• Ultraviolet Photography
  •  

UVIVF Possible Visible Light Leak?

Fluorescence White Balance
57 replies to this topic

#41 Cadmium

    Member

  • Members(+)
  • 3,031 posts

Posted 25 February 2019 - 15:52

And UV 400 graph too
Attached Image: Firecrest_UV_400.jpg

#42 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 7,541 posts
  • Location: USA

Posted 02 March 2019 - 19:56

I heard from various people that I don't "really need" it and I keep feeling like I should really get one but I was hearing conflicting things, or at least that's what I was taking from it, maybe I was misunderstanding some people

A stock Nikon D810 has a UV/IR blocking filter. But we do not know how much UV or IR that filter blocks. So if that filter is letting any UV light pass between 390-400, then you might be getting some blue/purple contamination. Our cameras record stray UV around 400nm as either blue or purple.

If anybody has told you it is OK to shoot UVIVF without a UV-blocker on the lens, well then, they are not correct. Proper fluorescence photography always requires double filtration: a filter on the illumination source and a filter on the taking lens. This is certainly not to say that we here on UVP all follow proper procedure all of the time. :D I experimented with UVIVF for a long time without using a filter on my UV-LED flashlight. Always did use the Baader UV/IR-Cut on my lens.

Scientific fluorescence photography uses very narrowband filtration for UVIVF. Our foto filters are very broad by comparison. So we will never get an absolutely, totally "pure" fluorescence photo. But we can and should attempt to minimize the leakage.


1) Your blue/purple overload may result from the converter you are using? What are you using to convert your raw NEFs from the D810?

2) For shooting, set the D810 white balance to K10000 and use a Standard [0] picture control. (Okay to add sharpening).
Yes, I know that internal settings only apply to the jpg, but you do need to check what you are recording while doing it. So getting as close as possible to fluor neutrality is helpful.

3) While in the dark shooting fluor, your vision may change. So be sure to observe the emitted colors at the very beginning of your shooting session while you still have some photopic vision working. I don't really know if our eyes ever convert to total scotopic vision when shooting fluor in the dark because, after all, there is some emitted visibile light from the subjects. But best to observe at the beginning, not the end of the session. And then hope your mental color map holds well enough so that you can aim for it during raw conversion. :rolleyes:

Hope this helps. Don't give up! We'll figure it out.
Andrea G. Blum
Often found hanging out with flowers & bees.

#43 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 7,541 posts
  • Location: USA

Posted 02 March 2019 - 20:04

Addition to preceding.

Todd, two things.

I have a D810. I could try to make a UVIVF white balanced NEF with my Convoy and make it available for you to load into your camera to use as a WB setting. (Do you know how to do that?) I wonder if this would be useful or help?

If you can arrange for me to download a raw NEF of your flower, I'll run it through Raw Digger and see whether any red is being recorded in the chlorophyll area. And I can try other converters for you to see if they bring out the red. Let me know.
Andrea G. Blum
Often found hanging out with flowers & bees.

#44 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 7,541 posts
  • Location: USA

Posted 02 March 2019 - 20:09

hmm.....turns out we can post raw nefs here on UVP.
TEMPORARILY.
They cannot be displayed. But can be downloaded.

Posting NEF files eats up a huge amount of space. So I would have to remove such a link after 24-hours or so. This does let us share NEFs or other raws if we want to.

So I might try to make that D810 WB foto and post it. Gotta wait til dark though. B)

Obligatory cat photo.
[Added: Linked as a NEF upload trial. TEMPORARILY.]
[attachment now removed]

Attached Files


Andrea G. Blum
Often found hanging out with flowers & bees.

#45 Cadmium

    Member

  • Members(+)
  • 3,031 posts

Posted 02 March 2019 - 20:38

Ulf posted this link, showing the stock D810 blocking.
https://diglloyd.com...sion-curve.html

#46 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 7,541 posts
  • Location: USA

Posted 02 March 2019 - 22:42

And we hope that the chart is accurate. Unless someone (Kolari or whoever) is willing to explain their spectro setup, then we take all such charts with the proverbial grain o salt because so many people do not make proper spectral measurements.

I know -- I'm always such a skeptic! Let's say that the chart probably is reasonably accurate.

In which case the problem lies in the finish not the effort. I.E., Todd's fluorescent colors, red in particular, are not being brought out by the conversion.
Andrea G. Blum
Often found hanging out with flowers & bees.

#47 Cadmium

    Member

  • Members(+)
  • 3,031 posts

Posted 03 March 2019 - 01:27

I agree with you, you are right technically, all I was saying previously is that I don't think his red problem is caused by the lack of a lens barrier filter given the lens he is using and the D810 graph.

Not sure what you mean about his red is not being brought out by the conversion?

#48 JMC

    Member

  • Members(+)
  • 829 posts
  • Location: London, UK

Posted 03 March 2019 - 13:39

View PostAndrea B., on 02 March 2019 - 22:42, said:

And we hope that the chart is accurate. Unless someone (Kolari or whoever) is willing to explain their spectro setup, then we take all such charts with the proverbial grain o salt because so many people do not make proper spectral measurements.

I know -- I'm always such a skeptic! Let's say that the chart probably is reasonably accurate.

In which case the problem lies in the finish not the effort. I.E., Todd's fluorescent colors, red in particular, are not being brought out by the conversion.
I'm just doing some work at the moment using a Canon EOS 5DSR (unmodified) for fluorescence. I am pretty certain the stock filters over the sensor are not blocking all the UV, and lead to a sizable contribution from reflected UV in fluorescence images, if a suitable UV blocking filter is not added to the lens. Once I have finished it I'll share my findings on here. Also, looking at the Kolari sensor filter transmission profiles, they look to be measurement points every 5nm. From what I'm seeing, that could lead to issues with averaging of the data, and be underestimating the potential problem of leaks. It'll make more sense when it's written up and shared (I hope).

Edited by JMC, 03 March 2019 - 15:42.


#49 JMC

    Member

  • Members(+)
  • 829 posts
  • Location: London, UK

Posted 03 March 2019 - 15:52

View PostJMC, on 23 February 2019 - 14:57, said:

Ok thanks Steve, and I fully I agree we need charts for these Firecrest filters. I have the feeling it isn't going to be a nice vertical drop at 400nm for their 400 one, and like wise for the 370 one. If I end up getting any I will of course post my findings.

EDIT - ordered one of the 400nm UV filters, so will see what it's like when it gets here.
My Firecrest 400nm UV filter turned up and I have measured the transmission on it. Used my Ocean Optics FX spectrometer. Measured spectra in two halves - up to 410nm with the Deuterium lamp only, and above 410nm with both lamps. Then pasted the two halves together (they matched up very well at the join). This was done to minimise stray light effects at the short wavelength side of the cut off. Here's how it looks;
Attached Image: Firecrest 400 UV filter trans.jpg

Steve - looks very similar to the graph you shared (where did you get from?). On mine there is a little more tail on the short wavelength side of cut-off. While I tried to minimise stray light effects, if I'm still getting any that would account for the tail I saw.

I stand corrected - I fully expected this not to be a sharp cut off, but I am pretty impressed. If I get one of the 370 ones I'll test that too.

#50 UlfW

    Ulf W

  • Members(+)
  • 848 posts
  • Location: Sweden, Malmö

Posted 03 March 2019 - 18:07

Jonathan-

To further examine the tail below 400nm of the Firecrest 400nm UV you could pre-filter the deuterium light with your SEU2-filter, and use that as the modified light source.
The stack of SEU2 and Firecrest 400nm UV will not transmit any unwanted light causing much stray light errors.
Naturally the data will only be valid for the transmission range of the SEU2.

Edited by UlfW, 03 March 2019 - 18:08.

Ulf Wilhelmson
Curious and trying to see the invisible.

#51 Cadmium

    Member

  • Members(+)
  • 3,031 posts

Posted 04 March 2019 - 01:06

Jonathan, Thanks for that scan. I got the graphs from them. They have them, just not anywhere in their webpages evidently.
Seems like neither of those are that great to use for a UV cut barrier lens filter for UVIVF ?
I also have a Zeiss T* UV filter. Have not tried it for much yet, can't find the graph for it online at the moment, but I think I saw it online once.

#52 JMC

    Member

  • Members(+)
  • 829 posts
  • Location: London, UK

Posted 04 March 2019 - 08:54

Ulf. Good idea. I'll try that next time I've got the equipment on.

Steve. Ah right. At least they have the data. As for use for a UV cut for UVIVF, I suppose it depends how the light source itself is filtered (and what the source is). But yes, I'd want something with a higher cut-off for fluorescence myself. Some Zeiss T* UV filter data;
https://www.lensrent...r-spectrograms/

Not the best graph, but might be worth a try.

#53 Cadmium

    Member

  • Members(+)
  • 3,031 posts

Posted 04 March 2019 - 20:59

View PostJMC, on 04 March 2019 - 08:54, said:

Ulf. Good idea. I'll try that next time I've got the equipment on.

Steve. Ah right. At least they have the data. As for use for a UV cut for UVIVF, I suppose it depends how the light source itself is filtered (and what the source is). But yes, I'd want something with a higher cut-off for fluorescence myself. Some Zeiss T* UV filter data;
https://www.lensrent...r-spectrograms/

Not the best graph, but might be worth a try.

Well, the light source will have UV in it, if you want to block reflected UV from entering the lens (for UVIVF shots) then it is best to do that above 400nm, even 420/435nm.
The KV-418 works nice, if you can find one.
If the Baader UV/IR-Cut filter leaks UV, and that is a concern, then it looks like you will get slightly more UV leaking with either of these Firecrest filters.
I tried Baader UV/IR-Cut stacked with GG420 and GG435 behind the Baader (between the Baader and the camera lens), and that worked good for me.

Here is where I first heard about the Zeiss T*.
http://www.ultraviol...dpost__p__25416

I saw a graph of it somewhere, which I can't find right now, which looked better than the one on the page you linked to above.
There is probably some UV/IR cut or UV cut filter that does cut slightly above 400nm, don't know.

Frankly, if some has the Baader UV/IR-Cut, I might try the GG420/435 between the Baader and the lens.

Edited by Cadmium, 04 March 2019 - 21:00.


#54 cdhgamer

    Member

  • Members
  • 24 posts

Posted 06 March 2019 - 02:39

View PostCadmium, on 04 March 2019 - 20:59, said:

Well, the light source will have UV in it, if you want to block reflected UV from entering the lens (for UVIVF shots) then it is best to do that above 400nm, even 420/435nm.
The KV-418 works nice, if you can find one.
If the Baader UV/IR-Cut filter leaks UV, and that is a concern, then it looks like you will get slightly more UV leaking with either of these Firecrest filters.
I tried Baader UV/IR-Cut stacked with GG420 and GG435 behind the Baader (between the Baader and the camera lens), and that worked good for me.

Here is where I first heard about the Zeiss T*.
http://www.ultraviol...dpost__p__25416

I saw a graph of it somewhere, which I can't find right now, which looked better than the one on the page you linked to above.
There is probably some UV/IR cut or UV cut filter that does cut slightly above 400nm, don't know.

Frankly, if some has the Baader UV/IR-Cut, I might try the GG420/435 between the Baader and the lens.



Thanks everyone for your continued interest in this, sorry I was away for a bit but I currently am taking care of two family members who are recovering from surgeries so I've been away from my computer.



Cadmium, I shot with the 400 nm Firecrest the other night and didn't see much difference honestly. The dark room I've been shooting in has a few walls that are white, would those be causing possible UV reflection then that's tinting the flowers a more purple/blue? Obviously I'd be interested in trying higher nm filtering, but just curious on contamination there.



View PostAndrea B., on 02 March 2019 - 20:04, said:

Addition to preceding.

Todd, two things.

I have a D810. I could try to make a UVIVF white balanced NEF with my Convoy and make it available for you to load into your camera to use as a WB setting. (Do you know how to do that?) I wonder if this would be useful or help?

If you can arrange for me to download a raw NEF of your flower, I'll run it through Raw Digger and see whether any red is being recorded in the chlorophyll area. And I can try other converters for you to see if they bring out the red. Let me know.



Andrea, that would be fantastic if you could make that white balance NEF, do you suppose that would also work if I color matched it in the curves tool of photoshop for the midtones? I can always do it the other way, but I'm just curious.



I also am not able to include the raw file in the forum's file system, there's a 30 MB global upload limit and the file is 36MB, so I'll include it in a Dropbox link. The red fluor wasn't crazy on the stamen by the way, but not too subtle to notice either.



https://www.dropbox....BGnq-UdhVa?dl=0


Edited by cdhgamer, 06 March 2019 - 02:45.


#55 UlfW

    Ulf W

  • Members(+)
  • 848 posts
  • Location: Sweden, Malmö

Posted 06 March 2019 - 05:07

Thank you for sharing the raw file.

I made a quick check with white balancing it with FastRawViewer https://www.fastrawviewer.com/ and posting screen shots.
The program is an excellent raw-viewer.
For good final conversion other programs are needed.

By selecting different points for the white balance I get different results.
There might be one similar to what you saw when shooting the image.

1:
Attached Image: Screen Shot 2019-03-06 at 05.41.23.png

2:
Attached Image: Screen Shot 2019-03-06 at 05.41.08.png

3:
Attached Image: Screen Shot 2019-03-06 at 05.41.42.png

4:
Attached Image: Screen Shot 2019-03-06 at 05.42.54.png

5:
Attached Image: Screen Shot 2019-03-06 at 05.42.05.png

Edit: Added number labels to the pictures.

Edited by UlfW, 10 March 2019 - 05:23.

Ulf Wilhelmson
Curious and trying to see the invisible.

#56 cdhgamer

    Member

  • Members
  • 24 posts

Posted 10 March 2019 - 00:09

What I saw was kind of a mix of 3 & 5 UlfW

#57 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 7,541 posts
  • Location: USA

Posted 14 March 2019 - 05:48

I finally had some time to look at the lily photo. I am sorry it took so long! I've had my hands full with domestic matters for the last couple of months.

Here is the original conversion in which a Daylight white balance was applied.
Attached Image: lily.jpg


This is the Raw Digger "raw composite" which has had no color corrections applied (i.e., no white balance). This shows us that there were really no visible red emissions recorded from the flower itself. In the deep shadows around the flower, there was some red fluorescence recorded from the chlorophyll of the lily foliage. This can only be seen by cranking an exposure slider up to max.
Attached Image: lily_rawComp.jpg


Here are the R, G and B histograms for the entire photo. (The two green channels are averaged.) I truncated some of the data representation on the left because it was for the black background. We are interested here only in the flower data. The red histo shows primarily noise and deeply shadowed red leaf fluor, as mentioned.
Attached Image: lily_rawComp_histo.jpg


The raw composite can be exported from Raw Digger as a TIF for further processing. Add a little bit more light, contrast and detail enhancement and it makes for quite an acceptable rendition of the lily's fluorescence. Of course, not having actually seen the fluorescence, I can't judge just how bright to make the photo.
Attached Image: lily_rawCompPn.jpg





Here are exports of the red, green and blue channels from Raw Digger. The blue channel is the brightest (no surprise), followed by the green channel. There is not much captured in the red channel.
Attached Image: lily_chanBlue.jpg
Attached Image: lily_chanGreen.jpg
Attached Image: lily_chanRed.jpg




I wanted to examine some selected areas of the photo. First I looked at a central selection taken across the petals and stamens. The selected area is shown by the white rectangle.
Attached Image: lily_selection02.jpg
Attached Image: lily_selection_histo02.jpg


Next I looked at a selection made only on the stamens. The selection is shown by the white rectangle. There is no red fluorescence on those stamens. The fluorescence is a blue-green, more blue than green.
Attached Image: lily_stamen01.jpg
Attached Image: lily_stamen_histo01.jpg


Finally I made a tiny selection over one of the fluorescent fibers. This was just for fun. (We are all always lamenting the presence of those dang-rabbity fibers and dust specks!) They fluoresce blue with some green, but not enough green to really call it cyan. That is one way to determine whether the white balance of your UVIVF photo is correct -- check whether the fibers and dust specks are fluorescing a bright blue with a slight move towards green. The fiber fluorescence is very strong. You can see that the blue and green channels have hit the right-hand wall and begun to blow out.
Attached Image: lily_fluorFiber01.jpg
Attached Image: lily_fluorFiber_histo01.jpg




So, the reason for the blue cast in your UVIVF photographs is because you have used a Daylight (or possibly an Auto) white balance setting. It doesn't matter what converter is used. Here is an example using Photo Ninja with a Daylight white balance applied. It produces the same blue. (I couldn't spend too much time on this conversion. It is not at all as good as your original for sharpening, detail and contrast.)
Attached Image: lilypnDaylight.jpg


Is it possible in Photo Ninja to move the white balance setting away from all that blue? Of course. But moving the Temperature slider to the highest value 15000 does not fully relieve the blue misery. Still, this is a little bit more accurate version of the fluorescence.
Attached Image: lilypnTemp15000Tint-50.jpg


If I turn off all color correction in Photo Ninja for the NEF, then I get a conversion much truer to the raw composite. There are some variations in contrast and white/black point settings, but this version is basically the same as the edited version of the above raw composite. (3rd photo above, just after the first histogram.)
Attached Image: lilypnNoColorCorr.jpg




Where does the "color truth" lie for this fluorescent lily? Somewhere in between the raw composite version and the Temperature 15000 version. B)


THANK YOU for letting me play with your photograph! :)


.
Andrea G. Blum
Often found hanging out with flowers & bees.

#58 cdhgamer

    Member

  • Members
  • 24 posts

Posted 17 March 2019 - 04:21

Thanks so much Andrea!