The (Limited) Chromaticity of Fireworks

[OlDoinyo]

Fireworks are designed to be showy to the eye, with vivid colors, intricate designs, and loud sound effects. One might be curious how showy they appear outside the visible band, and to this end, a few fireworks at the Lake Toya nightly firework show were photographed in ultraviolet. The Steinheil Cassar-S lens was mounted to the Sony A900 camera and set to f/16; filtration was the Baader U2 filter. The camera was aimed blind and triggered by wire on a tripod. ISO was set at 6400 and exposures varied from about 3 seconds to 10 or more. A standard white balance was used.

 

The initial example is of a simple mortar shell.

 

"White Urchin:"

 

 

No chromaticity at all is apparent in this image. The small dot at the center of the frame is the dispersal charge and the radial streaks are the luminous projectiles (called "stars" in industry parlance.)

 

A series of ground fireworks bursting at a low altitude produced the next two images.There is more resolved detail in these images than in comparable visible photographs due to the relative lack of halation in UV photographs.

 

"Fire Dandelion:"

 

 

 

"Fire Broccoli:"

 

 

The latter image shows some slight reddish chromaticity from the stars, but none from the vertical ascending streaks.

 

There are in fact reasons to expect what we see here. Fireworks emit light in two different ways. The first mechanism is simple incandescence, which is dominant in bright magnesium flares and similar sources as well as the explosive charges themselves. Such emission is broad-band in nature and not likely to generate large differential channel responses. The second mechanism is line emission from metal salts and other compounds which generate the bright visible colors. These substances are chosen for their visible emission lines, and may have few or no emission lines in the near UV. A possible demonstration of this lies in a depiction of another mortar shell; this bicolored round had contrasting axial and equatorial colors.

 

 

The equatorial stars are reasonably imaged in this (not very good) photo; the trails of the axial stars are almost invisible, as if they contained almost nothing which could generate UV light.

 

One might speculate, based on these results, that fireworks might also show very little chromaticity in the infrared; but consumer gear is not good at producing differential channel response in this range, so that it is unlikely that simple single-camera exposures could explore the truth of this conjecture.

[Cadmium]

Clark, I like that 3rd pic a lot.

[Andrea B.]

This is an excellent presentation of an interesting topic with lots of information. The photos are interesting in their own right aside from the UV record. I particularly like Fire Broccoli.

 

I am inspired to try this next year when I view the fireworks display in Maine where we spend some time each summer.

 

I wonder if using IR-bandpass filtration might be an interesting experiment for fireworks to artificially construct some chromacity? Do you expect that there might be different wavelength responses? (I'm thinking of the MaxMax IR-bandpass filters.)

[Andy Perrin]

This was very interesting. I was wondering last July what fireworks would do in UV.

 

Andrea, I'm not sure how you would set that up, because you need to capture with multiple bandpass filters AT THE SAME TIME? The fireworks won't sit still while you change the filter!

[Andrea B.]

I didn't think that one through did I??

 

I just ordered those filters recently. I'd always wanted to try them.

[OlDoinyo]

Yes, a side-by-side 3-camera array with simultaneous triggering would be best (I have seen comparable 2-camera rigs for IRG work.) You could of course attempt a tandem beamsplitter box, but...

[OlDoinyo]

Postscript: 4 July 2018, Lake Julian, NC.

 

The following obtained with the Sony A900 and the Asahi 35mm lens fitted with the Baader U2 filter. Exposure 8 seconds at f/11 or f/16 and ISO 3200. Display intent BGR.

 

"Fire Puffs:"

 

 

"The Porcupine's Bad Hair Day" (part of the finale:)

 

 

"Fire Thicket" (another part of the finale:)

 

 

"Fire Ikebana:"

 

 

"Fire Orifice:"

 

 

"Fire Lily:"

 

 

With this last one I had some fun and formatted it in the manner of an Oriental scroll or poster.

 

I think the Asahi lens obtained somewhat sharper images than the Steinheil got, and the modestly reduced bandpass is not much factor given the nature of the subject matter. The chief merit of UV here is the ability to resolve fine detail with little halation, even in the dense displays. Otherwise, I must admit that there is little advantage over visible photography. It is interesting that some of the dispersal charges produce much dimmer central dots in the images than others; perhaps this is due to differences in magnesium content. Intermittent trails are due to sparkling projectiles, while trails that begin far from the dispersal point may be due to projectiles that changed color and emitted UV only in one phase of their trajectories.

[Andy Perrin]

Pretty!

 

Otherwise, I must admit that there is little advantage over visible photography.

Hm. How sure are we that the UV isn't being overwhelmed by massive amounts of IR?

[OlDoinyo]

Hm. How sure are we that the UV isn't being overwhelmed by massive amounts of IR?

 

It is a difficult thing to prove, but IR leakage 1) tends to look distinctly greenish with the white balance used here, and 2) shows increased halation, due to the penetrating nature of IR. Neither appears to be the case in the above images.

[Andrea B.]

The Porcupine's Bad Hair Day and Fire Thicket are my favorites. So cool!

Much thinner trails for sure than I got with my recent visible attempts on the 4th of July.

 

Nice presentation with the scoll. (What does it say?)

 

I think that running at ISO-3200 for 8 seconds, even at f/16, would blow out the frame if there were much IR present? It takes longer than 8 seconds to pick up much IR, if present, with a BaaderU. And, as Clark said, the scarcity of 'blobs' anywhere also indicate we are seeing something else than Vis/IR. Well, just my 2$ worth. Doesn't matter anyway because the photos are so interesting. :D

[OlDoinyo]

What does it say?

 

If Google Translate has not misled me too badly, the mauve text reads 'shigaisen' (ultraviolet;) the green text reads 'hi no yuri' (fire lily.)

[Andrea B.]

Cool!

 

When I rework the Visitor's Gallery (time give me more time please universe), I would like to include one of your fireworks photos.