Exp 4: UVIVF White Balance with the UV-Grey Target, Stock Cam/Lens

[Andrea B.]

• Exp 1: First Look at Target-UV & UV-Grey for UVIVF [see Post 7]
  
• Exp 2: Target-UV with Stock Cam, Unfiltered Lens & Unfiltered UV-LED
  
• Exp 3: Target-UV with Stock Cam, Filtered Lens & Filtered UV-LED
  
• Exp 4: UVIVF White Balance with the UV-Grey Target, Stock Cam/Lens
  

---

 

UVIVF White Balance with the UV-Grey^TM Target

 

Experiment: With a stock camera and filtered stock lens, make a white balance preset against the UV-Grey^TM target under filtered 365nm UV-LED illumination. Shoot a Gazania flower. Using various other white balance settings, re-shoot the floral subject. Try to find a K setting which matches the preset white balance.

 

Comment:

• We have already seen in previous experiments that Daylight white balance is not accurate for UVIVF. Daylight WB might have worked for film, but digital is different.
  .
  
• The D810 has a strong internal UV/IR blocking filter. Previous results seem to indicate that it is probably not necessary to use a UV/IR block on the lens used with the D810. But I'm going to stress Good Practices in UVIVF photography and use such a filter anyway.
  

.

Equipment:

 

Camera: Nikon D810, stock

Lens: Micro-Nikkor 60/2.8G

Lens Filter: Baader UV/IR-Cut

Lighting: Nichia 365nm UV-Led Flashlight

Lighting Filter: BaaderU

 

Settings:

 

Neutral [0] Picture Control, no sharpening.

Nikon ADL = off

f/4 @ ISO-400, various speeds

Metering = Matrix

White Balance Series:

_{In Nikon cameras, a white balance may be fine-tuned on a 2-dimensional blue-red/green-magenta grid.}

^{The bracketed [0,0] indicates no fine-tuning was applied.}

• Preset UV-Grey -- against the fluorescent grey/white target.
  
• K2500 [0,0] -- the 'coolest' K.
  
• K10000K [0,0] -- the 'hottest' K.
  
• K6250 [0,0] -- a bit above Daylight, halfway between.
  
• Cool White [0,0] -- fluorescent tube emulation.
  
• Auto1 [0,0]
  
• Auto2 [0,0] -- but keeps warm tones on a Nikon.
  

.

Description:

 

The experiment was conducted in darkness in my hallway coat closet. The filtered UV-Led flashlight was aimed above and to the left of the flower with an attempt to evenly illuminate the fluorescent flower center. LiveView with Matrix Metering was used to set exposure times.

 

Results:

 

As you will see in the photos, the D810's K10000 white balance setting came the closest to the Preset UV-Grey white balance, but still produces a blue color cast. When I drag the Photo Ninja white balance sampler over the UV-Grey target, I get an even higher temperature reading of K15000. At this point I am not optimistic that any of the manufacturer supplied white balance settings on our cameras will produce a color accurate UVIVF photo under 365nm UV-Led illumination.

 

Observations:

 

The photos are shown in their unconverted, unedited SOOC Jpeg format with no sharpening. The Nikon Neutral picture control is indeed very neutral. If I were presenting these photos elsewhere I would use a slightly higher contrast, saturation and sharpening. One example of this is shown so that everyone doesn't get the idea that UVIVF photos are flat messes. :D

I managed to get a bit of finger shadow in the upper left of some photos.

 

The Standard

An in-camera white balance preset was made against the UV-Grey target.

The first photo is the SOOC Jpg made in Neutral [0] with no sharpening.

(Aperture f/4 does not give much depth-of-field, does it?)

 

The second photo has more typical contrast, saturation and sharpening added.

This is the approximate fluorescence level I saw from the Gazania.

 

Almost but Not Quite

The K10000 white balance setting still shows a blue color cast.

But, as you will see, it's not as blue as other white balance settings.

 

K6250

On my Nikon D810, this is the K setting half-way between the endpoints of K2500 and K10000.

Bluer than K10000.

 

K2500

Very, very blue.

 

Cool White

I had thought there might be a chance with this one, but no. Very blue-violet.

 

Auto1

No surprises here, it's blue.

 

Auto2

On a Nikon DSLR, the Auto2 setting preserves some warm tones. Still blue.

 

Visible

Here is the Gazania in its normal Visible appearance.

[Andrea B.]

Any comments or suggestions?

 

It was very disappointing that I was unsuccessful at finding a good UVIVF white balance setting using Nikon supplied settings like Daylight, Sodium Vapour, Fluorescent or Kelvin. Only with the white balance preset tool could I neutralize the UV-Grey target.

 

I'll keep working at this at a later time.

[Guest]

Andrea, just a question: does the preset measured from the UV-grey target make much of a difference in the exposure metering for an image? I'm curious as to how much the may affect/help to get proper exposures.

[Andrea B.]

I don't have any observations about that from the first experiments. I'll check it out in later trials. Too busy trying to get the hang of the thing!

 

I have seen that it is easier to get a good reflected UV exposure if there is a preset white balance in use together with a Neutral color control. Even if the WB in use is not quite accurate (such as with my ever pesky Nikon conversions), any WB setting which reduces the red overload in a Bayer filtered camera always helps a little to get a better reflected

 

With UVIVF we typically do not have quite the heavy channel overload under an Auto or Daylight white balance used in conjunction with a Neutral color control, so I don't expect there to be a big change in metering when using an accurate WB for UVIVF. But we will see. It would be the Blue channel that might cause trouble? So keep blue blowout, if any, under control. And of course if your subject has some ultra-strong visible fluorescence, then that would need to be taken into consideration. [A while back I had this bright, bright green fluorescent liquid for which I simply could not get an accurate photo.]

[Oliver Krohn]

Andrea, I am currently thinking about white balance for reflected UV and UVIVF, again. A UV grey standard (sintered PTFE, Spectralon) should be perfect for white balancing reflected UV, but for UVIVF? Maybe I did not get the point, but with UV lighting and UV grey standard, how can there be any visible light you can use for white balancing, since the standard should only reflect the light? And I wonder how important WB for UVIVF really is. Of course, it is important to use the same WB color temperature to retain the same fluorescence hues in various shots. But beside this?

 

And in reflected UV: the colours are pseudo colours and I am pretty sure the colours vary from camera to camera, depending on the Bayer matrix, sensor etc. So, I personally treat UV images like a b/w image. And I think setting the black and white point while post-processing is more important than the white balance, as long as I do not want the pseudo colours from the Bayer filtered sensor.

 

Back to WB UVIVF. Wouldn't a standard colour temperature set for WB be sufficient? Let's say we would agree to 7000 K, wouldn't this be fine? But maybe I am wrong...

 

Oliver

[Andrea B.]

The UV-Grey^TM target mentioned in this topic is a fluorescent target designed to emit grey-white visible light under UV stimulation. For reflected UV work we use instead the Spectralon or PTFE targets. Spectralon does not fluoresce and so cannot be used for UVIVF work. (We have had an ongoing discussion about whether PTFE fluoresces, but that is a discussion for another place.)

 

I can see that it is not immediately obvious in the initial post that the UV-Grey^TM target is a fluorescing target unless one is already familiar with the name. I'm going to add a remark in the first post to remind readers not familiar with it that the UV-Grey^TM is used only for fluorescent work.

 

Here's a link to the manufacturer: http://www.uvinnovations.com/target-uv

Lots of good info there.

 

 

Wouldn't a standard colour temperature set for WB be sufficient? Let's say we would agree to 7000 K, wouldn't this be fine?

 

Yes, setting a baseline color temp would be another way to standardize the results in UVIVF photos. However, when I experimented with various tempature settings, I could never quite get them to match the color of the fluorescence I was seeing (that is, "seeing" before eyes adapt to the dark). Tint plays a role in this along with color space. (Not all colours can be shown/captured in a camera gamut.) The UV-Grey target was able to provide a much better match of the visible colours.

 

 

 

And in reflected UV: the colours are pseudo colours and I am pretty sure the colours vary from camera to camera, depending on the Bayer matrix, sensor etc

 

Surprisingly perhaps, we have found that white-balanced false colours in reflected UV photos are basically the same across the various Bayer cameras. Some minor variation does exist because false colour is dependent on the converter and WB algorithm used amongst other variables. But everyone gets the same blue/yellow/grey/white/black false color palette after the WB step. Sometimes a little blue-green shows up. I don't have time in this brief reply to discuss the nuances of white balancing except to say that some converters are not able to take the false colour temperature low enough to eliminate the pink/magenta tones.

 

In false colour discussions I try to always add that white balancing a false colour, reflected UV photo is the photographer's choice. It certainly is not mandatory in any way. On UVP, only for our botanical postings we do request that WB be applied in order to show the floral UV-signatures in a standardized manner across the various gear platforms.

 

 

 

And I think setting the black and white point while post-processing is more important than the white balance, as long as I do not want the pseudo colours from the Bayer filtered sensor.

 

A reflected UV photograph cannot accurately represent the actual tonality of the photographed subject. For one thing, our cameras lack the dynamic range to be able to do this. And during the conversion step the raw data has contrast enhancements applied so that the raw data becomes human-visible.

 

However, I think that it is OK to set the BW points in any photograph having a wide enough dynamic range for this to be meaningful. (If tone range is narrow, then setting BW points can posterize the photo into a real mess.) In experiments with Spectralon standards -- under the assumption that digital brightness correlates in a reasonable way with the amount of UV reflectivity -- I've found that my 99/75/50/25/2 targets cannot be properly represented. The midtones get smushed (non-technical term there!!) because of the conversion contrast enhancements (usually called "gamma" and "autoscaling"). Our cameras are getting better though.

 

 

We have many topics on the subject of white balance. Use the white balance search tag to find more. :D

 

Thanks for your good questions. I hope I was able to clarify things a bit.

[Oliver Krohn]

The UV-Grey^TM target mentioned in this topic is a fluorescent target designed to emit grey-white visible light under UV stimulation. For reflected UV work we use instead the Spectralon or PTFE targets. Spectralon does not fluoresce and so cannot be used for UVIVF work. (We have had an ongoing discussion about whether PTFE fluoresces, but that is a discussion for another place.)

 

I can see that it is not immediately obvious in the initial post that the UV-Grey^TM target is a fluorescing target unless one is already familiar with the name. I'm going to add a remark in the first post to remind readers not familiar with it that the UV-Grey^TM is used only for fluorescent work.

 

Here's a link to the manufacturer: http://www.uvinnovations.com/target-uv

Lots of good info there.

 

Thanks for your detailed answer! You are completely right, I was fooled by the product name "UV-Gray". I should have looked it up when I noticed the TM, but however, I did not, ash on my main. :D

 

However, I think that it is OK to set the BW points in any photograph having a wide enough dynamic range for this to be meaningful. (If tone range is narrow, then setting BW points can posterize the photo into a real mess.) In experiments with Spectralon standards -- under the assumption that digital brightness correlates in a reasonable way with the amount of UV reflectivity -- I've found that my 99/75/50/25/2 targets cannot be properly represented. The midtones get smushed (non-technical term there!!) because of the conversion contrast enhancements (usually called "gamma" and "autoscaling"). Our cameras are getting better though.

That is true. My first thought was bracketing and tone mapping for HDR, but on the other hand this is not necessary for me.

 

We have many topics on the subject of white balance. Use the white balance search tag to find more. :D

To be honest, that is exactly the way I found this post. :)

 

Thanks for your good questions. I hope I was able to clarify things a bit.

Thank you, too! No questions left at this time! :)

 

Just found the sticky regarding this topic: http://www.ultravioletphotography.com/content/index.php/topic/1456-sticky-uv-induced-visible-fluorescence/

Again, ashes on my main.

[OlDoinyo]

I am not surprised that none of the canned presets worked very well--after all, they were not intended for this use case. Furthermore, the "whiteness" of your standard phosphor target may vary depending on the exact profile of the exciting light, if it uses more than one fluorescing substance (which it probably does.) I take it that you were exciting with a 365-nm LED. I suspect that daylight might produce slightly different results.

[JCDowdy]

This detail is likely in one of your previous experiments in this series, but what is the background and why is it visible?

[Andrea B.]

That is the UV-Grey fluorescent target, visible because it is fluorescing. B)

I don't recall the precise placement of the target/flower combo.

[JCDowdy]

I thought it must be, thanks for clarifying.