• Ultraviolet Photography
  •  

UV at night? Who knows?

17 replies to this topic

#1 Andy Perrin

    Member

  • Members(+)
  • 1,704 posts
  • Location: United States

Posted 05 October 2018 - 02:57

I decided it was time for another round of that perennial game we play, trying to take UV photos at night. The following image was taken with 2mm UG11, 1.5mm S8612, and another 1.75mm S8612 on top of that (a three-filter stack altogether). The camera was the Sony A7S, the lens was the Novoflex Noflexar, and the settings were F/4, 30s, ISO51200.

I had no idea what I should do for white balance, so I picked the sky.

Image has been denoised with Neat Image, and also a preliminary autotone in PS was used before the final white balance.

Attached Image: _DSC0105.jpg

Edited by Andy Perrin, 05 October 2018 - 02:57.


#2 Cadmium

    Member

  • Members(+)
  • 1,896 posts

Posted 05 October 2018 - 03:27

I know you are using the 'real stuff' too (per the graph below). ;)
Attached Image: UG11_2mm_S8612_3p25mm_stack.jpg

#3 Andy Perrin

    Member

  • Members(+)
  • 1,704 posts
  • Location: United States

Posted 05 October 2018 - 03:48

It looks promising, however I *also* did an IR photo (1000nm long pass) last night of the same scene, and the area under that light on the sidewalk is full of IR, even at 1000nm+.

Attached Image: _DSC0098 UVP.jpg

Edited by Andy Perrin, 05 October 2018 - 03:51.


#4 Cadmium

    Member

  • Members(+)
  • 1,896 posts

Posted 05 October 2018 - 05:44

Not all 1000nm longpass filters are created the same... ;)
However, I don't know that there is anything not to be expected with the 1000nm longpass shot, is there, do you think?

#5 OlDoinyo

    Member

  • Members(+)
  • 651 posts
  • Location: North Carolina

Posted 05 October 2018 - 06:58

I have wondered if the A7S might be good enough for UV star trails or starry sky night shots. I do not quite see that here, but perhaps at higher ISO it might be within reach.

#6 UlfW

    Ulf W

  • Members(+)
  • 392 posts
  • Location: Sweden, Malmö

Posted 05 October 2018 - 10:22

View PostOlDoinyo, on 05 October 2018 - 06:58, said:

I have wondered if the A7S might be good enough for UV star trails or starry sky night shots. I do not quite see that here, but perhaps at higher ISO it might be within reach.
It looks like it was overcast in the IR-image.

A 1000nm long-pass filter is not the best filter to check IR-leakage from an UG11 as that filter has the IR-peak at 690nm.
Do I misunderstand the purpose of the second image?

Edited by UlfW, 05 October 2018 - 10:23.

Ulf Wilhelmson
Curious and trying to see the invisible.

#7 dabateman

    Da Bateman

  • Members(+)
  • 371 posts
  • Location: Maryland

Posted 05 October 2018 - 11:12

Light pollution is and will be too great to get anything from the stars. Try specificly shooting Venus. Thus what the Baader venus filter was actually designed for.
Sodium vapor lamps and metal haloid street lights that are common output uv and you will pick them up. That is what your seeing in your image.
Also some street light are still high pressure mercury vapour lamp and we know they have excellent 365nm output even when filtered. Plus the full 405nm mercury line is unfiltered and will look uvish.

So I think the light pollution will be too high.

#8 Andy Perrin

    Member

  • Members(+)
  • 1,704 posts
  • Location: United States

Posted 05 October 2018 - 14:14

View PostUlfW, on 05 October 2018 - 10:22, said:

It looks like it was overcast in the IR-image.

A 1000nm long-pass filter is not the best filter to check IR-leakage from an UG11 as that filter has the IR-peak at 690nm.
Do I misunderstand the purpose of the second image?
The 1000nm pic was taken the NIGHT BEFORE for its own sake. After that, I subsequently made a UV pic, and wondered if there may still be IR contamination despite 3.25mm of S8612. It occurred to me that I could check the LOCATION of any IR light sources in the older photo. Since it wasn't actually taken with the intent of checking IR leaks, that's why it's not the ideal wavelength range for that purpose.

Anyway, just now I noticed that the brightest IR is coming from the chain of christmas lights (fairy lights for those across the pond) in the window of Johnny D's. But the light in the UV pic is not in that place, it's actually coming from the barber shop awning light next door. So I don't think there is much IR contamination.

Quote

Sodium vapor lamps and metal haloid street lights that are common output uv and you will pick them up. That is what your seeing in your image.
Maybe in the sky, but the pool of light on the sidewalk is not coming from a street lamp. You can see several street lamps in the picture and they are not emitting much UV. I believe they are LED.

Edited by Andy Perrin, 05 October 2018 - 14:18.


#9 dabateman

    Da Bateman

  • Members(+)
  • 371 posts
  • Location: Maryland

Posted 05 October 2018 - 18:42

Looking closely at the photo the street light and the car head lights are blue. I wonder if that is a little bit of 400nm sneaking through due to your really high ISO. Its interesting that the Christmas light and barbers pole appeared well lit.

#10 Alaun

    Member

  • Members(+)
  • 232 posts

Posted 05 October 2018 - 21:05

Andy, did you have a clear sky or a cloudy one? A clear sky at the blue hour offers some uv.
Werner

#11 Andy Perrin

    Member

  • Members(+)
  • 1,704 posts
  • Location: United States

Posted 05 October 2018 - 22:02

I think it was cloudy. Any UV from the sky is probably scattering from far away street lamps as dabateman said.

The streetlights are not glowing at all - I think the flects of blue there are reflections of the car headlights. The 400nm is blocked to OD5 per Cadmium’s graph, but that might account for the tinge. I think the bright light from the awning is UV from the fluorescent lamp on the awning.

Edited by Andy Perrin, 05 October 2018 - 22:13.


#12 dabateman

    Da Bateman

  • Members(+)
  • 371 posts
  • Location: Maryland

Posted 06 October 2018 - 03:09

Well looks like I was wrong about the high pressure sodium street lights.
The following figures are from:
Sensors (2010):10, 3961-3988, spectral identification of lighting type and character.

Attached Images

  • Attached Image: Capture+_2018-10-05-22-55-15.png
  • Attached Image: Capture+_2018-10-05-22-54-09.png
  • Attached Image: Capture+_2018-10-05-22-53-45.png
  • Attached Image: Capture+_2018-10-05-22-44-55.png


#13 dabateman

    Da Bateman

  • Members(+)
  • 371 posts
  • Location: Maryland

Posted 06 October 2018 - 04:40

Using the above reference spectra, it looks like your street lights are metal halide. Using a Rg1000 and R720 filters with your UV stack you can figure out your nighttime surrounding lights.
Leds will be black, fluorescent lights will show up in uv but not with a RG1000, and sodium lights will show up only with a R720. Since you see something with RG1000, I think you have metal halides. Mercury vapour lamp I think are rarely used, but you would see strong 365nm band.
The first little bump peak in the above specta is the 365nm line. Which seems to be absent in the sodium light.

#14 Cadmium

    Member

  • Members(+)
  • 1,896 posts

Posted 06 October 2018 - 06:13

Shoot the same shot again, using an exposure you are happy with, first with the UV pass stack, then with the UV pass stack stacked with the RG1000.
Shoot both shots with the same settings and exposure time.
That will tell you if there is anything out of band getting through the UV pass stack.
I rather doubt there is.
Try the same test using other longpass filters, starting with something like GG420 or above (even GG400 if you want). This will test a wider out of band range, visual and IR.
Do the tests, and it will give you confidence about what you are actually photographing with subsequent future shots.
Otherwise, this is just one big guessing game.

Edited by Cadmium, 06 October 2018 - 06:15.


#15 Andy Perrin

    Member

  • Members(+)
  • 1,704 posts
  • Location: United States

Posted 06 October 2018 - 06:41

I don't have any GG420 or GG400. Here is a retest (new picture, same scene) with the original stack, the following processing:
-F/3.5 ISO32000 30"
-Neat image to denoise
-In camera white balance from PTFT during daylight this afternoon in full sun.

Attached Image: retest UV alone.jpg

Then I added on a real Hoya R72 filter with the same settings and processing. There was nothing (not even if you push the exposure up in PN, it's just noise):
Attached Image: retest UV+Hoya R72.jpg

I conclude that if there is a leak, it is not in IR past 720nm. It could conceivably be in the blue, but I don't really believe it.

ETA: Here is a repeat with Tiffen#12:
Attached Image: retest UV+Tiffen12.jpg

Again, the conclusion is that there is no IR or vis over 520nm (or wherever Tiffen 12 cuts on).

As an aside, tonight was a pretty clear night. I could see stars with my eyes even in the city.

Edited by Andy Perrin, 06 October 2018 - 06:54.


#16 Cadmium

    Member

  • Members(+)
  • 1,896 posts

Posted 06 October 2018 - 08:27

Cool! The UV pass shot looks a little darker than your first one, but the other stacks are black, so that makes me feel good, your UV pass is showing only UV, that is good.
Sometimes I do a full frame 'marquess' white balance, and sometimes I might white balance on something neutral, like asphalt or cement... I don't think I would use the sky for a UV white balance.
Regardless, these tests should rule out any out of band contamination worries when using your said UV pass stack.

PS: Just a suggestion, you may want to do the same tests with half the S8612. You may not need that much S8612 in your UV pass stack.
Best way to know is to use just one of the S8612 filters in your stack, and do the same longpass tests.
If you don't need that much S8612, then it would make your needed exposure time shorter.

Edited by Cadmium, 06 October 2018 - 09:02.


#17 dabateman

    Da Bateman

  • Members(+)
  • 371 posts
  • Location: Maryland

Posted 06 October 2018 - 11:30

Wait, your newest image is during the day in full sun. That is really dark sky. Or was the WB only done during day.
Also I need more sleep your street lights can't be metal halide, as you have no uv.
If your second photograph above is at night looks like the light pollution bouncing off the clouds is much better.
Is that point light off in the distance from a stadium? They still use mercury vapour lamp and will have a strong uv signal.

Edited by dabateman, 06 October 2018 - 11:42.


#18 Andy Perrin

    Member

  • Members(+)
  • 1,704 posts
  • Location: United States

Posted 06 October 2018 - 15:12

No images were taken in sun. I WHITE BALANCED in sun off PTFE. Everything is UV here (or blue at worst). There were no clouds last night, as I said, which made the sky darker. I have no idea what that distant light was. No stadiums in that direction that I know of.

Cadmium, the original image had a white balance off the cloudy sky, the revised image was white balanced using sun on the slab of PTFE you sent me, so it should be our standard WB for UV. The WB was done hours before the photos were taken.

Edited by Andy Perrin, 06 October 2018 - 15:19.