• Ultraviolet Photography
  •  

White-balancing UV-photos in Lightroom

White Balance
17 replies to this topic

#1 Nico

    Nico Chalwatzis

  • Members(+)
  • 121 posts
  • Location: Bensheim, Germany

Posted 19 May 2013 - 15:59

Since I have been made aware in this forum that the colours in my UV-photos are not “balanced”, I have experimented around this and have found a solution that seems acceptable to me and does not change my workflow to much. Most of the previous discussion has taken place in this post:
http://www.ultraviol...aw-development/
I use to shoot in RAW-format and was hit by the fact that Photoshop Lightroom (ACR) does not allow the values for the white balance to fall out of a certain range. Therefore, my final images looked always different from the previews on the rear screen of my camera. Using the DNG-profile-editor (http://wwwimages.ado...cumentation.pdf) can change that behaviour by creating a custom profile, as I learned in an Adobe forum. After creating a suitable profile I was able to reproduce colours similar to those described here:
http://www.ultraviol...s-reproducible/
The colours are now also pretty consistent with the (white balanced) in-camera JPGs that I also record as a reference.
I’m planning to show my UV-photos using this profile in the future, once for aesthetic reasons but also to comply with the standard established here. For me it looks quite consistent, but I’m happy to discuss and re-evaluate this. Here are some examples that I have developed in Lightroom recently, some more can be found on my blog (http://bee-colours.blogspot.de/)

Ranunculus ficaria: The visible image shows a yellow flower with a different shade of yellow in the middle:
Attached Image: NCH_P1070793130401.JPG

The centre of the flower is UV-dark. The outer parts of the petals appear yellow.
Attached Image: NCH_P1070796130401.JPG

Taraxacum officinale: The visible image appears in a quite even, saturated yellow.
Attached Image: NCH_P1080308130503.JPG

We find a similar yellow as above in the UV-image while the stamina appear dark.
Attached Image: NCH_P1080309130503.JPG

Hepatica nobilis: Blue to violet flowers in visible light.
Attached Image: NCH_P1080010130417.JPG

Light blue petals and very dark stamina in the UV-image:
Attached Image: NCH_P1080017130417.JPG

Geranium sanguineum (pink) and Hippocrepis comosa (yellow) in visible light photography:
Attached Image: NCH_P1080493130518.JPG

While the petals of G. sanguineum reflect UV light extremely well and appear almost pure white (whit a dark flower centre and dark veins), H. comosa is (almost) entirely UV dark.
Attached Image: NCH_P1080499130518.JPG

Alliaria petiolata: A white flower with a bit darker centre in visible light.
Attached Image: NCH_P1080398130506.JPG

In the UV-image the petals appear in a very dark blue.
Attached Image: NCH_P1080402130506.JPG

All images were taken with the EL-Nikkor 80mm/f5,6 enlarger lens (mostly at f8.0).

As most people on this forum are probably aware, the colours of the UV-photos are not (completely) predictable, since some yellow flowers can appear entirely black and blue flowers can have different intensities of UV reflections. The latter is also true for white blossoms. However, until now I haven’t found a flower that we see as blue which would turn yellow in the UV-image or vice versa (it still might exist).

#2 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 4,635 posts
  • Location: USA

Posted 19 May 2013 - 17:58

The UV filters all leak a bit of violet-to-blue, so some actual violet and blue is always being recorded, and so perhaps that's the reason blue/yellow tend to be "recoverable" in UV photos? I don't think you will be seeing blue/yellow "switch" in the UV. A channel cannot invert itself in photography. You have already observed that, indeed, Visible yellow or blue can go neutral/dark/light in the UV.

Chittka has made some categorization of Visible colours with their UV appearance. I'll try to dig out that paper and reference it. It is not a predictive categorization, rather more a generalization of what is possible.
Eventually you will get very good at "predicting" the UV outcome by family, etc. :) There can be surprises however - all part of the fun.

Interestingly, Bjørn and I had noticed for years that there was a subset of Hieracium which did not exhibit the central diffused UV-darkness of the typical Hawkweed, but were instead almost uniformly dark. Recently this subset got moved from Hieracium to Pilosella. (Example: http://www.ultraviol...nge-hawkweed-1/) So the UV appearance of a particular genus might eventually have some actual value in identification and taxonomy.

I am happy that Lightroom is working better for your colour balance needs when you want to standardize a UV photo.
Andrea G. Blum
Often found hanging out with flowers & bees.

#3 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 4,635 posts
  • Location: USA

Posted 19 May 2013 - 18:37

Bee colours: UV (primary), UV-Blue, Blue (primary), Blue-Green, Green (primary), Green-UV.

So why do you use Red in your presentation of Bee colours??
I'm just curious.
Andrea G. Blum
Often found hanging out with flowers & bees.

#4 nfoto

    Fierce Bear of the North

  • Owner-Administrator
  • 1,766 posts
  • Location: Oslo, Norway

Posted 19 May 2013 - 18:37

"However, until now I haven’t found a flower that we see as blue which would turn yellow in the UV-image or vice versa (it still might exist)."

Yellow flowers turning blue are uncommon, but do exist. Thus I am looking at an example right now where the outer ligules of Taraxacum friesii (yellow to the eye) become intense blue. I will post this later. Other examples are the yellow flower of Brassica rapa that goes dual colour (blue and yellow).

While we might be better positioned to predict UV appearance of flowers there till are regular surprises. In analogy to the 'Black Swan' concept we could entitle them 'Black Flowers'. Thus no reason not to add to the knowledge base of UV flowers.
Bjørn Rørslett

#5 nfoto

    Fierce Bear of the North

  • Owner-Administrator
  • 1,766 posts
  • Location: Oslo, Norway

Posted 19 May 2013 - 18:39

View Postannedi, on 19 May 2013 - 18:37, said:

Bee colours: UV (primary), UV-Blue, Blue (primary), Blue-Green, Green (primary), Green-UV.

So why do you use Red in your presentation of Bee colours??
I'm just curious.

Probably because mapping onto an ( R G B ) space is so easy to perform. One could argue these colours are 'false' anyway.
Bjørn Rørslett

#6 Nico

    Nico Chalwatzis

  • Members(+)
  • 121 posts
  • Location: Bensheim, Germany

Posted 19 May 2013 - 19:04

View Postannedi, on 19 May 2013 - 18:37, said:

Bee colours: UV (primary), UV-Blue, Blue (primary), Blue-Green, Green (primary), Green-UV.

So why do you use Red in your presentation of Bee colours??
I'm just curious.

Hi Andrea,
First of all, thanks for your comments! - Concerning your question: We know that (honey)-bees are trichromatic like humans. The difference is that the spectrum they see is shifted towards the shorter wavelength and includes near UV, as you know. Now, the dilemma is that UV for us is black, since our eyes cannot detect it. On the other hand bees are unable to see dark red. So, in order to give a representation of the full colour-spectrum that the bees can see, I found it logical to use the complete RGB colour-space.
“bee space”: UV violet blue green yellow black
“human space”: black violet blue green yellow orange-red
“sim. bee colours” violet blue green yellow orange-red black
As you can see in the last line, I am basically moving the human colours to “the left”, towards the shorter wavelengths to represent the “bee space” within the colours known to us. So, what we see as yellow-orange becomes red in the “sim. bee colours”, since it is roughly the longest wavelength that they can detect. On the other end of the spectrum UV becomes blue in that schema.
I will write that up with some illustration and post it, soon.

#7 nfoto

    Fierce Bear of the North

  • Owner-Administrator
  • 1,766 posts
  • Location: Oslo, Norway

Posted 19 May 2013 - 19:44

"So the UV appearance of a particular genus might eventually have some actual value in identification and taxonomy."

The Pilosella subset of the original Hieracium genus is a good example as Andrea points out. Another example are the two closely related species Erysimum cheiranthoides L. and E. hieraciifolium L. (Brassicaceae). They are easily separated in UV by E. cheiranthoides having a bright yellow/black bull's-eye UV signature in contrast to E. hieraciifolium which appears very UV dark all over. In the genus Potentilla (Rosaceae) there are numerous examples of related species with highly different UV signatures as well.
Bjørn Rørslett

#8 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 4,635 posts
  • Location: USA

Posted 19 May 2013 - 19:56

Nico: I wasn't very clear. There have been several schema for bee-vision in the literature, and I am familiar with how the shift of colours is made for your bee vision simulation. I was just curious about why you chose this model. The answer tends to be: well, why not? To which I agree. :)
I have worked on various bee vision schema myself - currently I favor one which uses light green & dark green, light cyan & dark cyan, light blue & dark blue. But I gave up modeling bee vision because I felt I was neglecting bat vision, butterfly vision, insect vision and reindeer vision. Not enough time when collecting is the first priority.

Bjørn: I thought we were tallking about a complete channel inversion here, rather than yellow or blue going partially blue or yellow. Indeed I just posted a yellow Jerusalem sage which goes dual blue/yellow in the UV. And also that yellow Iris which does the same.
BTW, my Brassica rapa does not go dual coloured.

Bjørn: (Just added.) Yes, I do hope all this collection does sometime spur some botany graduate student or botany professor towards some taxonomic investigations. It would be so very interesting to do that myself, but alas I have no access to any lab in which various deeper studies could be performed - especially on the molecular level.
As it is said: "We also serve who simply collect UV-signatures".
Andrea G. Blum
Often found hanging out with flowers & bees.

#9 Nico

    Nico Chalwatzis

  • Members(+)
  • 121 posts
  • Location: Bensheim, Germany

Posted 19 May 2013 - 20:12

View Postannedi, on 19 May 2013 - 19:56, said:

Nico: I wasn't very clear. There have been several schema for bee-vision in the literature, and I am familiar with how the shift of colours is made for your bee vision simulation. I was just curious about why you chose this model. The answer tends to be: well, why not? To which I agree. :)
I have worked on various bee vision schema myself - currently I favor one which uses light green & dark green, light cyan & dark cyan, light blue & dark blue. But I gave up modeling bee vision because I felt I was neglecting bat vision, butterfly vision, insect vision and reindeer vision. Not enough time when collecting is the first priority.

Andrea: It was not exactly my intention to say: Why not … ;-) :)
I guess I was trying to say: If I don’t use red, then I narrow down the colours space which I think is not accurate if a species is trichromatic like us. Also I think that leaving the colours “in order” is a sensible thing to do, therefore the shift. Others are putting UV in the red-channel, which is another way of using the full colour space without so much red in most images.
However, you are certainly right, since there are many other species on the planet that have a broader or a narrower light sensitivity, one might argue how much sense this all makes …
In my mind it makes sense if we use the colour space of the organisms that a certain flower is trying to attract (that it has co-evolved with). I’m aware that it gets complicated if we consider flowers that are pollinated by both bird and insects …
So, yes whatever one does there are other ways how it could be done and one can find reasons to justify them … B)

#10 nfoto

    Fierce Bear of the North

  • Owner-Administrator
  • 1,766 posts
  • Location: Oslo, Norway

Posted 19 May 2013 - 20:24

I think any kind of colour coding scheme will only bring us half way to understanding the flower from the viewpoint of a visiting pollinator. The more I work with UV signatures the more I am struck by their diversity and the additional components of the signature that do not involve colour rendition per se.
Bjørn Rørslett

#11 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 4,635 posts
  • Location: USA

Posted 19 May 2013 - 20:29

Nico, yes, those are all excellent points!
And I thank you for taking the time to explain your model further. I appreciate it!! :)

It think our being able to photograph in UV helps us understand a little better how other organisms sense the world and deal with their portion of it. We cannot ever fully "see" as they "see", but we become closer to them.

****

Here's the link to the yellow flower which becomes dual yellow/blue in UV.
http://www.ultraviol...jerusalem-sage/

But we are still looking for a complete reversal example !!! :)
Whoever finds it first must promise to buy a celebratory round someday!!
Andrea G. Blum
Often found hanging out with flowers & bees.

#12 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 4,635 posts
  • Location: USA

Posted 19 May 2013 - 20:37

Bjørn, as we know the UV signature also references certain underlying chemistry in many flowers, yes? And that chemistry may or may not have anything at all to do with pollination strategies. It is very popular currently to focus on "pollination marks", but I think the Iridescence, Conical Cells and UV Chemistry and their interactions with the UV Signature might be an equally fruitful arena for further investigation.
(I do hope I have properly worded all that. We former mathematicians are not known for our botanical vocabulary. We can be taught, however. :) :) )
Andrea G. Blum
Often found hanging out with flowers & bees.

#13 nfoto

    Fierce Bear of the North

  • Owner-Administrator
  • 1,766 posts
  • Location: Oslo, Norway

Posted 19 May 2013 - 20:41

Well, I am not that well versed in flavonoid chemistry myself. However, given the fact that albino forms tend to respond in the same manner to UV as their normal coloured forms, the casual agent might not be visible flower colour at all. Conical cells etc. act more like modifiers to the overall UV signature. Most of the publications I have read on conical cells do not mention UV properties. So, evidently, in the field of UV signatures most of the work lies ahead of us.
Bjørn Rørslett

#14 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 4,635 posts
  • Location: USA

Posted 19 May 2013 - 20:49

Go for it !!!
*****

Nico, that Hippocrepis comosa is a new one for me. It is a lot like this US Lotius corniculatus.
http://www.ultraviol...dsfoot-trefoil/
Andrea G. Blum
Often found hanging out with flowers & bees.

#15 nfoto

    Fierce Bear of the North

  • Owner-Administrator
  • 1,766 posts
  • Location: Oslo, Norway

Posted 19 May 2013 - 20:52

Many members of tte Fabaceae encountered by me apparently are UV dark, but there are exceptions.

Let us join forces and add as many signatures as possible. That will make broader generalisations more feasible than at present.
Bjørn Rørslett

#16 nfoto

    Fierce Bear of the North

  • Owner-Administrator
  • 1,766 posts
  • Location: Oslo, Norway

Posted 20 May 2013 - 22:02

The Broom (Cytisus scoparius) is yellow in visible light and becomes all shades of blue in UV. Perhaps this is the plant to require a celebration?

http://www.ultraviol...coparius-broom/
Bjørn Rørslett

#17 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 4,635 posts
  • Location: USA

Posted 20 May 2013 - 23:19

WOW !!!! You got a complete reversal !!!! This is totally awesome !!!! :D :D :D :D :D
I really, truly did not think this could happen. :rolleyes:
But I'm happy it did. I like surprises. ;) :lol: B)
Andrea G. Blum
Often found hanging out with flowers & bees.

#18 nfoto

    Fierce Bear of the North

  • Owner-Administrator
  • 1,766 posts
  • Location: Oslo, Norway

Posted 21 May 2013 - 23:41

I just found another reversal for you, Andrea. This is Hyacinthoides non-scripta, the English Blue-Bell. Going from blue (as suggested by its vernacular name) in visible to yellow in UV. I'll post this species shortly.
Bjørn Rørslett