhard

 

[Dale Kelly]

that each photo is individually color balanced and
individually exposed when it gets printed

well, the solution to this could be to do away with a static encoding
space, and have a viewer and basic editing algorithms associated with the
encoding space