Jump to content
UltravioletPhotography

Color shifts and other issues


Stefano

Recommended Posts

Visual observation of a point light can be done even when a filter has very high OD, provided the light source is intense enough. The eye can easily span many magnitudes of light intensity.
Link to comment

Let's jump to the point here.

Put that red plastic on your UV+BG stack and see what it shows when using the same exposure.

 

It would be best to use a converted camera, with a changeable lens, and a verified UV-only filter., and a glass longpass filter as well (if the IR test is even still needed),

but since this is what you have, use the red plastic, and loose the LED's for these tests. Shoot the dandelion in sunlight, UV+BG stack vs UV+BG+Red-filter stack, and show the results.

Link to comment
I will do it when I will have a strong sun, without clouds. When I said that the filter was transparent in UV, I meant that paper was fluorescing strongly with the filter in front of the torch. It only attenuated UV a bit.
Link to comment

Yesterday I had a clear sun and I did some tests.

All images f/2.8, ISO 800, 1/8 s exposure.

 

ZWB2 (2 mm) + chinese BG39 (2 mm).

post-284-0-30681700-1582118797.jpg

 

ZWB2 (2 mm) + chinese BG39 (2 mm) + pen ink filter.

post-284-0-77276800-1582118906.jpg

 

ZWB2 (2 mm) + chinese BG39 (2 mm) + red plastic filter.

post-284-0-54693400-1582118928.jpg

 

 

ZWB2 (2 mm) + chinese BG39 (2 mm).

post-284-0-95471000-1582118980.jpg

 

ZWB2 (2 mm) + chinese BG39 (2 mm) + pen ink filter.

post-284-0-57818000-1582119010.jpg

 

ZWB2 (2 mm) + chinese BG39 (2 mm) + red plastic filter.

post-284-0-17465300-1582118998.jpg

 

As I said, that red plastic leaks UV.

 

Here are two other dandelion images:

 

F-stop: f/2.8, ISO 800, 1/8 s exposure.

post-284-0-55172300-1582119270.jpg

 

F-stop: f/2.8, ISO 1600, 1/8 s exposure.

post-284-0-23312400-1582119292.jpg

Link to comment
I am not sure you are drawing the correct conclusion about the red plastic filter. It seems more likely that your stack is leaking red/IR than that the red filter leaks UV. Unfortunately with homemade filters like this it is hard to be sure.
Link to comment

I am not sure you are drawing the correct conclusion about the red plastic filter. It seems more likely that your stack is leaking red/IR than that the red filter leaks UV. Unfortunately with homemade filters like this it is hard to be sure.

I agree with Andy.

Red filters normally transmits much more red and IR than UV.

Any attenuation in the UV area where the sensor have much lower sensitivity will have a a big impact of what UV you can see.

Link to comment
Look better at the red plastic filter images, they are UV images at low brightness. Also, my red filter is purely accidental, it is from a 3D goggle, never meant to be used like I did.
Link to comment

Stefano: If I take a lot of photos in "multishot" mode (you keep the button pressed and the camera keeps shooting), the first one usually has less vibrant colors than the others.

 

That is some kind of problem with your camera. I had a Nikon which had some exposure anomalies when shooting a rapid-shot series. Later a firmware fix was issued for the problem. (I forget which Nikon it was! D2X maybe? Long time ago.)

 

It is difficult to say whether your camera's problem is a fluctuation in the saturation setting or a fluctuation in the white balance setting, or some combination of both? Have you searched online to look for other reports of this multi-shot problem?


 

Stefano: In videos I usually get stronger colors than in photos.

 

Again that is a problem with your camera. It is setting a stronger saturation in videos. Why? Who knows!!?


 

Stefano: I have fringing at the sides. Maybe my UV filter is too thick (4 mm).

 

This is most likely because of the lens. Inexpensive lenses are not well corrected for chromatic aberrations. They may also be "soft" in the corners and along edges of photos.


 

The possible problems caused by stacking filters are:

  • slightly reduced focus
  • induction of flare or hotspots
  • reflections off of inner lens parts
  • "ghosting" of bright areas
  • vignetting and discoloration at edges/corners of photos.

When problems like those listed above occur, you can often fix the problem by changing the angle at which you are shooting the subject.

 

Dichroic filters can sometimes have similar problems -- although usually with a dichroic filter you get a problem of annular discoloration on the outer edges of the photo. (Just crop in or learn a monochrome processing technique.)


 

I am not sure whether your 4mm stack of Z glass & BG glass is leaking some IR -- or --

whether the camera sensor simply cannot record enough UV (or possibly, deep enough UV).

In either case you would get a slightly less dark center on the dandelion.


 

For now, the camera problem and any possible problems from Chinese filter stacks are not too important. You *are* able to capture a UV-signature for your subjects which is "good enough". Documentary "perfection" can come later if you retain an interest in this area until such time as you can acquire better lenses and filters. I enjoy your experiments (and also other members' experiments) with non-perfect materials. We all learn from these kinds of experiments.

Link to comment

The red plastic filter test didn't work in any way like I might have hoped for.

So I don't draw any conclusion from it.

However, I commend you on all the work you did to try that out. Well done.

The fact remains that in all your dandelion photos, the center of the dandelion is warm, not black like the surrounding background grass. So there is some kind of out-of-band (other than UV) light mixing with the UV.

Because of the warmth (brown/red) of the flower center, the leak is likely in the 700nm +/- area.

Whatever the BG glass is, and whatever it's true transmission/suppression plot is, it is not cutting off the Red/IR. This could be a factor of the U filter visual range suppression even.

You tested with BG 4mm, and the color was still there.

So that BG is somehow not cutting the U filter transmission at the right place. I would say.

So I suppose it could be the U filter has a wider visual leak than I would expect. Just a wild guess. I really would not expect that to be the case.

This is the thing with Chinese filters, we really don't have accurate data on them, they vary a lot, and so it is hard to say exactly what you will get.

Link to comment

"---, we really don't have accurate data on them, they vary a lot, and so it is hard to say exactly what you will get."

 

That, in fact, applies equally well to the dandelion Taraxacum genus. There are myriads of apomictic microspecies and their UV signature are all pretty variable. Some of them have very small UV-black basal patches on the centre phyllaries ("petals" which is a misnomer for all species in the Asteraceae). However, styles and anthers should be very dark for all.

 

I'd like to stress the above as taking a "bull's eye" appearance of a dandelion in UV is not a given assertion.

Link to comment

I guess what Birna is saying is that the dandelion warm brown center could be due to some variety of dandelion.

Hmm, I am not buying that idea myself, as far as what I am seeing here, after all, I think it was Birna who pointed out the situation herself:

"The dandelion picture is an indication there is a slight NIR leak. Or your lens doesn't go very deep into the UV(A)."

https://www.ultravio...dpost__p__33147

 

I don't believe this is because of some dandelion variety or because of white balance, but rather because of some kind or Red/IR/700nm +/- leak.

 

...but the only way to find out is to start with a known setup, camera, lens, filter.

This setup is rather unorthodox, and that is all fine and dandy, even interesting if you will, but there are a lot of loose ends, and quite hard to make heads or tails out of it.

So I almost don't see much point in it, unless you have some way to further verify.

Link to comment
No, you misread me. The point was that the bull's-eye of a dandelion can be very small and partly masked by the phyllaries ("petals"). however, what *is* visible should be dark/black.
Link to comment

Well, there ya go, I know nothing about flowers, parts of flowers, and so on. It is a given that Birna knows more about flowers, I am not a flower person.

I just think it looks like a leak is all.

Link to comment

Of course it's a leak. No uncertainty about that.

 

Likewise, I *am* a botanist by training and by work practice through many years.

Link to comment
I have a 736 nm far red LED and a 660 nm deep red LED (and other 12 wavelengths). I don't have anything at 700 nm at the moment (Thorlabs sells a 700 nm LED if you want one). Can they be helpful in finding a leak?
Link to comment
Well, any light source is very intense if looked at directly. If you have a very high OD, nothing can penetrate a filter. For example, a typical BG glass (such as BG39 or S8612) has an extremely strong blocking at 940 nm, which can exceed OD 9 at 2 mm thickness. My camera couldn't see a fully lit "10 W" (actually 4 W) 940 nm LED with a piece of 2 mm thick chinese BG39 on the lens. I have a solar filter meant to be used on a telescope to observe the sun safely (NEVER point a telescope at the sun, you will have permanent eye damage). It should be a OD 5 filter, since that is considered the safety threshold. Through it, I can see the sun (of course), a halogen lamp filament, my 10 W LEDs, a <1 mW red laser pointer, a ZWB2 filtered 365 nm LED (yes, that surprised me), and I can even barely see outside on a bright sunny day, if I am in a dark place.
Link to comment

LED intensity varies with wavelength quite a bit, unfortunately.

Also, does the peak wavelength shifts as a function of power? I see a slight color shift in my LEDs, but maybe it is just the Bezold-Brücke shift.
Link to comment

Also, does the peak wavelength shifts as a function of power? I see a slight color shift in my LEDs, but maybe it is just the Bezold-Brücke shift.

There is a thermal effect and when you run it at higher power, it will run warmer. But I don't think it's that big. The problem is more that if you have, say, a red LED and a UV LED, the red one is much more efficient at turning electricity into light.

Link to comment
Today there are 365 nm LEDs with efficiencies reaching 35%, and maybe 40%. The technology improved a lot. Now they should be the most efficient UVA source, because I think that both fluorescent and mercury vapour "blacklights" are around 30% efficient.
Link to comment
It doesn’t matter what the absolute efficiency is if you are trying to test a lens or filter using different color LEDs. The point is that different LED colors have very different efficiencies. If one is 35% and the other is 55% how will you know if a brightness differential is due to your filter or the LED? Never mind the fact that they tend to just saturate the sensor.
Link to comment
If I want to make a qualitative test, LEDs should work just fine. It isn't difficult to force a leak in a OD 5 filter, but comparing the exposure times you should have a rough idea of how much and where your filter is leaking. The problem is that LEDs are not very monochromatic (why?), and so you might see a 700 nm leak with a 736 nm LED. Laser diodes are much more monochromatic, but I don't have them (except for that 650 nm laser pointer from my infrared thermometer which I showed in other topics).
Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...