Filter Series :: Name-brand UV-Pass Shootout with LaLaU, LuvU-II, KolariU, Moon UVA Miami and More

[Andrea B.]

These are the ones I have. I'm omitting a few older, no longer manufactured UV-Pass filters

 

fwhm = full width at half-maximum

 

FILTERS:

I'm going to list the filters by their raw, non-white-balanced "color" on the color wheel in order of their distance from red which is 0°/360°.

 

Clockwise from red/0°, headed next to orange/30° but none get even half-way there.

• 3° UVIR*Optics LaLaU 363fwhm50, absorptive.
• 4° Kolari Vision KolariU 365fwhm40, absorptive.
• 7° UVR Optics AndreaU Mk II 351fwhm40 or 359fwhm48 or 360fwhm37(confusion!), absorptive.
  Two charts are shown. I'm not sure why both are labeled AndreaU Mk II. I will ask.
  And I'm not sure which one I have.
• 9° UVIR*Optics LuvU II 360fwhm53, absorptive.
• 9° Baader Planetarium BaaderU 350nm peak, labeled HWB=325-369, dichroic.
  The current BaaderU offered by BP is labeled "320-380". It appears to be a 350fwhm60.

Counter-clockwise from red/360°, headed next to cerise/330° then magenta/300°.

• 341° UVIR*Optics Moon Uva Miami 365fwhm55, absorptive.
• 335° UVR Optics StraightEdgeU dichroic/red.
  I am not sure whether this is the most recent version of the SEU or not. I will ask.
  The current StraightEdgeU 353bp61.
• 329° UVR Optics AndreaU absorptive, no longer made.
• 320° UVR Optics SEU Mk I 379fwhm52, dichroic/red, some still available.
• 307° UVR Optics SEU Gen 2 dichroic/silver, no longer made.
  This filter peaked at 395nm, but its (approx) 45 nm wide half-max was between (approx) 345-400 nm and not centered around the peak at all.

......

Camera: Nikon D610 full spectrum conversion

Lens: UV-Nikkor 105/4.5

 

Subjects:

• Spectralon Rectangle - used for exposure determination.
• Color Checker Card - we always seem to include one out of habit.
  We got into the habit of including a CC Card because a quick-and-dirty white balance
  in UV can be made against the black patch left-most in the top row.
• Orange Osteospermum Flower - no test is good without a flower!!
• Labsphere Spectralon 75% Reflective Standard - used for white balance.

Exposure:

• f/11
• ISO-100
  This is base ISO in the D610.

Exposure Times:

Please do not read too much into the exposure times or brightness differences in the raw color photos.

 

When shooting to compare raw colors, I increase exposure time against a white standard until one of the channels (either red or blue) starts to blow out. Then I back up by 1/3 stop. So there are some exposures which could be adjusted by 1/6 to 1/3 stop in order to be optimal. This method is necessary in order to use the white standards for white balance without any highlight recovery (which compresses some of the histogram).

 

As mentioned you will see the minor exposure differences as brightness differences in the raw color photos. When I converted the files, I adjusted the black/white points, thus exposure differences are less noticeable, if at all, in the converted, white-balanced frames.

 

Conversions:

• All raw color conversions in Raw Digger. No additional edits.
• Unless otherwise noted, all white balance conversions in Capture NX2
  with black/white point adjustment and
  USM (5,50,0) for minor local contrast boost.

 

---

 

Photos in the next two topics.

[Andrea B.]

Visible Reference

f/11 for 1/1000" @ ISO-100

Conversion in Photo Ninja

 

 

3° UVIR*Optics LaLaU 363fwhm50

f/11 for 1/6" @ ISO-100

 

 

4° Kolari Vision KolariU 365fwhm40

f/11 for 1/5" @ ISO-100

 

 

7° UVR Optics AndreaU Mk II 351fwhm40 or 359fwhm48 or 360fwhm37(confusion!)

f/11 for 1/3" @ ISO-100

 

 

9° UVIR*Optics LuvU II 360fwhm53

f/11 for 1/5" @ ISO-100

 

 

9° Baader Planetarium BaaderU 350nm peak, labeled HWB=325-369

f/11 for 1/8" @ ISO-100

[rfcurry]

Andrea,

 

You wrote -

• 320° UVR Optics SEU Mk I dichroic/red, no longer made
  

Actually, I have some of that dichroic left over, so I am offering them once again on my website. It is a good filter and does very well in blocking IR, as your tests showed. I moved along from it because the line was not the straightedge I wanted.

 

Just an fyi. :)

[Andrea B.]

341° UVIR*Optics Moon Uva Miami 365fwhm55, absorptive

f/11 for 1/5" @ ISO-100

 

 

335° UVR Optics StraightEdgeU 353bp61, dichroic/red

f/11 for 1/6" @ ISO-100

 

 

329° UVR Optics AndreaU absorptive, no longer made

f/11 for 1/8" @ ISO-100

 

 

320° UVR Optics SEU Mk I 379fwhm52 dichroic/red, some still available

f/11 for 1/5" @ ISO-100

 

 

307° UVR Optics SEU Gen 2 395nm peak, 345-400 hm, dichroic/silver, no longer made

f/11 for 1/15" @ ISO-100

[Andrea B.]

An observation....

The preceding 10 filters have peaks ranging from 350 - 395 nm, yet after white balance we see the same false blues in each photo with some variation in saturation. So this is why I don't think you can claim anything definitive about what false blue represents in a photo made with a wideband UV-pass filter.

[ulf]

An observation....

The preceding 10 filters have peaks ranging from 350 - 395 nm, yet after white balance we see the same false blues in each photo with some variation in saturation. So this is why I don't think you can claim anything definitive about what false blue represents in a photo made with a wideband UV-pass filter.

That is correct except for when the flower have little reflectivity close to 400nm, that is the main source of the blue colour. I do not think the reflectance variations of the ColorChecker, for different wavelengths makes it a good motif for this.

Some white flowers might show blue with filters that allow light close to 400nm, while filters that eliminate that upper band better will give a more purple tone.

[Andrea B.]

I've never gotten a false purple with any of the above filters.

[ulf]

I have, with the Baader U and with the U-360 stack:

https://www.ultravio...__fromsearch__1

[Andrea B.]

White balance: selection of the bottom half of the pictures below the flowers, from side to side.

 

An average white balance like that might not produce the same result as would a white balance against a UV "white" standard. So I don't think false purple would occur if a standard were used.

[Andrea B.]

Example: Soapwort Flower

 

An average white balance on the Soapwort foliage* pushes the flower false color away from blue 240° to around 255° which is halfway to purple at 270°. This foliage is not UV-neutral, so should not be used as a white balance area.

*I used the upper left corner of the photo for the foliage WB. Not all of that area is shown in this crop.

 

 

White balance against a Spectralon 99% standard (not shown here) gives a typical false blue of about 225°. Blue is at 240°. Cyan is at 210°. So this blue is about halfway to cyan and is 30° from the preceding color of 255°.

 

 

 

On a personal note, I would certainly be happy with either of these colors 225° or 255° because they both produce a very nice photo. As do all of Ulf's blue-to-purple Strawberry flower examples referred to above. (See linkie.) But for documentary repeatibility, one would most likely want to use some kind of standard for setting the false white balance.

[ulf]

White balance: selection of the bottom half of the pictures below the flowers, from side to side.

 

An average white balance like that might not produce the same result as would a white balance against a UV "white" standard. So I don't think false purple would occur if a standard were used.

White balance against virgin PTFE still show a difference for me.

Sony A7III and UV-converted Godox AD200 flashes.

 

This flower show less dramatic dark purple as it is rather UV-bright i the spectrum range close to 400nm

WB with a 17 x 17 pixels selection average on the PTFE disc only.

Screenshots directly from FastRawViewer.

 

UG2A + S8612 2mm stack

 

U-360 + S8612 stack

 

Vis with BG40

 

---

Could there be some camera profiling lurking in the background for you that removes the purple tones?

The difference is so big that I think something is fishy here.

[Andrea B.]

---

 

Hmmmm, let me think for a minute or two...... :grin:

 

---

 

Ulf, I remembered just why we often include a Color Checker card in a UV test photo. The black patch (in my photos, first row on the left) can be used as a white balance patch with fairly good results. Not perfect, but gets one started. I'll go back and add that info to the first post.

[Andrea B.]

This "thinking" business is tough work. Just kidding. I woke up sooooo late this morning because I was reading The Martian by Andy Weir and got caught up in it - thus not getting to sleep until like 3 AM or so.

Anyway, back to the white balancing matter at hand.

---

 

My Soapwort white balance examples above with a blue at 225° and 255° were both made in Photo Ninja with an added D610 color profile. So it is not just the D610 profile which makes the difference between the two photos.

The difference occurs because:

• in the 255° example I white balanced against a rectangular selection of the foliage containing no flowers.
  
• in the 225° example I white balanced against a rectangular selection over Spectralon.
  (The spectralon WB was then saved as a preset to apply to all photos in that session.)
  

Let me repeat this --->>> Either of those white balances is OK for all practical purposes. One WB steps towards cyan and the other steps towards purple. But they are only 30° apart on the color wheel.

 

Usually I would not recommend UV white balance against foliage. Some is UV-neutral, but some is not.

 

And for any n00bs reading --->>> There are so many factors affecting white balance in UV. (List them!)** But the results we get are all close enough to provide a standardized WB appearance IF that is desired. Remember WB in UV is not mandatory. We usually make white balances in UV to standardize the outcome for comparative purposes.

---

 

Experiment in UV white balance.

Select a photo which will produce some false blue and some false yellow.

(This might require two separate photos. That's OK.)

 

Using a converter with a good sampling tool (a marquee or a dropper which can be dragged),

make the following series of white balances:

• Draw the marquee to select the entire photo. This gives an Average White Balance.
  
• Marquee select over an area which typically produces false blue. This should give an overall Yellow Cast to the result.
  
• Marquee select over an area which typically produces false yellow. This should give an overall Blue Cast to the result.
  
• Marquee select over an area which is UV-neutral. This should give the typical blue/yellow tones with no color cast to the UV-neutral area. (Ideally your UV-neutral area in the photo is a piece of PTFE.)
  

Look at the results both with and without a camera color profile added. There usually are some minor differences.

Note: With a WB dropper an Average White Balance can be obtained by dragging the dropper diagonally across the photo, for example. I usually make a wavy diagonal drag to include as many different areas as possible.

 

**Here are the usual suspects in the WB game played in any light:

converter, converter white balance algorithm, camera profile or not,

camera, lens, filter, amount of illumination, kind of illumination, time of day....

---

 

Back to Ulf's examples.

[Andrea B.]

As I see it, or as I sample it, Ulf's examples above are both blue. The two blues are about 10° apart.

 

Before any color purists raise the red flag -- (255,0,0) of course -- I do understand that sampling color from converted, compressed Jpgs can be not the ideal. But the sampling effort still tells the tale.

 

Both flowers sample Blue.

I sampled four areas in each photo with a 5x5 sampler and noted the color wheel location.

The saturated square patch shows samples at full saturation and full brightness.

 

Yes, we can detect a smidge more red in the UG2A photo than in the U-360 photo.

But both are still basically a "blue" to me - and to the color wheel. There is no rule about when blue becomes purple, but generally speaking you need to move at least halfway from 240° to 255° before you can say you have blue-purple or purple-blue, however it is called. (Purple at 270°, Magenta at 300°.)

 

BTW, either of these examples would be absolutely acceptable in the botanical section.

((And I don't recognize the flower! Polemonium?))

 

UG2A + S8612 2mm stack

 

 

U-360 + S8612 stack

You could say this photo shows about a 12% move towards purple (128,0,255).

 

---

 

I am *not* a "color expert". If anything I've done/said above is flakey, you know that I want to know about it!!

 

---

[Andrea B.]

Something not discussed with my examples in the first post.

The NON-white-balanced colors from some of these filters can be so beautiful that I often use them.

Look at the raw colors for the MoonU and the StraightEdge/SEUs above. So lovely!

And lots of nice variations off of from them with a few degrees turn of the color wheel or an alternate WB in incandescent or sodium.

[ulf]

As I see it, or as I sample it, Ulf's examples above are both blue. The two blues are about 10° apart.

 

Before any color purists raise the red flag -- (255,0,0) of course -- I do understand that sampling color from converted, compressed Jpgs can be not the ideal. But the sampling effort still tells the tale.

 

Both flowers sample Blue.

I sampled four areas in each photo with a 5x5 sampler and noted the color wheel location.

The saturated square patch shows samples at full saturation and full brightness.

 

Yes, we can detect a smidge more red in the UG2A photo than in the U-360 photo.

But both are still basically a "blue" to me - and to the color wheel. There is no rule about when blue becomes purple, but generally speaking you need to move at least halfway from 240° to 255° before you can say you have blue-purple or purple-blue, however it is called. (Purple at 270°, Magenta at 300°.)

 

The original files are not compressed Jpgs and the difference there looks even bigger.

 

When you look at fully saturated colours the 245° and 235° colour patches indeed look blue, but with less saturation they would definitely appear differently.

If instead of looking at the angles on the colour wheel, you look at the colour on the screen or the RGB composition in the channels, at least if your screen is able to show a reasonably good gamut, 38 in the red channel is much more than a smidge and 4 in the green channel for the other image expands the difference. The R-G difference is 42

Small differences due to WB on foliage, instead of PTFE are much smaller.

 

I wonder if I am the only one, to perceive the U-360 based flower as blue with a rather big purple tint, when looking at the flower, not the synthetic colour patch?

[nfoto]

Puple hue, yes I concur.

[Stefano]

Desaturated blue looks more violet. It is the same thing Andrea noticed here.

[ulf]

No Stefano, that was about using profiling or not when processing the RAW files.

Andrea has posted about how desaturated colours elsewhere.

 

Colour management is very complex and can be confusing.

I cannot say that I have mastered it, but I have been trying a bit.

 

Profiles are correction tools that are applied both in the beginning of the image handling process like the one Andrea used in your linked post and also to correct how colours are displayed on a screen or printout media.

I have used profiling for my monitors for a very long time and am now using my third screen calibration dongle to get proper colours on my screens.

My Dell UP3017 monitor has a 99% Adobe RGB, sRGB, REC 709, and DCI-P3 rating.

I calibrate it from time to tome with a measurement dongle to correct for ay drift.

 

I think I can see correct colours in my images we have been discussing as I think my RAW process flow do not force any profile for the camera /input data.

[Stefano]

Yes, the discussion in the post I linked was about profiles, while here we are discussing filters. But both Andrea and you noticed that fully saturated colors do look different than "normal". In this case, your U-360 image is more violet than the other (and I too see it as more violet than the corresponding color square), and this is kind of the same thing Andrea observed, when she saw desaturated blue as appearing more violet than it really is, but due to the profile being used.

 

In general, when you have a pure blue (almost Andrea's case) and when you have a blue-purple (your case) they both appear more purple when desaturated, because our color vision is very weird.

 

I hope I am understanding properly, please correct me if I am missing a point or two.

[Andrea B.]

There is a name for this color phenomenon.

Let me go find it.

[Andrea B.]

Abney Effect

Link is to Wikipedia. There are lots of other interesting articles about this if you Google around.

 

The Abney Effect is the perceived hue shift when white light is added to a color.

 

When blue is gradually desaturated, it begins to look like lavender/purple.

 

 

 

I'm not sure this next one is also called an Abney Effect.

But there can also be a perceived hue shift when some colors are darkened.

When yellow is gradually darkened, it begins to look like olive green.

 

 

I wanted to see how the blue Abney effect looked in a white frame.

[Stefano]

Interesting. And this is just one of the many "aberrations" our eyes and brains have.

 

There is a difference between darkening and brightening a color, and as you know it is saturation. The dark yellow is still a pure yellow, so it should always look yellow but it doesn't, probably because it is surrounded by brighter yellows and white (it tricks me too, I too see it greener), while the whitish-blue is no longer a pure blue, and so the illusion should work always.

 

Brown is dark orange for example, and you see brown only because the background is brighter. With a black background you would see orange.

[Andrea B.]

I just happen to have an orange to brown strip.

((I've been playing with these color illusions for years now.))

This particular orange begins at (255,128,0), halfway between red and yellow.

 

 

Orange to brown in White Frame.

 

Orange to brown in Black Frame.

I see brown in this also.

 

 

The Abney effect doesn't work for every color. I don't see a hue shift with green, for example.

(Sorry this has some jpg conversion artifacts.)

[Andrea B.]

Here is actual purple which I'm taking to be halfway between blue and magenta (128, 0, 255).

The desaturation steps are 10%.

BTW, there is some variation across the world in what this color is named. That's why I gave you its beginning RGB value.

 

After white balancing our UV false colours we should not have any of this colour.

 

For some people this purple strip will produce an Abney effect. For others it will not.

I see a bit of a hue shift towards pink on the right.