I’m interested in making a camera profile -- in
particular, a daylight profile for a digital back -- and I am looking for
pointers.
The existing Outdoor Daylight profile from Phase One lacks
highlight definition, saturation in light colors and have a hue shift in the
blues toward cyan. I made a “quickie” profile which did a
much better job. By “quickie” I mean I stuck the target in
the sunlight and took a shot without worrying about such things as color
reflection from the tan table it was sitting on, specular highlights,
etc. Even so, when comparing the colorchecker numbers in the image using
my quickie profile vs. the theoretical/average numbers provided by Bruce
Lindbloom and others, I got a much better match than with the supplied
profile. When applying to various images, the result was also better,
with much better highlight detail, better saturation of light colors (like
rainbows) and no hue shift in blue values.
So, I think I’m on the right track. I’ve
made lots of printer and scanner profiles but never a camera profile. So
I’m looking for pointers.
I realize I need to light the target as evenly as possible,
eliminate glare, and eliminate color reflections from surrounding
objects. I also realize that if I should shoot the target at mid-day sun,
it will technically not be correct for any other time of day (different
light). But that’s no different than, say, using daylight film at sunrise
or using the existing/supplied daylight profile at sunrise. So I realize
it won’t be perfect. But I’d like to get a better result than
with the existing/supplied profile.
A rep for the back said that making my own profile
wouldn’t work because of IR and/or UV affects. I haven’t had
the opportunity to question him about why he thinks this will be a problem or
how they dealt with it in the profiles that they produced. But I
don’t see why that would be a problem. I mean, of course they are
present in daylight. But that’s what I’m trying to
profile. It seems to me that the camera sees what it sees and, regardless
of what parts of the spectrum cause the camera response, the profile should do
its job and characterize the camera’s response.
Should I try to diffuse the sunlight or place the chart in
direct sunlight? For my quickie experiment, I placed the chart in direct
sunlight. I realize that direct sun would introduce an undesirable
specular component which may fool the profiler into thinking the colors are less
saturated than they really are. A diffuser tent could reduce/eliminate
spectral highlights and reduce contrast but may also slightly alter the color
temp (mine subtracts about 200K) but if the image is white balanced, I’m
thinking that it shouldn’t be a problem. Using a tent would also
likely help with evenness of illumination. So I’m thinking
that a diffuser tent is the way to go. Comments?
GretagMacbeth has many controls for making the camera
profile. Other than reading the help-file instructions and experimenting,
does anyone have any general or specific recommendations for how these controls
might be optimized for a daylight profile?
Thanks in advance,
Michael