Friday, March 20, 2015

Ultraviolet Photography on an Amateur Budget

I recently became curious to try ultraviolet photography. Unfortunately I found very little information online, and an even smaller market for relevant products. The problem is that most people either aren't that curious, or have seemingly endless budget and patience.

So after three weeks of research and experimentation, I decided to publish my results for anyone impatiently curious like me. Turns out that, with care, you can get acceptable results for about $100 in eccentric, used camera equipment; however to ultimately discern between ultraviolet and infrared reflectance, I may need to look into spectrometry.

I wrote this as someone extremely conscious of how sensory overload, money, and fuss conspire against curiosity. This is about the bare minimum investment to get satisfying answers, not about tinkering. I actually hate tinkering.

First I had to sort through an enormous amount of confusion over infrared photography, induced flourescent photography, and ultraviolet contamination in ordinary photography. What remained was approximately three levels of interest:

  1. The popular level, where people use household equipment to produce pictures with lots of purple and blue in them, and call it "ultraviolet." Plenty of how-tos and Youtube videos here, which were useless to me.

  2. The novelty level, where people tear apart cheap webcams, remove internal filters, and apply filters of their own. I spent about $26 to try this myself, and was disappointed.

  3. The professional level, where people with disposable income make it a lifelong hobby. These were the people I learned from, because they tested a wide range of products and published informed results.

A lot depends on your goal. The very best set of equipment costs a terrifying $6,000, but to answer basic questions about the world it's completely unnecessary.

I have a background that includes a lot of color theory. I knew that almost all imaging technology works on a standard three-color system of red, green, and blue. This is because these primary colors correspond to the three types of cone cells in the retina of the human eye, and thus follow our "trichromatic" color perception.

Then I learned that birds and insects have the same three cone cells, plus one for ultraviolet light, making them "tetrachromats." I began theorizing about tetrachromatic color perception, and soon I was curious what colors normal, everyday objects are to tetrachromats.

If I could take a UV picture alongside a normal RGB picture and carefully measure the relative exposures, then I would have an image representing all four channels of tetrachromatic vision, and could calculate the color of anything in the image.

This was my angle. I found everyone pointing back to Dr. Klaus Schmitt, but much of his information is outdated or unclear. Shane Elen's page was also indispensible. But the real solutions are below.

I needed three basic parts: a UV-capable camera, a UV-only filter, and a UV-capable lens. At first the lens sounded unnecessary, but after failing twice I listened.

UV-Capable Camera

Nikon D70 is the answer. I paid $70 for one on eBay. Newer cameras might work if you don't mind having to modify them, but this means either more money or more fuss – not to mention risk of low UV response in the sensor itself, which requires more research to avoid.

As I understand it, no camera gives better functional results, which is lucky if you like throwing cosmetics to the wind for shabby, used equipment. This was me.

UV-Only Filter

This is where prices start to diverge. The top choice by perfect consensus of all my sources is the Baader U astronomer's filter, which has to be special-ordered from a German manufacturer. Their website isn't even English. At current exchange rates, the price is at least $200.

Schmitt and others' next suggestions were discontinued and now rare, so I went solo. With carefully chosen search terms I uncovered a $10 lighting filter at B&H and called on my childhood paper and cardboard crafting skills to make a holder for it. I also purchased a $5 IR-cut filter for good measure.

I suspect there are plenty of other reasonably priced solutions out there disguised as something other than a lens attachment. This item is more to illustrate than to dictate. Just be careful of non-UV contamination.

UV-Capable Lens

The only lenses still manufactured specifically for UV photography go for $5- or 6,000, like the Jenoptik CoastalOpt. A number of ordinary lenses have been tested to accidentally admit sufficient UV light, but these were by far the hardest item to research. There's a dizzying multitude of technical variables, none of which I had a full grasp of at the start.

Schmitt's blog was anything but clear on this point, whereas the list these folks provide is positively overwhelming. It wasn't until I found Enrico Savazzi's page that things started to make sense.

I was looking for an easy answer, and the easy answer is any EL-Nikkor lens, e.g. the EL-Nikkor 50mm f/4. These go for $20 used. The downside is that they're enlarger lenses.

I didn't understand this at first. Apparently these are lenses not made for cameras, but for enlargers, which are similar to cameras but used in dark room development.

I would read casual comments about these requiring manual focus, not realizing that this meant no variable zoom or variable focus assembly at all. "Manual focus" here means either entirely static focus, or devising or purchasing an entirely separate means of varying the distance from lens to sensor, even if it's just your fingers holding the lens in open air (which does work, although extension rings are obviously preferrable).

You can't change the zoom per focal distance except by buying a different one. The only variable built into the lens is the size of the aperture stop, which is actually delightful to play with (it's hidden inside the shutter action on most DSLR lenses). No electrical components also means no lens settings recorded in image metadata.

The 50mm one is limited by the Nikon D70 camera body to a maximum focal distance of two-and-a-third feet, pictured below. For all lenses, as you inch it away from the camera, the focal distance drops. I'm nearsighted, so I'm comfortable with two or three feet anyway. I didn't need good equipment for closer shots, so I opted against extension rings for now.

In the foreground above is a U.S. dime leaning against an acorn cap. The general equation is graphed below for a variety of lens focal lengths. The shots above are marked on the graph below by crimson dots.

The last factor is lack of a standard camera mount. Instead of the Nikon F-mount, EL-Nikkor lenses come with screw threads – usually the "Leica M39" (check it here). Adapters go for $4 used, but if you can find the lever a lens would press and wedge it open (see photo below), the camera software will let you take pictures with nothing in the mount. I've had continuing success with the EL-Nikkor taped in place while I wait for the adapter to come in the mail.

Results, Apr. 18th, 2015

The image at the top of the page, showing sunscreen on my left cheek, was promising enough that I wrote this blog immediately after I took it. But I had to wait to write a "results" section, because the best way I knew to gauge the infrared contamination was Shane Elen's dandelion pictures, and this is the first week dandelions are in bloom in my area.

This was obtained using both my UV-pass and my IR-cut filter, and required a heavy contrast adjustment in post-processing to bring out the gradient in the flower. These results are far too dirty and uncertain for me to measure tetrachromatic colors like I originally wanted. Ultimately I just can't know from these pictures which frequencies are showing in the image.

What I can say for sure is that the images are from primarily invisible frequencies, probably including a decent amount of ultraviolet. Patterns are sometimes discernible with the filters where they aren't without, and vice versa. In sunlight these UV photos require an exposure increase of about 150 times.

A scar in the pansy petal, visible only in UV, is highlighted. Next are English daisies, a plain yellow daisy, and a red pansy whose pattern vanishes in UV and which becomes translucent.

This is one way to do ultraviolet photography, but it seems the optimal approach depends dramatically on your needs. I actually learned far more about lens optics and macro photography than about the ultraviolet world I set out to study. But more importantly I know to try spectrometry next.

$6,000 might also do the trick, but that would approach my problem backwards.

No comments:

Post a Comment