• Ultraviolet Photography
  •  

Profiled and non-Profiled White Balance of UV False-Color

10 replies to this topic

#1 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 9,116 posts
  • Location: UVP Western Division, Santa Fe, New Mexico

Posted 13 May 2021 - 21:59

I was working on Oenothera pallida (Pale Evening Primrose) today and was curious to review the white balance appearance of this false-blue flower both with and without the application of a camera color profile.

A visible color profile photo (not shown) was made using a D610 + UV-Nikkor 105/4.5 + Baader UV/IR-Cut filter to photograph a Color Checker card. Three minutes of work in Photo Ninja (tutorial reference 1st link below) then provided a saveable set of accurate visible colors to apply to any visible photos made with this gear.

I would say, roughly speaking, that such a camera profile for a converted camera certainly gets you into the visible color-accuracy ball park and very likely well on your way to third base. Those wishing a color-accuracy homer will need to understand and conquer a great deal of additional technical variables. I'm happy stopping at 3rd.

However, when rendering a reflected UV photograph, we have no real definition of "color-accurate" for the false colors. I think that this gives us license to either apply or not-apply a visible color profile prior to white balancing the UV photo. In earlier years (see 2nd link) I tried to make the case for why we should use these camera visible profiles in UV photos. These days I am not so insistent. Use or do not use a converted camera vislble profile for UV photos as you please, OK? :grin: :cool:

Reference: Photo Ninja: How to Make a Custom Light Profile
See also: Colour Calibration in a Converted Camera

The Primrose raw photo used for this demo was converted in Photo Ninja with the usual simple tweaks for black/white points and some detail enhancement. Sometimes I also use the shadow slider, iillumination slider, and/or contrast sliders. I keep the color intensity between 50 - 60 for the Plain setting. White balance was made against a Spectralon background. (It is so fast and easy to convert in PN that I rarely use anything else these days. I only wish for a brush so that I could selectively apply the detail slider settings.)

After white balance was made in PN's Color Correction tool, I selected No Profile from the Light Source drop-down. This result is on the left. After saving that, I selected the D610 profile to get the result in the middle. The 3rd result on the right used the generic PN Daylight profile.

I always like to sample colours and compare them to their fully saturated versions, so I proceeded to do that for my 3 Primrose renditions. Obviously sampled results will vary depending on where you sample, but we still can get a feel for how the profiles are behaving. The specific Hue/Saturation/Brightness and RGB numbers are from 5x5 samples made at the same place on each rendition.

The predominant channel in each rendition is blue. Blue is considered to be 240° on the color wheel. Here we get 225°, 239° and 237°. So the first color sample is about 15°, one quarter of the way closer to Cyan than the other two samples which are almost "pure" blue, although desaturated.

So what color should our UV false-blue be? You decide. :lol:

Attached Image: oenotheraPallida_uvBaad_sb14_20200609aveLasNubes_20620_studyRes.jpg
Andrea G. Blum
Often found hanging out with flowers & bees.

#2 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 9,116 posts
  • Location: UVP Western Division, Santa Fe, New Mexico

Posted 13 May 2021 - 22:09

Color is just so weird. Desaturated blue is so weird.
To my (good) eye, the left-most flower looks like actual blue, the middle flower looks like blue with a dash of purple and the right-most flower looks like blue with two dashes of purple. I just cannot get my head to look at that middle flower is being the bluest.

Oh well.

Aside: For those of you who remember that I currently have (not by choice) a good eye for color (the left) and a bad eye for color (the right), I note that with the good, color-accurate eye these three renditions look distinctly different. With my bad color eye, it is more difficult to distinguish between the three. All three look duller (less saturated) with the bad eye. I'm being thankful to have eyes, so not worried about this.
Andrea G. Blum
Often found hanging out with flowers & bees.

#3 Stefano

    Member

  • Members(+)
  • 2,187 posts
  • Location: Italy

Posted 13 May 2021 - 22:11

View PostAndrea B., on 13 May 2021 - 22:09, said:

Color is just so weird. Desaturated blue is so weird.
To my (good) eye, the left-most flower looks like actual blue, the middle flower looks like blue with a dash of purple and the right-most flower looks like blue with two dashes of purple. I just cannot get my head to look at that middle flower is being the bluest.
True for me too. Color vision is extremely complex, way more than one would think.

#4 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 9,116 posts
  • Location: UVP Western Division, Santa Fe, New Mexico

Posted 13 May 2021 - 22:16

I would like to hear from any color blind members who could comment on the differences, if any, that they see in the three renditions above depending on their type of color blindness.
Andrea G. Blum
Often found hanging out with flowers & bees.

#5 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 9,116 posts
  • Location: UVP Western Division, Santa Fe, New Mexico

Posted 13 May 2021 - 22:54

So I should try this with a yellow flower.
Andrea G. Blum
Often found hanging out with flowers & bees.

#6 Andy Perrin

    Member

  • Members+G
  • 4,425 posts
  • Location: United States

Posted 13 May 2021 - 22:59

Heh. I'm probably the only reply you'll get on that one, Andrea, but my take is:

Left: very different, Center and Right: almost but not quite the same (in fact if you hadn't told me they were different, at a glance I would have said they were identical).

My color blindness is of the red-green variety, although I have a little bit of red-green sensation and I can tell the difference about 80% of the time. I have the most trouble with pastels. Pink vs. light green is hard, and sometimes fluorescent green looks like fluorescent orange to me.

By the way, bear in mind that your color profile is specfic to the Baader IR/UV cut filter. You would have a different profile with the original BG glass that came with the camera, since the Baader has no red attenuation (IIRC).

Edited by Andy Perrin, 13 May 2021 - 23:02.


#7 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 9,116 posts
  • Location: UVP Western Division, Santa Fe, New Mexico

Posted 13 May 2021 - 23:56

Andy, thank you for the input.

By the way, bear in mind that your color profile is specfic to the Baader IR/UV cut filter. You would have a different profile with the original BG glass that came with the camera, since the Baader has no red attenuation (IIRC).

Yep.
I always try to be careful to mention what gear is in use for such examples. "-)

That Baader UV/IR Cut has a very "squared-off" right-hand transmittance. And that is exactly why a color profile is needed to restore the proper vislble colors for a conversion. It is not always enough to just white balance. I will add that sometimes the differences can be subtle between a white balanced visible photo and a white balanced, color profiled visible photo. And sometimes not so subtle. (So many variables at play!)

With whatever UV/IR Cut filter one uses on a conversion, color profiling should be done. You can't just slap any old BG filter on a conversion and think you are necessarily getting the original colors back.

I was always hoping that all those internal filters removed during conversion could perhaps someday be measured so we would know more about which manufacturer was using what kind of IR blocking.
Andrea G. Blum
Often found hanging out with flowers & bees.

#8 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 9,116 posts
  • Location: UVP Western Division, Santa Fe, New Mexico

Posted 14 May 2021 - 04:58

I tried the same thing with a false-yellow and blue flower. The false yellow also varies between the Profiled and No Profile versions .

Gear: D610 full spectrum + UV-Nikkor 105/4.5 + BaaderU UV-Pass Filter
Flower: Verbesina encelioides, Cowpen Daisy
VIew: Abaxial

Conversion with No Profile
Attached Image: verbesinaEncelioidesAbaxial_uvBaad_sb14_20200702trenViaAveEldor_21895pnNoProfile.jpg

Conversion with D610 Color Profile for UV-Nikkor & BaaderU
Attached Image: verbesinaEncelioidesAbaxial_uvBaad_sb14_20200702trenViaAveEldor_21895pnProfiled.jpg

Conversion with generic Daylight Profile
Attached Image: verbesinaEncelioidesAbaxial_uvBaad_sb14_20200702trenViaAveEldor_21895pnDaylight.jpg
Andrea G. Blum
Often found hanging out with flowers & bees.

#9 Andy Perrin

    Member

  • Members+G
  • 4,425 posts
  • Location: United States

Posted 14 May 2021 - 05:08

Profiled look nicer. Both.

#10 UlfW

    Ulf W

  • Members+G
  • 1,837 posts
  • Location: Sweden, Malmö

Posted 14 May 2021 - 11:50

The rear side of flowers is interesting. I must add some focus on them i the future.

Profiles are normally used for correcting images in the visual range and quite useful for full range converted cameras when using some BG glass only.

I do not think there are any correct profiles for false colours generated by filters like the BaaderU.
Visually I agree with Andy and think the images processed via the visual D610 profile are the nicest looking.

Sometimes breaking the intended rules still leads to visual improvements.

I sometimes do similar things like when I apply a geometrical correction profile from a Sigma Fisheye lens on my Sunex images, as there are no correction file available for my lens.

Edited by UlfW, 14 May 2021 - 11:51.

Ulf Wilhelmson
Curious and trying to see the invisible.

#11 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 9,116 posts
  • Location: UVP Western Division, Santa Fe, New Mexico

Posted 14 May 2021 - 18:53

Profiled look nicer. Both.

Interesting, Andy. I thought so also, but I'm not exactly sure why. After white balance, the Daylight profile isn't too bad either in PN.



Ulf, yes, the abaxial side of flowers can be different from the top side (adaxial). Quite interesting and worth taking the time for for anyone who enjoys photographing UV floral signatures.


I do not think there are any correct profiles for false colours generated by filters like the BaaderU.

Indeed. We are free to deal with UV false colors in any way we want to.
Unless, of course, one is posting in the UVP botanical section where we ask for a "standard" white balance in order to be better able to compare & contrast different flowers. :lol:

Many converters do "force" a color profile either directly or indirectly.
I'm thinking that Photo Ninja is unusual because it does offer a No Profile option and because it offers such an easy profiling option using the CC card.
Andrea G. Blum
Often found hanging out with flowers & bees.