Jump to content
UltravioletPhotography

IR leakage?


montanawildlives

Recommended Posts

montanawildlives

I took this pic a few weeks ago (Fujifilm xt3, full spectrum, Fujifilm 60mm f2.4, Kolarivision UV pass filter). I set white balance in post off the sky (which was blue in real life). I know this is far from optimal and have since purchased a piece of virgin PTFE. (I also found that asphalt worked as one member suggested).

 

I just wanted to get a little feedback on a couple of things. I had kind of decided that the Kolarivision filter had serious IR leak, but when I look at this pic the foliage and grass is quite dark, not bright as I would expect with IR. On the other hand, you can see the lack of contrast in the center of the image, which in fact IS what I would expect with IR leak because this lens is known to have a diffuse hot spot in IR, which I have been told does not really materialize in UV photography (hot spots are not a common or typical problem in UV, so I hear). The hot spot gets worse and more defined as I stopped down the aperture as would happen with IR (leak). So, I guess my question is...if this has IR leak, why are the foliage and grass dark? This was a full sun day and shot with the sun at my back.

 

Thanks!

 

post-356-0-44156700-1622993398.jpg

Link to comment
If the filter leaks some IR, the foliage is still dark, but non as dark as it should be. Also, seeing green in a white balanced UV image (unless you are below like 350 nm) is an alarm bell. I think you have a leak.
Link to comment
Andy Perrin
If this is an image from FILM, film will not behave like a digital Bayer image so I don't know that we can judge by the colors at all here. It's possible there is an IR leak or visible leak, especially given that the Kolari filters seem to have issues in recent times. If you have a leak, the image could still be mostly UV (it looks mostly UV) but with lighter grass than a pure UV image would give you. I would get some S8612 2mm and put it on there if you want to be certain there is no leak.
Link to comment
Bill De Jager
If this is an image from FILM, film will not behave like a digital Bayer image so I don't know that we can judge by the colors at all here.

 

The Fujifilm XT-3 is a digital camera.

Link to comment
I think this boils down to the combination of an unsuited w/b, a lens not passing UV well, and substantial IR leakage. In some way this reminds of the very first version of the Baader U (1/2" size) on an non-modified camera. The overall outcome was nothing like UV at all, yet not IR either.
Link to comment
If you already have IR capacity, why not just put a longpass 720nm filter over the Kolari. If anything comes through, you have an IR leak.
Link to comment
Uv hot spots are not so incredibly rare. The Minolta A 50/1.4 has a particularly egregious one (not that I use that lens for UV much anyway.) The Contax/Zeiss 24 shows an annular hot spot when stopped down, due to back-reflection off the iris!
Link to comment
Green leaks would be very unusual for a Wood's-glass based filter--that is one of the ranges of greatest opacity for such a filter.
Link to comment

I have seen green foliage before in tests with various UV-Pass filters, even the Baader U.

I have not see this for years, but below is an example with a UG11 stack (left) and a Baader U (right). It may have something to do with white balance. And no, this was not using the older Baader U that Birna is talking about above. I can't say what causes this, so I can't blame it on the the filter, or the lens. The lens used here was a Kuribayashi 35mm.

One could try to isolate the green, but it may have something to do with something entirely different.

 

post-87-0-01321600-1623031157.jpg

Link to comment
I have seen a faint green tint on foliage in some of my photos as well. But I assume actual green light probably has little to do with it. The absorbance of chlorophyll is not perfectly uniform across the UVA spectrum; it declines to a minimum somewhere near 325 nm. Perhaps this explains it.
Link to comment

Hard to say. UV-green is a very hard color to see, unless you have a "UV-green" bandpass filter or a "UV-green" LED. The sensitivity of our cameras at 325 nm is much less than at 360-370 nm, so it's unlikely that a weak signal there is visible in UV images.

 

Here (especially in the printed boxes) you can see how things appearing UV-yellow in TriColour UV (absorbing at around 323 nm, peak of Bernard's bandpass filter) are just slightly lavender if not white in the typical UV images. So I doubt any UV-green around 325 nm would be that visible.

Link to comment

The only stacks I have detected any visible light leaks with are those using thin U-340, like 1mm U-340, which does leak more green than the UG11 at the same thin thickness, but the U-340 green leak is still fairly hard to see,

and doesn't look green as much as it just slightly fogs up the blacks. It is hard to demonstrate, but I would not recommend U-340 1mm. Use 1.5mm or 2mm instead and the leak will be gone (assuming stacked with S8612...).

The green shown in photos above 'I don't think' is that kind of green leak, unless you can isolate it and prove that it is.

Link to comment

Based on Jonathan's graphs here (won't share the raw data as he didn't want to), at 0.5 mm U-340 has a 0.2% peak in green light, and U-360 is below 0.1%. This means that at 1 mm thickness they have OD 5+ blocking (U-340) and OD 6+ blocking (U-360) in visible light. Perhaps OD 5 is not enough?

 

Anyway, it is either a leak or a white balance issue, in my opinion. So trying the usual longpass test is the best thing.

Link to comment
It's all about poor lens for UV, filter leakage in IR, and unsuitable w/b. There is no magic in getting "green grass" for assumed UV captures with these limitations in mind. I managed that twenty years ago :smile: thus not surprised my experience can be duplicated.
Link to comment

I fully agree with Birna´s assessment of this problem with this Kolari UV-pass filter.

 

Please let me expand the explanation a bit.

  1. A poor lens for UV pass less UV light, but still pass much IR, making any IR leakage more visible, as IR suppression depends on the UV/IR ratio.
  2. Normally grass and other vegetation with chlorophyll are very UV dark, almost black, while in IR they are bright or white. An IR leakage will make those parts less dark.
  3. A typical IR-leakage by the used glass types for UV-pass filters is located around 700nm.
  4. When doing a white balance for IR-images with different cut on wavelength image features can get very different colours depending on both filter type and used WB object
  5. This marginal Kolari UV-pass filter gives an odd greenish tint of the normally more colour neutral grass due to a combination of all above, but mainly due to the WB.
  6. If WB had been done of the grass instead the headstones and sky would lose the purple-ish tint.
  7. It is more likely the green colour is due to an IR leakage/WB combination than an actual green leakage as normal Ug-type glass has a much bigger peak around 700nm.

Link to comment

Also the pink sky is odd, it should look either white or yellowish (with a better reach it looks more yellow).

 

If the white balance had been done on the grass, the sky would actually look even more pink/purple, as magenta is the complementary color of green.

Link to comment

If the white balance had been done on the grass, the sky would actually look even more pink/purple, as magenta is the complementary color of green.

 

Yes you are right about that. I turned the change the wrong way. Thanks for correcting me.

Link to comment
Andy Perrin
I think the ultimate conclusion is that we need to continue to warn people away from these Kolari filters or make it clear they must be stacked with an additional IR blocker filter. Or suggest they return them to Kolari, except the only person who tried that previously (Nemo Andrea) had to accept a new filter in a different size.
Link to comment

I think the phenomenon is more usual than one might think. Not caused by the filter or the lens. If I can do it with a Kuribayashi 35mm and a Baader U, then you can probably do it with anything.

Of course one can test the situation, and one should do that before assuming anything. A lot of things can be assumed as per those of you who have said, but that doesn't define what is going on.

Best to test and define before assuming.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...