Ultraviolet photography at night

[kogakunippon]

My topic title says it already, I am working on ultraviolet landscapes since the last weeks, and I can tell you it's very interesting what i have figured out so far. I started with my modified Nikon D7100, an EL-Nikkor 63mm F3.5 and the Baader U-Filter. Then modified a Kyoei W.Acall Type II 35mm to work at infinity on my Nikon DX camera, going straight to a Noflexar 35mm with Exacta Mount. Next lens I have tested was the Lithagon 28mm..... with Little success as i have to admit.

 

Al thoose lens are good when using them at bright sunlight during the day. But at night when you can't step down to f/11 they are simply the wrong choice. Needless to say that the exposure times at night go through the Roof.

 

30 seconds at ISO 6400 is not enough at f/5.6, as you need around 180 sec to illuminate a building correct at this aperture, and most lens dont work well below f/5.6

 

The usable light sources are also very limited when you plant travel by plane.

 

Very interesting for me was just the fact, that the ordinary Ai Nikkor 50mm f/2.0 works better at open aperture then my Noflexar 35mm f/3.5 at f/5.6!?

 

Somebody ever tried UV-Nightshots?

[nfoto]

Not surprising as these older 'fast' Nikkors are pretty good in the upper end of UV-A, perhaps usefully responding already from 370 nm upwards. That in combination with the faster aperture would cut exposure times significantly, I'd reckon.

 

In my recent informal UV lens test; I included the successor to your Nikkor, namely, the AIS 50/1.8 'Longnose'. It delivered more than acceptable in UV provided one is willing to restrict the bandwidth coverage in UV correspondingly. For a lot of natural subjects in UV, this might not be the hurdle one could imagine in terms of UV response from the subjects themselves.

 

I do have several copies of the 50/2 Nikkor spanning at least the production years 1966 to 1977, thus I might try them later to learn whether they differ in their UV response.

[Andy Perrin]

Somebody ever tried UV-Nightshots?

 

 

Yes, I tried here, although people wondered if I may possibly be getting out-of-band signals too (e.g. infrared or visible).

[OlDoinyo]

You may be interested in my post:

 

http://www.ultravioletphotography.com/content/index.php/topic/1089-the-challenge-of-shooting-at-night/

 

Film still outperforms most digital sensors where light is limited. I don't know if you have any access to it.

[Cadmium]

Hi Wolfgang, I am curious about the three round 'yellow' circles on the left edge of the image. I am also curious about your white balance?

[kogakunippon]

The yellow circles came from a street lamp beside the building. No White balance was made for this Picture as ist just a test how far the MTE U-303 can illuminate. In this Picture the distance was 53 meters.

[Cadmium]

Wow, an illuminated building with an MTE?!

I would expect a fairly monochrome white balance form an MTE, and the 'pink' is the MTE, and the 'yellow' is the street lights. Not sure how much of the street lights or other lights play a role in building illumination,

but the yellow would suggest a broader/deeper UV band of light, than the MTE.

Even the painted window trim seems to have slightly more color to it than I would expect from 365nm light.

It would be interesting to see this shot white balanced, either with Photo Ninja or NX2 or NX-D.

Still hard for me to fathom illuminating a whole building with an MTE, but then 180 seconds is a long exposure.

Thanks Wolfgang, very interesting.

[kogakunippon]

Hi Steve,

I did a lot of research and testing with this torch because i want to make some night shootouts in Tokyo this weekend. One week later I will be able to tell you guys what’s possible at night and what’s not.

 

The white balance was not my major interest (during the tests), as the lens makes the big difference. All slow lenses are not really useful when you have close to no UV-emissions at night. On location I will use my PTFE board and place it somewhere inside my picture for easy and correct white balance correction afterwards.

 

And the slight focus shift from the 50mm Nikkor is not a big issue either because its suitable at f/2.0 already. My best other “non-dedicated” UV-lens must be stopped down min. to f/5.6 which means 8 times longer exposure times. If we talk about 1sec to 8 sec it’s not an issue, but if f/2 means already 240 sec it’s getting weird.

 

The very long exposure times are also a serious problem for the camera sensor and limit the photographer about how many subjects he can make in the few hours during the night time (the metro is not operating in Tokyo after midnight).

[colinbm]

Yes that is good to know Wolfgang, that the MTE can illuminate a building at this range & with a long enough exposure you are able to capture it.

I would never had thought this was possible. Thanks for showing us B)

Col

[OlDoinyo]

Can there be any question of white balance with a monochromatic light source?

[kogakunippon]

Every non-contaminated ultraviolet picture contains just monochromatic light.

[Andy Perrin]

Every non-contaminated ultraviolet picture contains just monochromatic light.

That isn't true. We typically get between 340-400nm pretty reliably, depending on how good your lens and filter are. There has been some discussion of how the UV false colors correspond to wavelength, but in general the consensus seems to be that you start to get yellows (in our standard white balance scheme on this forum) if you get out lower than 370nm. Higher than that you see blues and greens in our standard WB. There would be little point in our discussions of filter stacks and so on if all the light were a single wavelength!

[kogakunippon]

I don't know anything about the consensus from other forum members, but I believe what you are telling me. False colors can be interpreted in many ways, depending on the WB settings and post-processing.

 

But what I know is, that even in the visible spectrum everything from 620 nm to 750 nm is red, with slight differences, but its still red for human eyes. Which means there are no other colors within a bandwith of 130 nm.

 

While my Baader U-Filter let just wavelenghts pass between 330 nm and 380 nm.

 

I really don't know what could be the difference within this 50 nm if its produced by the sun or my torch. If there is no leakage or contamination from visible or IR light the result should be quite the same, or not?

[Andy Perrin]

kogakunippon, even if you take just 620nm to 750nm, and use a filter for that range, your camera will produce a colorful image from that if you white balance it against an appropriate standard. That is, the colors the camera produces will roughly map to the range of wavelengths you choose, if you rescale things properly. You have to have a REALLY narrow band (like 10nm-20nm or so) to get inside the range where there is no variation in the surroundings due to wavelength. Obviously that also depends on what you are photographing.

 

Also, if you are using the famous Baader with a suitable lens, you should definitely be seeing color variation if you take photos of flowers (which tend to be the only things in the surroundings that reflect much of the shorter wavelengths, hence the interest around here...). I think you are not white balancing the camera, based on what I see above?

[Cadmium]

The image is not entirely monochrome. My results with 365nm MTE shots in UV-only is that they are fairly monochrome. That being said, the images above do show some faint amount of blue (like around the window trim).

I can run the image through PS levels, which shows most of the image in Black+White with some blue around the trim (see example image below),

however that is not a very good way to do it, and white balancing the image from RAW would be preferable.

 

With the Baader U, the only possible contamination I might consider is from the very long exposure and IR, but IR with a Baader U would look white/gray/monochrome, because the Baader would show IR in the higher range (~900nm~), which is hard to distinguish from the rest of the image usually, because it doesn't have a false color.

 

What I am really curious about are those yellow lights. Are those outside the window? Are you shooting through a window? Are they a reflection off the inside of a window?

Why are they yellow? They must have some slight amount of UV in a lower range than the MTE.

 

I think white balance would be something interesting here, and you don't really need the target to get a general idea with the programs I mentioned.

You just need a NEF version, and one of those programs.

This would show the blue much better, however faint, and I wonder what it would do with the yellow too.

 

Thanks Wolfgang, very interesting!

 

[Andy Perrin]

My results with 365nm MTE shots in UV-only is that they are fairly monochrome

Yeah, if you use a monochrome light source, your results will indeed be monochrome! It's not clear what the real spectrum of the lights are in the nighttime photos (my bridge pics included). We are not even certain of the type of lights in those. In sunshine, though, there is certainly color to be found!

[Cadmium]

That is kind of what I am curious about, what the mix if light is, most of it looks like it is the MTE however.

MTE 365nm light is not entirely monochrome, just 'fairly monochrome'. Some blue can be seen in this rough test I did a while back.

I was trying to 'paint' the frame with the MTE light in this test, and the exposure time was not very long, so the MTE shot may not be very evenly painted, but it still gives me a general idea of differences

(there other two shots were not 'painted', because those lights covered a larger area).

 

Wolfgang, I may try this myself some night soon. Did you point the MTE at one spot, or move it around to 'paint' the scene?

Thanks.

[Andrea B.]

Don't forget the Spoon Test (Search) to determine whether your UV-LEDs are leaking visible light. Most leak a bit in the violet/blue end of the spectrum. Such torches should be filtered with a Vis Blocker if a purer UV output is desired. For example, when making UVIVF fotos.

[Cadmium]

All of my MTE torches have U-340 2mm on them to prevent visual leaking, but even so, with UV-only shots there is normally no need to filter the MTE.

There is the situation of the long exposure time, but still I would expect no visual leak with the Baader U.

Given Wolfgang's set up, I would expect everything in those shots to be UV-only light.

However, for UVIVF filtering the MTE is advisable.

I have not compared the visual leak between the MTE 301 and 303.

[Andrea B.]

Wolfgang: Every non-contaminated ultraviolet picture contains just monochromatic light.

 

Don't forget that the UV light has to pass through the Bayer filter on your converted camera. So with a wideband UV-pass filter and a wideband UV-capable lens, you are not usually going to get a monochrome UV image unless you are shooting a monochrome subject. (You can make the same statement with Visible light.) Typically your UV image will have some false colour variation. If you look at the raw data, all three of the R, G and B channels capture some of the UV light in different amounts (given a non-monochrome subject).

 

And also, don't forget violet. Where does violet really "end" and UV begin?? Some say violet ends at 400 nm. Some say violet ends at 380 nm. I think that if you are capturing anything at all between 380 - 400 nm, then you are likely capturing some portion of violet light. It is not exactly clear to me which part of the Bayer filter that violet light goes through. Mostly thru the Blue cell, but looking at the few examples we have of sensor sensitivity, it appears that small amounts of violet light can also pass through the R or G cells (or both).

 

The point being -- that whatever violet is captured also contributes to the non-monochrome-ness of the resulting photo.

 

[sorry about the Violet Ramble there. Lately I've been obsessing over Violet light and problems recording it.]

 

Does anyone know the FWHM for a Nichia 365nm UV-LED?? I think I recall it to be approximately 10 nm?? So if your UV illumination is only that UV-LED, then you will probably get a monochrome UV image to some degree.

[Andy Perrin]

Andrea, regarding the violet problem, we know that the UV bullseye markings (in flowers that have them) can certainly be seen in 380nm-400nm, but we often can't see those with our eyes. So if "violet" is partly a matter of defining a sensible threshold, the 400nm one seems more reasonable to me.

[Cadmium]

I don't think the violet in Wolfgang's pics (around the window trim) is coming from 400nm+, I think it is under 400nm.

 

Wolfgang, you might be gone on your trip now, but if you are there, are the yellow street lights at the left side of your first shot between the camera and the tree, or behind the tree?

[Andrea B.]

Andrea, regarding the violet problem, we know that the UV bullseye markings (in flowers that have them) can certainly be seen in 380nm-400nm, but we often can't see those with our eyes. So if "violet" is partly a matter of defining a sensible threshold, the 400nm one seems more reasonable to me.

 

I agree that 400 nm is a useful and reasonable demarcation for all practical purposes. :D

 

But such demarcations are about reflected UV and violet. Whatever capabilities we may have had to see anything between 380 - 440 nm do decline with age. Given that the bullseye patterns in flowers are absorbed UV (and maybe, in some cases, a bit of violet also, who knows?), I have to think for a minute about how that might be affected by the yellow eye lens problem. I'm not sure.

 

**********

 

If anyone is not familiar with the yellow eye lens problem, here it is:

 

As we age the lens of our eye begins to turn yellow. A yellow lens blocks violet/blue. Thus, whatever natural capability we may have had to detect colours in the region between [APPROXIMATELY] 380 - 440 nm declines with age. We are not aware of it on a daily basis. Our eye pigments and brain and experience still tell us that the blue sky is blue.

 

I have had this brought home to me in a big way after surgery on my left eye to remove a damaged lens which had formed two cataracts, one inner and one on the surface. [Old story, mentioned before.] The lens was replaced with an implant which protectively blocks UV and some violet/blue. The implant curve approximates more youthful vision. And so, with the repaired eye I can see much more violet and blue than I can with my right eye.

 

With my left eye, the sky looks blue. With my right eye (the old eye), the sky has a cyan tint.

With my left eye, I see a larger diameter beam from the Nichia 365nm UV-LED when shined on a wall.

With my left eye, I can see what I think is a spectral violet reflection when performing the spoon test with the Nichia 365nm UV-LED.

 

Now let those be very informal observation, OK? I don't have any way to actually measure the actual colours I am seeing - or not seeing. I can only tell you that I have observed first hand the effects of an aged, yellowed eye lens.

[Andrea B.]

Reference: There are many, but I liked the chart and comment in this one. (And the paper is interesting also. We have a harder time maintaining circadian rhythms as we age.)

 

NOTE: Chart is RELATIVE, not absolute.

 

Protecting the Melatonin Rhythm through Circadian Healthy Light Exposure

by Bonmati-Carrion, et alia

17 Dec 2014

International Journal of Molecular Sciences.

 

6.1. Aging

The human circadian system, like other cell, organ or systems in the organism, undergoes processes of maturation and aging. In the circadian system, profound alterations occur at all levels, from the inputs to the outputs and in the circadian clock itself. Regarding the input pathways, there are some structural and functional changes. Thus, as absorption in the crystalline lens for shorter visible wavelengths (400–500 nm) increases substantially with age, at the same time that the pupil diameter tends to decrease (miosis), the effective retinal exposure received under the same ambient lighting conditions is lower in the aged, as compared to the young, eye. Very elderly individuals retain just 10% of the photoreception of a 10-year-old, and therefore would require ten times brighter exposures from identical light sources to maintain youthful circadian performance. Fig 5 shows the decline in circadian photoreception over the decades and its improvement following cataract surgery, implanting various intraocular lenses (IOLs).

[Andy Perrin]

Yeah, so the "edge" where UV begins goes toward longer wavelengths as you get older. For the 95 year old, it looks like the 50% threshold is like 475nm even! But even for the ten year old, their eyes will not see much of the <400 stuff unless all the other visible light is removed first, just like our cameras, because it will be drowned out.

 

Depressing to think I'm already seeing only 60% of the light I could see when I was ten, even on the red end.