11:09 AM 1/8/2007 -0500, you wrote:
When shooting photos documenting something is it better to use film,
where an unaltered negative can be shown vs a digital image that can
be argued that it was made to show what the shooter wanted by software?
Thanks,
Bob
An interesting forensic question.
I believe that both film and digital images can be modified after the
fact to represent things not actually there.
With film, the path would be to scan at a very high resolution,
modifiy the image appropriately and then remake the film image via a
film printer thus yielding a "new negative.'
The digital path is more obvious.
In either case, however, I think a determined and skilled forensic
examiner could discover the alteration.
Things to look for:
Discontinuities in the image density gradients - change the contrast
range on a digital image and compare the before and after levels
histogram to see a gross example.
For film, differences in color balance between consecutive frames
could mean alteration.
Minute examination of the grain in the film image could show patterns
related to the film printer. Film printers are scanning CRT based
and you might be able to see the remnants of scan lines in the
image. An image that is slightly out of focus everywhere would
suppress these lines but would photographically be extremely unlikely
to occur in a real image - one would think that somewhere in the
image it would be in focus.
Differences between the shape of the image at the margins and the
actual shape of the camera opening / sensor chip in the camera
purporting to be the taking camera. In the duplicate, the size of
the image may change as well as the contour of the image in the
negative or digital image. Size differences may indicate a different
camera / lens in terms of focal length, etc.
Differences in spacing between negatives ( film camera only ) between
newly taken negatives from the purported camera and the proffered
negatives. Frame spacing is very idiosyncratic and may be like a
fingerprint in identifying the original camera.
Differences between 'dead' pixel in the purported digital camera and
the offered digital image. Ditto with dust artifacts on the
sensor. Missing / additional spots could be an indicator of alteration.
Examination of EXIF data in sequential images might show time
sequence difficulties.
I'm sure there are other features that could indicate alteration but
these are some that come to mind.
I can't speak for other manufacturers, but Nikon has taken this into
consideration in their latest upgrade firmware for the D2x and for
the original release in the D2Xs cameras. They provide a method of
encryption and certification that the digital image has not been
modified. Will most folks have such a feature routinely turned
on? Probably not, I don't.
Cheers,
James
James Schenken
