Exp 2: Target-UV with Stock Cam, Unfiltered Lens & Unfiltered UV-LED

[Andrea B.]

^{Update: 14 June 2017}

^{Added links to other experiments}

• 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
  

---

 

Target-UV^TM with Stock Cam, Unfiltered Lens & Unfiltered UV-LED

 

Experiment: Shoot the Target-UV^TM under unfiltered 365nm Nichia UV-Led illumination using an unconverted stock camera and unfiltered lens. [Added-->] No later in-converter white balance adjustments will be made in this first post.

 

Conclusion: The shot made with the Preset white balance (made against the UV-GreyTM target) was the most accurate in colour in this unfiltered test. The other five WB choices showed color casts.

[Added-->] Note that color casts are because no in-converter WB adjustments were made. However, please seePost #7 and Post #10 for alternately processed versions of this composite chart.

 

Equipment:

 

Camera: Nikon D810, stock

Lens: Micro-Nikkor 60/2.8G

Lens Filter: NONE

Lighting: 365nm Nichia UV-Led Flashlight

Lighting Filter: NONE

 

Settings:

 

Neutral [0] Picture Control, no sharpening

Nikon ADL = off

f/2.8 @ ISO-400, various speeds

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.}

• In-camera Preset White Balance
  made against the UV-Grey^TM Target under 365 UV-Led light in darkness.
  
• 10000K[0,0] -- the hottest K.
  
• Direct Sunlight[0,0] -- that is, "daylight" or "sunny" white balance.
  
• Incandescent[0,0] -- aka "tungsten".
  
• 2500[0,0] -- the coolest K.
  
• Auto1[0,0] -- the "normal" auto wb.
  

Description:

 

The experiment was conducted in darkness in my hallway coat closet. The unfiltered UV-Led flashlight was aimed onto the Target-UV from the top of the camera and maintained at the same distance from the Target-UV for all shots. LiveView was used to set an exposure time. Photos were made of the Target-UV under 6 different white balance settings.

 

Photo Preparation:

 

The raw files were converted to TIFs in Photo Mechanic.

In Capture NX2, minor LCH curve edits were applied on the Luminosity layer to adjust the same neutral patches to approximately the same brightness on each strip. This was needed because of slight differences in exposure and minor variations in flashlight illumination. (I was hand-holding the torch.)

The composite was constructed in PSE 11. WB labels were added in PSE 11.

The labeled composite was resized to a high-quality JPG in Photo Mechanic.

 

Observations:

 

Preset White Balance - the Winner: The white balance made against the UV-Grey target under the 365nm UV-Led was the clear winner. Its neutral patches are the most neutral of the series.

 

The worst WB setting: Auto white balance got confused. No wonder when it was trying to see Visible emissions through UV light.

 

Exposure Times: I didn't agonize over the "proper" exposure this time. I just used the LiveView matrix-metered exposure.

 

The Usual Caveat: Converting the raw file and stuffing it into a resized JPG/sRGB box to be posted on a website may alter the colors. This is a known factor.

 

Wobbly Torchlight: Tonight the Nichia 355nm UV-Led Flashlight was bright when I first turned it on but then faded a bit. Is that normal? Well, it certainly is annoying. Therefore I was careful to turn the flashlight off after each shot so that I could shoot the next shot when the flashlight was at its greatest brightness just after turning it on. I wonder if I was entirely successful in this effort? All the clicking and clacking was detrimental to the uniform application of torchlight to the same area of the subject each time. Geez, it's always something.

 

Next Up: Repeat experiment with filtration on the flashlight to cut visible leaks, if any.

 

See Also: Post #7 and Post #10 for alternately processed versions of this composite chart.

 

D810 (stock), unfiltered Micro-Nikkor 60/2.8G (stock), unfiltered 365nm UV-Led flashlight

[JCDowdy]

It may just be my eyes or monitor but Sunny seems to have a better looking red.

[Andrea B.]

The Sunny Red does look good.

Some of the "Reds" have an orange cast (too much yellow).

And some have a magenta cast (too much blue).

 

Important to note: I did not apply any colour profiling nor any white balance steps to these images.

 

So let's see what happens when I filter the UV-Led. Will those Reds change?

And, what would happen if I white balanced these images?

I think I'll work that up also.

 

The Target-UV has an interesting learning curve. How best to use it to obtain accurate fluor colour is one of those things I hope to discover with these little experiments. So far we can see that making an in-camera preset white balance is definitely a good thing. :)

[Cadmium]

I received some of this, the white version.

eBay item: 111080373214

It is very white when illuminated with my MTE. Bright also.

Would it be OK if I painted a little on something and sent it to you for a test?

[Andrea B.]

Sure, send it on.

[Andrea B.]

Update: 03 June 2017

Added info about proper metering for WB accuracy.

***********************

 

I took the all the UV-Target raws into Photo Ninja -- except for the Preset strip -- and white-balanced each one on the bottom left fluorescent white patch. Then I reconstructed the composite. The white balance step considerably improves the RGB patches but does not entirely remove the color cast from the neutral patches.

 

Added 03 June 2017:

Actually it does not make a difference whether you set the WB in-camera or in-converter as long as you photograph the white balance area carefully to prevent blowout for in-converter WB use.

In my example here, the brightest white patch (2nd row, left) was not completely accurate for setting an in-converter white balance because that brightest white patch was not spot metered for later white balance accuracy. Without spot metering on the brightest areas, those areas can partially or fully blow out or nearly blow out and render later white balance inaccurate in varying degrees. I was using matrix metering when photographing the UV-Target and trying to hit a "happy medium" on the lights/darks exposure. Please see Post #10 for a more accurate white balance using the UV-Grey target.

 

D810 (stock), unfiltered Micro-Nikkor 60/2.8G (stock), unfiltered 365nm UV-Led flashlight

[Cadmium]

Why are you not filtering the light?

[Andrea B.]

I'm experimenting.

We have to see what happens if we don't filter the UV illumination in order to know why we should filter the UV illumination. :)

As noted above, my next experiment will filter the torch.

[Andrea B.]

Update: 03 June 2017

Added info about proper metering for WB accuracy.

***********************

 

Here is a repeat of the white balanced composite, but his time the same white balance obtained from the UV-Grey card was used for each raw file. Then the composite was constructed as usual.

 

This white balance result is more like what I expected in the preceding post but did not quite get. The white balance is improved on the neutral patches although there are still a few color casts. [Added -->] Not surprising when incorrect WB settings were used and brightest areas were not spot metered -- sometimes you cannot recover all WB errors. But I will investigate further, later.

 

Added 03 June 2017:

In my example here, the UV-Grey target was more accurate in producing a fluorescent white balance than was the brightest white patch (2nd row, left) possibly because the brightest white patch was not spot metered for later white balance accuracy? Without spot metering on the brightest areas, those areas can partially or fully blow out or nearly blow out and render later white balance inaccurate in varying degrees. I was using matrix metering when photographing the UV-Target and trying to hit a "happy medium" on the lights/darks exposure.

 

[Cadmium]

Yes, sounds good (about not filtering light first).

Those last two are a little hard to tell apart. The whites seems pretty much the same, and the red/green/blue look the same, just maybe the grays are a little less blue in this last one, is that what you are seeing also?

[Andrea B.]

I had to think about this for a while. But then it turns to be fairly obvious!!

 

In my example here, the UV-Grey target was more accurate in producing a fluorescent white balance than was the brightest white patch (2nd row, left) because the brightest white patch was not spot metered for later white balance accuracy. Without spot metering on the brightest areas, those areas can partially or fully blow out or nearly blow out and render later white balance inaccurate. I was using matrix metering when photographing the UV-Target and trying to hit a "happy medium" on the lights/darks exposure.

(I'm also going to add this note to the posts above.)

 

It's early days. I've got more learning to do. :D

[Andrea B.]

Some updates to Post #7 and Post #10.

 

I hope I'm not making a bosh of this.

[aphalo]

Looks like a very useful tool! :)

What a pity that the target is so expensive! :( (Maybe at some point I can have it bought at work...)