Jump to content
UltravioletPhotography

The Challenge of Shooting at Night


Recommended Posts

I have only done this a few times, and some of my efforts are too dismal to be worth sharing here. In what follows, artificial illumination from mercury-vapor lamps (365.4 nm) ends up playing a big role, although high-pressure sodium lamps have displaced many of them in recent years (and those emit some UV as well, it seems.) I find it interesting not only because of what shows up, but because of what does not--certain light sources are completely blacked out by UV filters, giving a distinctly different look from what the eye sees, and often a distinctly different mood.

 

post-66-0-75221800-1414969888.jpg

"Dinnertime." B&W film, 403 filter, 5-minute exposure.

 

Note that the stack of signs at the left appears completely dark, as do the celebrated Golden Arches of McDonald's. Mercury vapor dominates the scene, but the sodium-vapor entryway light is clearly visible, as is a distant neon sign between the legs of the stacked signs. The pavement is characteristically pale and shiny, and the restaurant windows are dark despite the busy hour.

 

post-66-0-94693900-1414971851.jpg

"Snowy Night." B&W film, Baader U2 filter, 10-minute exposure.

 

It really was snowing when I took this. All mercury light here, including the box sign with internal fluorescent tubes. A surprising amount of the 365.4 line makes it past the phosphors and tubes and into the surroundings (have you ever noticed faster lampshade deterioration when you changed from incandescent bulbs to compact fluorescents?) I conceived this as an exercise in subject-space dialogue, a balancing between the central subject matter and the black void space surrounding it. It was shot across a road. but there was little traffic due to hour and weather, so no headlights appear. The one evergreen bush below the nearer light standard has its foliage suitably darkened by the filter, which gives greater contrast with the snow lying atop it.

 

post-66-0-88072400-1414973389.jpg

"Midnight in Times Square." Sigma SD14 with hot mirror removed, Baader U2 filter, ISO 200, 2-minute exposure; considerable detinting performed in workup.

 

A very different look from the film shots, and also from what the eye saw. Visually, the scene was dominated by very bright, garish colored LED billboards on the lower walls of the buildings. The Baader filter blacked these out very efficiently, shifting the visual emphasis to the older-style billboards lit with mercury illumination; and even those look somewhat smoky and obscured, due to their being made of UV-absorbing paper. I see this as revealing a darker and more contemplative vision of Times Square than what a visible-light photo would show. Another point of the interest is the light sources that show up red in the picture; I believe these are mainly incandescent lights, and that the red represents leakage of IR, to which the SD14 is highly sensitive. Even the Baader leaks some IR. In any case, it provides an interesting accent color.

 

post-66-0-47830200-1414977074.jpg

"Give My Regards to Broadway." Same location, looking a different direction, same exposure parameters.

 

The annoying radial ghosting of the Baader filter is readily apparent here when examining the point light sources. I tried to counterpose the statue of George M. Cohan at left against the inward-leaning stack of billboards on the right. Again, the incandescent lights provided a color accent. The SD14's sensor is more than a bit noisy under these conditions.

 

The previous probably represents the best of my nocturnal achievements to date. I have always wondered how lightning would look on film, but nocturnal lightning is uncommon where I live, and my storm-chasing skills are poor. I have made a couple of unsuccessful attempts to capture fireworks. Night photography by purely natural light is a challenge for me even in the visible, and has utterly defeated me in the UV so far; between the wretched sensitivity of digital sensors and the reciprocity failure of film I have not captured so much as a single star-trail. If any of you others have images, tales or experience to share about the short wavelengths at night, I would be delighted to see them here.

Link to comment

Nice examples Clark

'Snowy Night' is very iconic.

You have handled the difficulties of the SD14 very will in these examples. I am surprised to see this much IR leak in the Baader UV filter. I am getting close to getting one.......

I do have trouble with dichroic filter's reflections in the Sigma Foveon cameras, whether placed inside or outside.

Col

Link to comment

Really fascinating night captures, Clark. Thanks for posting them.

 

Col, yes, there can be a bit of IR leakage in the Baader-U. The newest ones are better than the older ones. Such leakage might manifest itself in long exposures.

 

It would be interesting to see comparison shots with other UV-pass filters or filter stacks.

And to see a comparison shot with a BG filter stacked over the Baader-U.

Good future experiment for some of you to try. "-)

Link to comment
  • 2 weeks later...
  • 11 months later...

Postscript:

 

I happened across a deserted playground at the south end of Fairview, Utah the other night, and thought that it might make good subject matter,

 

"At Play in the Time of the Owls." Sigma SD14 (hot mirror out,) Asahi 35mm lens, Baader U2 filter. 122 seconds @ F/11 and ISO 200. Tinted monochrome worked up in Photoshop.

 

post-66-0-65799700-1444792014.jpg

 

The noise level of this image is partly because it was worked up from the X3F version of the file rather than the JPEG (which did not look very good to me) and partly because I had to use ISO 200 to get decent image brightness.

 

The amount of halation seen here would be very unusual in a pure UV image, and I suspect that a lot of the halation seen around the mercury-arc lamps is visible bleed-through. The lamp housings may be engineered to attenuate the 365.1-nm line, which makes such contamination more likely. I find the owl-like appearance of the light standards with their halos striking.

Link to comment

I'm addressing another aspect here, viz. what kind of "IR leakage" is to be expected. The salient point is the outcome depends on the incoming light field and its spectral distribution as well as the filter transmittance as such. We cannot deduce much, if anything, about the spectral features unless both conditions are known. The final exposure is set by the level of spectral energy passing through the filter system with that light field as illumination source. If the illumination is say 104 stronger in IR than in UV, even the best UV bandpass filter will struggle to block IR completely. For sunshine, or a flash with uncoated Xenon tube, where the IR is perhaps 10 to 50 times stronger than UV, the filter functions flawlessly.

 

So it all depends on the circumstances.

Link to comment
What I take for visible bleed-through is apparent mostly at or around bright point light sources. I do not think much visible light reflected off surfaces is recorded in these images (these features show up mostly in the blue channel, whereas the blown-out lights show up in all channels in the color original,) The complexity of the channel processing in these cameras is, of course, a potential confounding factor. The grass in the foreground is much darker than the neighboring basketball court surface; that does not say IR to me, and I don't know if Hg-vapor lamps even emit much IR. My point is that in and around overexposed image areas, the sensing medium will record a greater proportion of stray photons, especially in areas that would be dark if the incoming light were pure UV. Halation is mainly caused by in-plane light-piping at the sensing surface. IR is good at traveling long distances through various transparent media, so IR images, especially on film, show a lot of halation. UV is not so good at this in most media, so UV images tend to have relatively little halation.
Link to comment
  • 2 months later...

Postscript 2: Hendersonville, NC.

 

Last night I took the Steinheil out at night for the first time, with the A900 and the Baader U2 filter. There was some amazing fog earlier in the evening, but I was tied up and could not shoot in it. By the time I got out, it was dissipating rapidly; I did, however, get two frames of enough interest to share here. I also wasted time trying to shoot a parking lot full of truck trailers at the local General Electric factory under what I thought was mercury illumination. The joke was on me: I think that the parking lot lights were actually LED, because those frames turned out totally black. Both frames were exposed 30 seconds at approximately f/8 and ISO 400. Display intent is BGR.

 

The Steinheil's infinity focus is noticeably degraded at this aperture, and the A900's sensor is noisy at 400 for long exposures. If I repeat this, it would be wise to leave the lens at f/16, back off on the gain, and stack several exposures. The A900's palette is wider and more nuanced than that of the SD14.

 

"Shady Dealings:"

 

post-66-0-07945800-1452353779.jpg

 

This is a veritable feast of light sources. In the background we see a mercury-arc lamp (greenish) and a high-pressure sodium vapor lamp (pinkish). Another mercury-arc lamp is visible just underneath the canopy. The light standard at upper right had the visual appearance of a mercury lamp, but it shows dim ruby red--what gives here? This is the wrong color for infrared leakage (which would be expected to show show either bluish or white in this color mapping.) An LED lamp would not be expected to show at all. I think the most likely hypothesis is that it is a mercury lamp, but that the combination of the lamp cover and the Baader filter have isolated the 404/408 line pair rather than the 365 line, exciting the camera's "blue" channel but not registering in the other channels. Some of the under-canopy lights above the fuel pump show a similar color.

 

Part way through the exposure, a second automobile pulled in to the left of the first parked vehicle. The component of its headlights show different colors. The yellowish component is presumably incandescent; but what of the greenish-yellow one? Is it a xenon/HID lamp? The Krispy Kreme doughnut sign in the background is more perplexing yet. The lower neon tube portion is completely blacked out except for an odd bright cyan dot which corresponds to no visible feature I know of in such signs. The upper portion shows a red "bowtie" pattern about where the lettering is, but again, I recall nothing in the sign with that shape and position. Around the windows of the building is a faint blue outline which may correspond to deep blue LED decorative lights whose spectrum might have some overlap with that of the filter.

 

"Ingles Supermarket, East Flat Rock:"

 

post-66-0-22808800-1452353803.jpg

 

Many internally lit signs (including the large one in the right foreground) do not show up as such. The mercury-vapor lamps on the underside of the canopy above the fuel pumps show the expected color. The main parking lot lights look like sodium vapor in this picture, but to the naked eye they appeared more like mercury vapor. This could be due to attenuation of the 365-nm line by the lamp covers to the point where contribution from the 404/408-nm line pair starts to be significant. The Steinheil exhibits a lot of radial flare from these point sources. The internally lit fuel price sign at the right is the more extreme case, showing bright red, most likely a fluorescent tube shining through white plastic. Different colors may be seen on the vehicles in the parking lot.

Link to comment

I am fascinated with the Ingles Supermarket shot and wondering also about all the lights.

 

Perhaps stacking with blocking filters would reveal which kind of "leaks", if any, play a role in a 30 second exposure?

 

I do hope to see more UV night shots in this thread over time. Very enjoyable.

Link to comment
  • 6 months later...

Postscript 3: Sugamo

 

The following are from a recent trip when I had a few moments to take some photographs. They are taken with the Sony A900 and the Steinheil Cassar-S 50mm at f/16 with the Baader U2 filter. I experimented both with stacked exposures (noise reduction off) and with single, longer exposures (noise reduction on.) Each approach had its pitfalls, as will be seen. Furthermore, as more and more outdoor signs are converted to energy-efficient designs such as LEDs, the opportunity to record signage in the UV (and possibly IR) decreases. I attempted UV in one alleyway that seemed ablaze with signs--but the UV image of that scene showed only a few dim features, not enough to make an interesting image.

 

The first example depicts another alleyway which does, in fact, show better features. This is five stacked 2-minute exposures at ISO 400.

 

"Fairy Dust Alley:"

 

post-66-0-80452500-1470322332.jpg

 

The stacking does seem to control random noise well; I had to mask in the highlights with a separate image stacked in Darken mode to control blowout. Hot pixels were a real problem as there were so many that spotting them all out was infeasible except in the dark void areas. Finally, I embraced them as a pictorial element. A mixture of sodium vapor, incandescent, and possibly other sources is seen. Orange-red on the large sign at top right is probably titanium dioxide; other deep red features are (fluorescent) light shining through white plastic and may represent the 404/408nm mercury line pair. There were some pseudowhite LEDs in the scene that are efficiently blocked by the filter and do not show. Extensive pedestrian traffic which was present did not register at all in the final image.

 

The second example is shot from a nearby pedestrian overpass facing oncoming traffic (on the right) and consists of four stacked 2-minute exposures at ISO 800.

 

"Nakasen-Dori at Night:"

 

post-66-0-39758700-1470323415.jpg

 

The distant streetlights look like sodium vapor to me; a few cyan lights may be mercury vapor or fluorescent. Traffic signals were LED and do not show. The distant sign on the top of the Hotel Mensels is white in this photo, as it is to the eye; it could conceivably be an old xenon-arc glass-tube sign. The approaching headlight traces are interesting. Some are deep red and continuous, whereas others are pale pinkish and show pulsation. A single one (motorcycle?) is pinkish but without pulsation. My guess is that the first two are incandescent/halogen and xenon-arc HID, respectively; I am not sure about the last case. The collection of greenish lights visible in the left lanes corresponds to the positions of vehicle taillights which stopped at a signal there; these sources were too dim to leave moving traces and it is not clear what wavelength is being recorded. The problem with hot pixels is also painfully apparent here, although I spotted out a few of the most egregious ones.

 

The final example was taken at ground level on the same street and consists of a single ten-minute exposure at ISO 800 with noise reduction on:

 

"Invisible Shopping:"

 

post-66-0-56673700-1470325038.jpg

 

There are no hot pixels, but there is a lot more coarse color noise and the image looks less polished. The cyan illumination is bare fluorescent tubes leaking at 365 nm (I do not think the phosphors are contributing much to this image.) The dim little yellow lights around the sign at right may be incandescent--I am not sure. The nature of the whitish lights in the middle distance is a matter for speculation. It is surprising how empty the scene looks--hundreds of pedestrians and cyclists passed in front of the camera during this exposure, but left nary a trace of their presence.

 

The scenes depicted here are dimmer than many of those earlier in the thread, and required long exposures. This is always a challenge with consumer digital gear due to the noise and hot-pixel issues. The stacking technique is one I have successfully used for star trails, but at lower ISO settings than here, and usually with 1-minute exposures. Perhaps the best technique to use here would have been one with more, shorter exposures at lower ISO and possibly with the camera's noise reduction on (although that requires camera downtime equal to the exposure time on each exposure.)

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...