D
Dale Kelly
the last project I had while working at Kodak was to create a
characterization target for color negative films, these was geared toward
the portrait and graphics markets
granted with more digital cameras selling then optical cameras, and with
digital projection in cinema, the use of color negative films has gone
down considerably over the the last ten years
but, one advantage of negative films, color or B&W, over transparency
films or digital cameras is exposure latitude, there are certain
situations where you just don't know what the right exposure will be, or
the exposure is going to vary, where you might choose to use a negative
film
I see no one has taken this issue to a manufacturing stage
my approach was to design a target that covered the gamut of the film,
using a film recorder directly on the negative, or using a film recorder
on a transparancy film, that could then be optically printed on the
negative for mass manufacture of the negatives, I believe I used a
10x10x10 RGB grid that was randomized before recorded
the intent of the negative was assumed to be a print on optical paper, so
in order to get colorimetry to mathematically map the RGB scan of the
negative, you would have to print the target on each paper you wanted a
model for
the target had a full range near neutral scale, getting a near neutral
scale required some fundamental characteristics and assumptions of the
behavior of the film with respect to a paper and printer
it was printed such that any expsoure on the negative would register on
the paper, in other words, you would print the target such that the full
dynamic range of the negative would register on the paper
secondarily, the neutral scale on each paper would be balanced to neutral
then the RGB values of the scan could be mapped to the colorimetry of the
scan and encoded in PCS
this approach would allow normal, under and over exposures to be
characterized on the same target, considering printed, it does mean that
ann over or under will not result in full 24bit encoding, but that is the
same as ANY scanner that does not have analog prescan capabilities, which
FEW have
characterization target for color negative films, these was geared toward
the portrait and graphics markets
granted with more digital cameras selling then optical cameras, and with
digital projection in cinema, the use of color negative films has gone
down considerably over the the last ten years
but, one advantage of negative films, color or B&W, over transparency
films or digital cameras is exposure latitude, there are certain
situations where you just don't know what the right exposure will be, or
the exposure is going to vary, where you might choose to use a negative
film
I see no one has taken this issue to a manufacturing stage
my approach was to design a target that covered the gamut of the film,
using a film recorder directly on the negative, or using a film recorder
on a transparancy film, that could then be optically printed on the
negative for mass manufacture of the negatives, I believe I used a
10x10x10 RGB grid that was randomized before recorded
the intent of the negative was assumed to be a print on optical paper, so
in order to get colorimetry to mathematically map the RGB scan of the
negative, you would have to print the target on each paper you wanted a
model for
the target had a full range near neutral scale, getting a near neutral
scale required some fundamental characteristics and assumptions of the
behavior of the film with respect to a paper and printer
it was printed such that any expsoure on the negative would register on
the paper, in other words, you would print the target such that the full
dynamic range of the negative would register on the paper
secondarily, the neutral scale on each paper would be balanced to neutral
then the RGB values of the scan could be mapped to the colorimetry of the
scan and encoded in PCS
this approach would allow normal, under and over exposures to be
characterized on the same target, considering printed, it does mean that
ann over or under will not result in full 24bit encoding, but that is the
same as ANY scanner that does not have analog prescan capabilities, which
FEW have