• Ultraviolet Photography
  •  

Colorchecker

31 replies to this topic

#1 bobfriedman

    Member

  • Members
  • 200 posts
  • Location: Massachusetts

Posted 12 January 2018 - 11:38

i don't expect too much here other the pigment variation in the UV but i thought i post the set anyway.

UV - Illuminated by two Blak-Ray 100A
Nikon D800M ,Nikon El-Nikkor 105mm f/5.6 ,Ultraviolet Baader U
0.62s f/11 at 105.0mm iso250
Posted Image

Monochrome Conversion Visible - D55 light source

Nikon D800M ,Carl Zeiss Apo Sonnar T* f/2 135mm
1s f/5.6 at 135.0mm iso100
Posted Image

D800 RG830

Nikon D800E RG830 ,Nikkor 105mm f/2.8 Ai-S Micro
1/60s f/5.6 at 105.0mm iso320
Posted Image

D800 Stock Camera

Nikon D800 ,Carl Zeiss Apo Sonnar T* f/2 135mm
0.60s f/5.6 at 135.0mm iso160
Posted Image

Edited by bobfriedman, 12 January 2018 - 11:39.


#2 JMC

    Member

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

Posted 12 January 2018 - 11:46

The UV one grey tiles confuses me slightly - I've measured UV reflectance of these things and while they are different in UV to what they are in visible, they rank the same way. From left to right they should get darker, but the 3rd one is lighter than the second. How even was the UV lighting across the image? The reason I ask is that the border of the chart seems to get lighter towards the centre, so i'm wondering if the UV was more intense towards the centre.

#3 bobfriedman

    Member

  • Members
  • 200 posts
  • Location: Massachusetts

Posted 12 January 2018 - 13:37

No I don’t think it was very even at all

#4 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 6,216 posts
  • Location: USA

Posted 12 January 2018 - 21:23

That grey patches on my Color Checker Passport do not remain in a brightness rank order when photographed in UV. Here are some fotos of it under various UV-pass filters. Look at the top row -- that's the grey patch row. The black patch is on the left, the white on the right. The patch which is 2nd from right next to the white one is brighter in UV than it "should be".
http://www.ultraviol...h__1#entry13622

It may have somewhat to do with what version of Color Checker is being used - card or passport? old or new?

Added: The entire sun is shining on my CC Passport so there is no question of uneven light causing this.
Andrea G. Blum
Often found hanging out with flowers & bees.

#5 JMC

    Member

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

Posted 12 January 2018 - 21:33

Hmm interesting Andrea. I'll see if I can dig some pictures of mine out over the weekend.

#6 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 6,216 posts
  • Location: USA

Posted 12 January 2018 - 21:45

I'm happy to reshoot it a patch at a time if it helps.

Given that the CC Passport or Cards are not designed for stable reflectivity under UV or IR, perhaps we should not be surprised. OTOH, it does seem a bit odd that there is some kind of underlying non-color-related chemical difference in the pigment of one gray patch as compared to the others.
Andrea G. Blum
Often found hanging out with flowers & bees.

#7 bobfriedman

    Member

  • Members
  • 200 posts
  • Location: Massachusetts

Posted 12 January 2018 - 23:56

when the weather gets better i will shoot some outside with both the passport and the A-size card. lighting will be more even then

#8 JMC

    Member

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

Posted 13 January 2018 - 10:00

This was a photo I took outside in sunlight (Monochrome converted EOS 5DS R, Asahi 85mm UAT lens, Baader U, 1/30s, ISO 800, f8). It shows on the top line 6 squares from an Xrite colour checker chart, left to right - white to black, in the original order. The second line are some custom UV calibration samples I've been playing around with which vary between 35 and 3% reflection.
Attached Image: Figure 4b 0I8A4578 Baader U 30th ISO800 f8 small.jpg

The Xrite ones were cut down so I could fit them in a UV Vis and measure the reflectances. This is what that showed.
Attached Image: 300 to 400 chart Xr.jpg

While the 2nd, 3rd and 4th grey tiles are very close together, especially below 360nm, overall the tiles would be expected to rank in the same order in UV as they would in Visible light.

My colour checker chart is from 'pre' the change in Nov 2014.

#9 bobfriedman

    Member

  • Members
  • 200 posts
  • Location: Massachusetts

Posted 13 January 2018 - 11:58

nice even illumination which i will redo when possible.. how were the curves calculated Jonathan?

also where do the custom calibration targets come from... or how did you create them (custom?)

Edited by bobfriedman, 13 January 2018 - 11:58.


#10 JMC

    Member

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

Posted 13 January 2018 - 13:16

The curves were done on a UV visible spectrometer I had access to in my last job, Bob. Thankfully I was able to do some fun stuff as well as work on it.

The custom targets I made from a mix of plaster of Paris, carbon black and magnesium oxide. Can't go through it all here, but basically mixed together in different ratios (to give 'greyness') and some water then allowed to set. Obviously not the same as Spectralon, but a tiny fraction of the cost. I'm currently writing it up in a paper and hope to have it submitted in the next few weeks. Once that's out I can share more details.



#11 bobfriedman

    Member

  • Members
  • 200 posts
  • Location: Massachusetts

Posted 13 January 2018 - 13:35

thx.. so your custom targets are not calibrated with the spectrometer like the x-rite chips..

i really didn't expect much from the colorchecker but i have it around and use it for my visible photography to do color profile correction in PS which was the x-rite design criterion.

Edited by bobfriedman, 13 January 2018 - 13:36.


#12 JMC

    Member

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

Posted 13 January 2018 - 15:35

I do have spectral reflectance curves for my targets Bob, but as that is the crux of my paper I can share once it's accepted for publishing. They do not show the same drop between 400 and 360nm as the commercial ones, but they are not as flat as spectralon. Given they cost about $10 to make try Willem adequate for most needs, especially where $5000 isn't available or 8 commercial ones.

#13 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 6,216 posts
  • Location: USA

Posted 13 January 2018 - 16:28

I have a question. In the photo, the brightness (ref HSB model) of the round standards ranges from about 72% on the left to 22% on the right. But you say the reflectivity ranges from 35% to 3%. I've never been exactly sure how we should correlate reflectivity to brightness, but perhaps some tonal adjustments are needed? Because the left-most standard does not appear to have a 35% reflectivity. What's your thinking on this?
Andrea G. Blum
Often found hanging out with flowers & bees.

#14 JMC

    Member

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

Posted 13 January 2018 - 16:46

No effort was made to set the conditions of this picture such that the 'greyness' in the picture exactly correlated with % reflectance values. The picture is to show the difference between the commercial and custom standards. The aim is to include these standards in with the sunscreen films I need to assess so I can see how much UV is being reflected within the image. Make sense?

#15 Andrea B.

    Desert Dancer

  • Owner-Administrator
  • 6,216 posts
  • Location: USA

Posted 13 January 2018 - 19:48

As mentioned, it is difficult to figure out whether or not we should manage brightness when displaying standards in a photo. There are complexities. Does reflectivity correlate linearly with brightness in a digital photo given that the conversion of a raw file includes the application of a midtone curve lift? We have to think about that and be prepared to answer why we have or have not performed any brightness or contrast adjustments because someone reviewing our work may ask.

I think we need to be sure that the standards' tonal steps in a photo are "large enough" to ensure a proper correlation between a subject's reflectivity and the standard's reflectivity. (I'm hoping I said that clearly.) If you drop brightness/contrast too low in such a photo, then because converted digital files don't behave linearly, you might get some tones "jammed together" and not easily separated by sight -- or separated by too few degrees of brightness. I have had conversions where my 25%, 50% and 75% standards are showing brightnesses much closer together than their reflectivity would seem to indicate. In short, I think that it might be best to adjust the brightness/contrast in any photo such as yours above to better indicate the tonalities.

Also I'm not sure that comparing your standards (which BTW I really really like and am impressed by !!) to a Color Checker card is appropriate for UV because nobody uses the unstable Color Checker card as a standard in UV work ?? Granted you will see many photos of these CC cards here on UVP but that's usually because we have the standards scene set up for recording under many different filters -- IR-pass, dual bandpass, UV-pass, Visible, BG, etc. And I'm also concerned that there is something unusual going on in the newer cards which seem to not display consistently the expected brightness ranking in UV.

Well, enough.... :D
Jonathan, you know I'm not writing this to beat on you! So I'm hoping others understand that it's always useful to raise a few questions prior to publication of a paper* just to be prepared for the reviewers. I'm genuinely interested in the use of UV reflectivity standards and the best display of them in a work photo. All opinions always welcomed! We all enjoy talking about such stuff here.



*My own paper writing and tech writing days are over. But I certainly remember the reviews at the Labs and how we always tried to anticipate questions. My papers got sucked up into the giant jaws of "proprietary work product". Oh well. I had fun.
Andrea G. Blum
Often found hanging out with flowers & bees.

#16 JMC

    Member

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

Posted 13 January 2018 - 21:35

Andrea, I do not consider it to be 'beating on' me at all. Science demands question and without question there cannot be change. My work is currently working on samples which reflect between 20% and about 4% of the UV which falls on them, so I needed to make calibration standards which covered that range, and go beyond the extremes of my samples to give me confidence in the measurements. The Spectralon ones which would have given me 20%, 10%, 5% and 2%, I didn't feel gave me enough spread in the region of interest. I approached them to make me some custom ones but cost would have been prohibitive. So I made my own to cover my region of interest. I'm not interested in above 40% - it can all be the same that high up from my perspective when I'm trying to measure amount of reflection from my samples.

I cannot go through the whole paper in a couple of paragraphs but the aim is to show through measurements and photos why the commercial charts are no good for UV work, and demonstrate a simple way people can make UV suitable standards. I'm certainly not the first to use the method of carbon black, plaster of paris and magensium oxide. But the methods for making them which have been reported before do not give sufficient information to easily repeat the work. They also raised questions which were not addressed about the standards themselves and how they reflected light at different wavelengths. So I have presented something more systematic with data to back it up. Should I need to do the same for high reflectance samples one day, I've no doubt I'll approach it in the same manner.

Something you mentioned above about differences between the old and new ones, it would be interesting to check, and yours and Bobs experience would suggest there is one. While I no longer have access to the spectrometer I used previously, even a photo under direct sunlight of the two (old and new) would be enough to see whether there is a difference.

Edited by JMC, 14 January 2018 - 09:20.


#17 Cadmium

    Member

  • Members(+)
  • 1,492 posts

Posted 14 January 2018 - 09:14

Jonathan, Very cool! When you say "flat", do you mean the flatness of the surface? Which could be sanded flat...
Have you tried these in fluorescence? Not that they would really need to work in florescence, but would be curious how they compare with Spectralon, say... or other target materials.
Very interesting and well done! :-)

#18 JMC

    Member

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

Posted 14 January 2018 - 09:24

View PostCadmium, on 14 January 2018 - 09:14, said:

Jonathan, Very cool! When you say "flat", do you mean the flatness of the surface? Which could be sanded flat...
Have you tried these in fluorescence? Not that they would really need to work in florescence, but would be curious how they compare with Spectralon, say... or other target materials.
Very interesting and well done! :-)
Thanks Steve. When I refer to 'flat' above, I'm talking about their reflectance variability across the wavelength range of interest, not their shape.

No, not tried in fluorescence, as it's not currently part of my research plan. Also, with regards to something you mentioned in another thread, I do not have a UV light source which completely blocks visible light.

#19 Cadmium

    Member

  • Members(+)
  • 1,492 posts

Posted 14 January 2018 - 11:37

Also somewhat discussed here, old 2009 posts:
https://www.fotozone...indexing-in-uv/

#20 bobfriedman

    Member

  • Members
  • 200 posts
  • Location: Massachusetts

Posted 14 January 2018 - 12:27

as i originally said.. i don't expect too much from colorchecker cards except some pigment variation.. i only used that card as it was available and i had the other wavebands...

i probably shouldn't have posted it..... will wait until i can spend 1.8K USD on some spectralon targets