It will take me a month of Sunday to digest all that but, in the interim...
The QP target was recommended on another photography list in response to
someone who was having difficulty getting proper CMYK conversions done
by her publisher for shots taken in a particular area inside a large
building. Shots outside and inside elsewhere would convert OK.
Strangely, the problematic shots (converted to ProPhoto in Lightroom)
looked OK to her on her monitor but gave the publisher fits on the CMYK
conversion.
Because they looked OK to the photographer on a (presumably) sRGB
monitor I wondered if there was some out of sRGB gamut (and therefore
invisible to the photog on her monitor) discrepancies between ProPhoto
and CMYK. The person who responded with the QP target suggestion was
obviously looking for a complex white balance solution.
It might also be that the room in question has a combination of
tungsten, halogen, fluorescent and sunlight and there's no simple
solution. I don't know what the actual lighting is.
Chuck Norcutt
On 10/18/2010 6:19 PM, Moose wrote:
> On 10/15/2010 3:46 PM, Chuck Norcutt wrote:
>> Anyone know anything about this? I saw it mentioned on another list today.
>> It supposedly builds a custom profile rather than a simple white balance.
>>
>> <http://www.bhphotovideo.com/c/product/286652-REG/QP_Card_GQP201.html#features>
>
> It looks to me like a variation on the ideas behind the ICC IT8 targets and
> ICC color profiles. In fact, Coca offers to
> build ICC profiles from QP card shots. Whether the QP software builds ICC
> format color profiles or it's own version, it
> is simply another of several ways of mapping the results of image capture to
> a known subject/source.
>
> For that use, I can't see where QP offers anything new, or particularly
> useful:
>
> 1. QP targets -
> a. Cost $16.65 apiece, incl. shipping from B&H in the US.
> b. Have 30 B&W and color reference patches on a 5.6 x 1.6" (142 x 40mm) card
> c. Do not appear to have any reference between their colors and correct
> colors. Either they have very special printing
> technology or just don't care about that level of detail.
> d. May require proprietary software to use, at least as intended.
>
> 2. IT8 targets:
> a. Cost $10 (or €10), incl. shipping (and get to me quicker from Germany than
> ground shipping from NY) from Wolf Faust.
> <http://www.targets.coloraid.de/>
> b. Have 228 color and 24 B&W reference patches on a 4x5" sheet of
> photographic paper.
> c. Come with a standardized color reference file documenting the differences
> between the target colors and correct
> colors and readable by apps that create ICC profiles.
> d. Are an international standard supported by many, many apps.
>
> I know which I would use. ;-)
>
> But let's back up a bit and look at the uses for such technology.
>
> 1. Correcting for inherent bias in an imaging system.
>
> This is the way in which I use ICC profiles. By profiling a film/scanner
> combo, sensor system or sensor system and
> RAW converter combo, I can correct for imperfections in their capture of
> reflective color in images.
>
> Used this way in light other than that in which the profile was made
> will show color differences from the same
> subject in the reference lighting. This way, subjects shot in 'magic
> hour' lighting looks like magic hour light
> shots, and so on.
>
> 2. Adjusting an image so that it appears to have been captured in different
> light than was actually illuminating it.
>
> In this use, one may, for example, shoot in tungsten light, including an
> image of the standard target in the same
> light, create a profile and use that profile to correct color so the
> resultant images appear to have been taken in
> daylight, with a highly color accurate camera.
>
> This way, subjects shot in 'magic hour' lighting looks like they were
> shot in midday light - except for the angle of
> the shadows, etc.
>
> 3. One may combine the two basic techniques.
>
> For example, one may use a simple neutral reference item to correct for
> difference from daylight in an image or set
> of images and a color profile to correct for camera/film specific
> irregularities in color response.
>
> So here comes the question. What do you want to do?
> ===============================================
>
> Correct for camera/film inaccuracies?
> -----------------------------------
>
> This is simple. Use an IT8 target and create ICC profiles to be used in
> scanning or in post. It's the cheapest, simplest
> and most accurate.
>
> Correct natural subjects?
> -----------------------------------
> Most of the time, I don't want to correct shots in other light to look like
> midday light. I carry a WhiBal around in my
> camera bag, but very seldom use it. Actually, I tend to forget it's there,
> but that' at least in part because of my
> prior experience with it.
>
> To me, the problem is simple to state, somewhat trickier to deal with. I
> shoot mostly outdoors, by natural light. Much
> of that shooting is in light from various degrees of cloud/overcast, shade,
> often colored by the foliage it has passed
> through and light from open sky or sunlight at times other than midday.
>
> The result is colors that aren't correct, in the color profile sense, but are
> perceptually correct. It's possible to
> correct a shot of a color target shot in any of these lights to look just
> like one shot in midday sun. But when I apply
> such correction to shots of natural subjects, they end up looking unnatural.
>
> There are all sorts of clues other than color alone that let us know in what
> kind of light a subject was shot. When a
> familiar subject is color corrected to a light much different from that in
> which is was captured, it tends to look
> "off", although it's often hard to say exactly why.
>
> My experience with the WhiBal is that it tends to overcorrect, at least
> perceptually, so the image is just wrong looking
> , to at least some extent, in other than relatively ordinary, midday light.
> When I have used it, I usually ended up
> applying it to a layer, then adjusting the opacity to find a middle ground
> that looks "right", like some combination of
> how I remember the scene and how millions of other, similar, subjects have
> looked to me.
>
> Since I can do essentially the same thing with the WB sliders in ACR, without
> taking extra shots with a reference item
> in the shot, I don't find much need for the WhiBal.
>
> I've thought for a long time that Dpreview's endless rants against Auto WB in
> tungsten light JPEGs from Canon cameras is
> misdirected. I'm sure Canon is capable of making cameras that output
> "correct" WB in incandescent light. So why don't
> they? I think the answer is that their research has shown that most users
> don't like the look of such shots fully
> corrected. We expect some shots to be warm, and like it that way. Other
> visual, perceptual and memory clues let us
> 'know' that daylight colors in those settings aren't right, aren't what we
> saw.
>
> In summary, I just don't see the need for anything more sophisticated than a
> neutral reference card or other gadget in
> any but a few specific lighting situations.
>
> Correct studio work?
> -----------------------------------
>
> Not my area, portrait, product or what have you photography. Still, I can't
> imagine the QP system can be as accurate as
> the ICC profile system, nor is it likely to be as good as the Greytag/Macbeth
> systems in use by so many pro studios.
>
> In Summary
> ===========
>
> I may be missing something, but I don't see where it offers anything
> worthwhile that isn't already available cheaper and
> likely more accurate. I can understand someone looking for a way to make an
> honest bit of dosh, but I don't see where
> this product succeeds in offering value.
>
> A. Critical Moose
>
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|