[Filter Test] Blocking UV for Visible, Full-Spectrum Photos: Part #2 UV Stress Test

[Andrea B.]

16 Dec 2021:

Part #1 is here: LINKIE 

Part #3 SUMMARY is here:  LINKIE

 

 

Okie dokes, The Zeiss T* UV-blocker has arrived. And the new Kolari Hot Mirror Pro 2, which is a UV/IR-blocker, has arrived. Testing will begin soon.

The Kolari HM Pro 2 is beautifully packaged in a large, square, plastic filter box with a magnetic closure. The filter is seated into a protective black foam so it will not shift around inside the storage box. The foam has fingertip cutouts so you can lift out the filter easily. Included is a separately packaged black microfiber filter cleaning cloth marked with "Kolari Vision" and their K logo. All in all this was a very cool & spiffy presentation.

 

ADDED 20 Dec 2021:

The Kolari Hot Mirror Pro 2 has a blue-green absorption substrate

which is interference-coated with a UV/IR-blocking coating.

The Kolari HM Pro 2 appears to be very well made. I like the knurled filter ring which makes it easier to handle when attaching to the lens. The ring is not too tall, about 3 mm, thus less to cause vignetting.  The ring seems thick enough not to be easily ding-able. There is a blue-green cast to the glass. The coated surface is very smooth.

The Zeiss T* was mentioned briefly in the first test. To review, the Zeiss T* is also very nicely made. It is packaged in a clear, round, metal-hinged, plastic filter box having a press-click opener. There is a layer of white foam under the filter. The filter is clear. It is difficult to detect a color cast unless held "just so" to see some blue-green. The coated surface has a slight "drag" to it, whatever that may mean.

Side Note:  I paid for both filters.


First Test: LINKIE
In the first test I don't think my subject was well chosen to give meaningful results.
So I'll try again. And this time I have two new filters.

Test Goal
When used on a UV/IR-capable lens on a full-spectrum camera,
do the listed filters produce differences in the photos?
If there are differences, can we attribute them to UV-blocking?

Filters, in 𝛂𝛃-ical order

Note:  I'm going to block IR because I don't want IR interfering with results.

Baader UV/IR-Cut
This Baader represents the square-shouldered type of UV/IR-Cut filter.
These are known to pass too much red, which is difficult to repair even with color correction profiles.
 

Kolari Hot Mirror Pro 2 (UV/IR-Cut)
This is the newest UV/IR-Cut filter which looks *very* promising because its transmission curve
matches the typical transmission curve in unconverted DSLR or mirrorless cameras.

Any minor color variations from the original color can be easily tweaked using a color correction profile.
 
Schott BG38 x 2.0 mm (IR-Cut)
The BG38 passes a lot of UV between 310-400 nm (and also Visible between 400 - 705 nm).
Used alone, this filter acts as a control for the other UV-blocking filters.
 
Schott GG395 (Longpass) + Schott BG38 x 2.0 mm (IR-Cut)
This filter combo, which passes some UV near 400 nm, is also added as a control.
 
Schott GG400 (Longpass) + Schott BG38 x 2.0 mm (IR-Cut)
This filter combo includes the all the visible violet.
 
Schott GG420 (Longpass) + Schott BG38 x 2.0 mm (IR-Cut)
This filter combo cuts some of the visible violet
and matches the cut-in of most internal filters in unconverted cameras.
 
Zeiss T* (UV-Cut) + Schott BG38 x 2.0 mm (IR-Cut)
This filter combo is similar to the GG400 + BG38 stack.
The Zeiss T* has a very steep left shoulder.

 

SUBJECT
OK, I could use some help here. What should I photograph? 
I was thinking of perhaps a Color Checker Card and the Labsphere 99% with a UV-torch shining on them because I want to ensure a supply of UV to determine whether the UV-blockers really do block UV.
On the other hand, that is not a particularly "natural" scene. So maybe I should just photograph some outdoor, sunlit, medium or short distance scene?

CAMERA:  Nikon D610 Full-spectrum Conversion
I'm using the D610 in this test so that I can create color correction profiles in Photo Ninja.
PN does not yet support my Panasonic S1R.

LENS:  Probably the Coastal 60/4.0 for maximum UV capability.
If there is some UV getting through a filter, then it really needs to also get through the lens for proper testing of these filters.

 

 

[Doug A]

@Andrea B. looking forward to this test. Not sure what weather your having? Could you shoot a flower arrangement and have the color checker in the corner of the picture? If the weather is nice, perhaps place the flowers outside on a table, with mountains in background.

 

Does the Coastal pass more UV than the Nikon Rayfact?

 

Thanks,

Doug A

[Andy Perrin]

Andrea, I would like some day photos but also night photos with bright lights in them. Maybe one each. These could be separate tests if that’s too much work. 

[dabateman]

Its probably hard to photograph now were you live. But one of the best test subjects I found for testing out BW486, Bg38, bg40 were a collection of various degree of yellow, orange and red maple leaves. 

Something with a good range there would be ideal. 

[colinbm]

The Kolari Hot Mirror Pro 2 gives a good restoration of visible colours in Auto WB on my converted Sigma fp

[Andrea B.]

STRESS TEST

In which I'm giving these UV-blockers something to block!

 

Test Goal

Do the listed UV-blocking filters block UV?

 

Camera:  Nikon D610, full-spectrum conversion

Lens:       Coastal Optics 60/4.0

Subject:  Color Checker Card, Spectralon 99% & 75% Standards

 

Procedure:

For each filter an in-camera white balance was made against 5x5" square Spectralon.

Then CC card and standards were photographed twice:

1) In a mix of ambient daylight and tungsten overhead lights.

2) Lighting in A with added Convoy 365nm UV-LED held about 10 inches from the card.

All photos made at f/5.6 with ISO-400.

 

Filters, in 𝛂𝛃-ical order

I'm going to block IR because I don't want IR interfering with results.

This info is repeated from my previous post.

 

Baader UV/IR-Cut
This Baader represents the square-shouldered type of UV/IR-Cut filter.
These are known to pass too much red, which is sometimes difficult to repair even with color correction profiles.
 

Kolari Hot Mirror Pro 2 (UV/IR-Cut)
This is the newest UV/IR-Cut filter which looks *very* promising because its transmission curve
matches the typical transmission curve in unconverted DSLR or mirrorless cameras.

Any minor color variations from the original color can be easily tweaked using a color correction profile.

 

Schott BG38 x 2.0 mm (IR-Cut)
The BG38 passes a lot of UV between 310-400 nm (and also Visible between 400 - 705 nm).
Used alone, this filter acts as a control for the other UV-blocking filters.
 

Schott GG395 (Longpass) + Schott BG38 x 2.0 mm (IR-Cut)
This filter combo, which passes some UV near 400 nm, is also added as a control.
 

Schott GG400 (Longpass) + Schott BG38 x 2.0 mm (IR-Cut)
This filter combo includes the all the visible violet.
 

Schott GG420 (Longpass) + Schott BG38 x 2.0 mm (IR-Cut)
This filter combo cuts some of the visible violet
and matches the cut-in of most internal filters in unconverted cameras.
 

Zeiss T* (UV-Cut) + Schott BG38 x 2.0 mm (IR-Cut)
This filter combo is similar to the GG400 + BG38 stack.
The Zeiss T* has a very steep left shoulder.

 

PHOTOS

These photos are straight-out-of-camera.

The embedded JPG was extracted in Photo Ninja.

Labels and resizing were made in Photo Ninja.

 

The order of presentation does not match the filter list order.

 

 

 

 

 

 

 

 

[Andrea B.]

dang-rabbit. The forum software won't let me put them side-by-side.

[Andy Perrin]

No guesses except that number 4 is probably the Schott GG395 (Longpass) + Schott BG38 x 2.0 mm (IR-Cut). They all look like they did a decent job except that one.

[dabateman]

Andy 4 would have to be BG38 only. As its the only one that's not like the others. 

It also shows that this isn't the best test subject. You need something with more variation on red and orange to get a better idea.

[Andy Perrin]

David, you are right, I actually missed that she put BG38 alone as one of them. I picked what I thought was the one with the weakest UV cut but missed the “none” option. 

[Andrea B.]

I think that the experiment shows that the UV-blockers all did a good job!

 

1. GG400 + BG38

2. Baader UV/IR-Cut

3. Zeiss T* + BG38

4. BG38 alone

5. GG395 + BG38

6. Kolari Hot Mirror Pro 2

7. GG420 + BG38

 

There are some nuanced color differences here and there, but that is somewhat difficult to analyze meaningfully. Remember, I used an in-camera white balance for each filter & filter stack. But as everyone probably knows by now, I have always thought that the best white balance is obtained by creating a color correction profile.  I'm pretty sure that if I created a cc profile for each filter/stack, the resulting colors would be very, very close.

 

ADDED LATER:  Good color profiling presupposes that there is even lighting which is not different colors in different directions. I just learned this the hard way!.

 

Side note:  I make these photographs with neutral camera settings, so the JPGs lack appropriate contrast. For example, the black areas should be darker. But that isn't what the experiment was about.

 

I'm going to try to rearrange the results so we can scan the "with UV" results separately from the "no UV" examples.

 

Another comment:  Using a 365 nm UV-Led probably does not fully stress test #5, the GG395 + BG38 combo, which has the possibility of a small UV passage between 395-400 nm. 

 

Purple Patch Comment:  The purple patch in row3, column3 has some color differences. For example, compare Purple Patch in #2 with Purple Patch in #1. The Baader UV/IR-Cut (#2) is passing a bit too much red, as always. 

[Andrea B.]

I made color correction profiles, applied them and adjusted black & white points in Photo Ninja.

Labels and resizing were made in Photo Mechanic.

 

Here are the photos made without the 365 nm UV-LED.

 

PROBLEM: 

After resizing and saving as JPGs, the photos seem to have lost their white balance.

What the heck ??

 

See my post below for the explanation. It was my fault!!

 

I am leaving only one photo posted here to show that even after color profiling, my uneven light could not be properly profiled leaving a cyanish cast in the neutral patches.

 

 

 

[Andrea B.]

I have no idea what is going on here.

In the initial set of unprofiled photos there is not a cyan cast as there is in the profiled set of photos.

I'm at a loss to figure out what happened.

 

I'll walk back thru the processing and try to figure it out.

 

Added Later:  Got it. See next post.

[Andrea B.]

OK, I think I was photographing in visible light which was too weird. There were overhead tungstens mostly on the left and ambient daylight on the right. You can see the color casts in this uncropped photo.  It is unreasonable to expect that any color profiling could repair this. If I take WB from the Spectralon on the left, then the white patch on the CC Card and the right-hand Spectralon are not in balance. And if I take WB from the Spectralon on the right, then the white patch on the CC cart and the left-hand Spectralon are not in balance.

 

I screwed up the lighting !!

 

But I do think we can still conclude that the filters quite successfully cut the UV.

 

I will certainly try to run this test again with proper lighting.

[Andrea B.]

Oh well.

Please learn from my errors! That's why I leave them posted. La!

 

[Doug A]

@Andrea B. this reinforces something I ran into long ago. Testing is hard, complicated work! Doesn't matter what is being tested.

 

Thanks for sharing,

Doug A

[Cadmium]

I would like to compare BG38 (alone) vs BG38 + GG420 stacked. I assume those are both there above, but I am not going do the guessing game.

And maybe also compared to BG38 + KV 418 (instead of GG420) and BG38 + Zeiss T* (instead of GG420).
Those are the 4 I would like to compare. I don'[t think you have all of those, so in that case just the BG38 vs stacked with GG420, which I think you have there.

So... wake me up when things are determined, then I would like to see.

 

I like the last photo above, BG38 + GG420, of the Color Checker and the Spectralon. 

I agree with Doug, testing can be laborious, hard and complicated, care must be taken, and often things get mixed up and confused.

[Andrea B.]

.....but I am not going do the guessing game.

 

It is very important to judge test results without a preconceived bias. 

That is a bit difficult for the tester (me in this case) to do, but stripping the file names helps.

 

And, I listed the filters in the test results here: https://www.ultravioletphotography.com/content/index.php?/topic/5088-filter-test-blocking-uv-for-visible-full-spectrum-photos-part-2/&do=findComment&comment=52732

You must not have seen that?

 

BG alone is 4A, 4B.

GG420 + BG38 is 7A, 7B.

Zeiss T* + BG38 is 3A, 3B.

 

******

 

To be clear, the UV-blocking test went well. All the filters, except for (of course) the unstacked BG, did cut UV quite well. For 1,2,3,5,6 and 7, the A photo looks like the B photo.

 

It was the attempt to create color profiles which did not go well. Even though it was staring me in the face, I did not not see that my mixed lighting was weighted towards red on one side and blue on the other. Sometimes we just do not see what we are looking at !! 

After conversion, if I white-balanced using the right-hand Spectralon, then the left-hand Spectralon had a slight color cast. If I white-balanced the left-hand Spectralon, then the right-hand Spectralon had a slight color cast. La!   

 

[Andy Perrin]

Mixed UP lighting!