UV-Green object

[Hornblende]

EDIT: oups, should have created the thread in the UV/IR experience section..

 

Hi all,

I would like to make an inventory of all objects appearing green (what I call "Uvert") in UV pictures, under sunlight. We don't see this color very often.

In my case I use an EL-Nikkor 80mm f/5.6 with a Baader-U filter. Custom white balance is made on either PTFE, snow or concrete.

 

Red lights:

 

More to come.

[OlDoinyo]

It would be interesting to speculate how uvert (BGR) might differ from uvert (RGB).

[Hornblende]

Sorry I don't really understand what you mean, do you talk about channel swapping?

If this is the case, here is a red-blue channel swap:

 

Here is another UV green "object": brake lights

[OlDoinyo]

The traffic signal is so invariant under the channel swap that one might think that only the green channel is registering signal. But at least for most cameras, there is no UV signal capable of doing this, because there is too much channel overlap. On the other hand, most UV filters are extremely opaque to actual green light, so this is not easily explained as visible leakage. I admit to being puzzled by this. I assume that these are incandescent light sources rather than LEDs.

[JCDowdy]

This is an intriguing thread.

 

Stop lights or brake lights, whether red-LED or red-filtered incandescent, would not be expected to be emitting UV.

 

The other colors in your images look otherwise typical of UV neutral WB, but I am wondering if this is some kind of WB anomaly. What kind of PTFE do you have?

 

I assume these images are from your Canon 6D full spectrum, is this UV-green "out of camera" or the result of how you post processed?

 

This may be unique to your Canon, which would also be interesting, I will see if I can replicate the green brake lights with my Panasonic.

[Hornblende]

WB anomaly could be plausible. But even then, it means that compared to a neutral target these green-UV objects are more recorded by green pixels than blue and red pixels.

My PTFE comes from here.

 

This UV green comes straight out of camera. For processing, I only adjust the levels and increase sharpness and contrasts. It makes no differences for UV-green.

[JCDowdy]

That PTFE is not claiming to be "virgin" but I am not sure that would differ enough to cause your green.

 

Looking at your brake light image again, it seems to have a greenish cast most noticeable on the asphalt street surface.

 

These greens seem to pop up unexpectedly sometimes and are often cause for much speculation. If your observation is reproducible across camera brands it may provide a clue as to the cause of this.

[Hornblende]

Here the WB has been adjusted on the asphalt to get ride of the greenish cast.

You can see the green value (V) of the brake light is significant compared to red and blue value.

[OlDoinyo]

I noticed now that there are some green-looking light sources in the Nakasen-Dori picture I posted a few months back.They are dim, and occur in a place where they would be taillights of vehicles stopped at a traffic light. The moving light sources are too faint to leave trails. After thinking the matter over, I believe I may have some insight.

 

1) To appear green, the camera's "green" channel should register the strongest signal; the "red" and "blue" signals should be weaker and approximately equal. This observation is borne out in post #8 above.

 

2) The relative signal strength at a given location in the image is influenced by the light itself and the intrinsic channel sensitivities, but also by the white-balance parameters chosen to achieve overall scene balance. Since the green dye of the typical reseau is more opaque to UV light than the red or blue dyes, a fairly extreme boost must be given to the green channel to achieve scene balance in daylight. Infrared also requires such adjustments, but the green boost needed for a hard black-IR filter such as B+W 093 is typically less severe.

 

3) To appear green,then, the light registered must be in a frequency range where the overall sensitivities of the red and blue channels (taking WB into account) are roughly equal and where that of the green channel is greater. Such a UV range probably exists; the B+W 403 filter appears greenish-blue (in BGR; RGB might render it yellow-green.) However, taillights and red traffic lights are unlikely to emit the needed range: red LEDs certainly do not emit UV, and the thick red plastic of a taillight lens would appear quite sufficient to quench any UV emanating from an incandescent filament. This leaves us to consider out-of-band leakage.

 

4) Green leakage is improbable due to the very high opacity of typical UV filters and taillight-lens plastic to green.

 

5) Deep blue leakage might occur, but it would not register as green; it would appear deep blue (RGB) or deep red (BGR.)

 

6) This leaves us with a third possibility: a range out beyond 750 nanometers, where all the reseau dyes become fully transparent. Incandescent filaments emit plenty in this range; red LEDs may have some tail emission here. Taillight plastic does not absorb these wavelengths at all, and UV filters are known to have a minor transmission window here. Because the white balance is set for UV scene balance, there will be excessive green boost for this IR signal,causing it to appear greenish. This hypothesis at least seems to fit known facts.

[Andy Perrin]

It seems like you could test that by doing a UV white balance in-camera, then switch filters to a >800nm IR long pass. IR-white things should then look green in daylight, if I follow you, OlDoinyo.

[Cadmium]

Hornblende, This topic has been of interest to me since you posted your first example of the green UV stop light.

With all the UV shots I have floating around my hard drives, I can't think of a single shot that has stop lights or tail lights in them, so I have been thinking of going out and shooting a few to see what I come up with.

I have encountered green before in UV only shots, but on foliage and such. Also sometimes I get what I call 'the gold look', which I can not exactly explain.

Both have something to do with white balance, but then that it true in any UV shot, and I tend to use the same white balance methods, several or more to choose from,

but even so the tendency for green or gold in certain photos isn't simply white balance.

So I am fascinated with your green UV lights.

[Andy Perrin]

Is the gold thing any different from our usual false-yellows? The sky is yellowish in UV, so you see the reflection of the sky in the leaves? (By the way, I love both those photos, particularly the one on the left with the spots.)

[OlDoinyo]

Just tried it--UV white balance (2700G9) shooting through a Zomei 850 filter definitely gives a green image.

[Hornblende]

Thanks for your interest :)

I recall seing a UV flower picture with green leaves in this forum, I don't remember the name unfortunately.

 

Cadmium, in your "gold look" picture we see that the grass is not black but yellow, what would happen if you adjust the WB to make the grass black? By the way these images are sublime.

 

OlDoinyo, in conclusion, your best bet is that IR is leaking through the Baader-U?

I was wondering something: in my portait image, my scarf and hat appears yellow, and I know that these objects strongly reflects near-IR. Could this yellow color be also the result of IR contamination?

[Cadmium]

It would be highly unlikely that your Baader U is leaking anything in the Red/IR 600-700nm range. That is not where a Baader U leaks, they leak up higher, around 900nm, but only when pushed.

[Andrea B.]

Everything affects false colour in UV. Lens, light, camera, converter, white balance method, white balance tool type and -- most especially -- the filter and its transmission peak and its bandwidth.

 

So, it would be best to also look at the raw colours in an app such as Raw Digger before any white balance is applied. And for standardized comparisons, show your white standard in the shot and describe how you converted and white balanced the photo for presentation here.

 

It doesn't take much deviance from a true "click-white" (directly on a diffuse, white reflective standard which is stable in UV, Vis and IR) to produce interesting false colours which are not the typical blue or yellow. (Here I'm assuming a broadband UV-pass filter which does not leak violet.)

 

When shooting UV landscapes at medium to long distances with lots of foliage using the BaaderU, I quite often get some green tones after such a standardized white balance. When shooting very close (such as botanical specimens), we less often see greens. But, like I said, it only takes a tiny nudge in one direction or the other to drift into dark cyan/green tones. Or to drift the other way into violet or purple tones.

 

That said, I have no experience with shooting tail lights or stop lights in UV. But I will be sure to investigate when I get the chance. (Off on a trip right now.)

 

*****

 

Martin, test for IR leak by putting an IR-pass filter over your UV-pass filter and making a long exposure. Some of the filter test posts here show examples of IR leak discovered this way.

[Andrea B.]

Continued.......

 

Here are some photos made with 4 different filters and for which white balance was NOT made in the standardized fashion (on a diffuse, reflective, stable white standard). Intead, the white balance was made in the first set on a leaf. And in the second set, an average white balance was applied. Both these methods produced greens, golds, dark cyans and some violets.

http://www.ultraviol...ons-per-filter/

[OlDoinyo]

I believe that the 340BP10 and 325BP10 filters show a green-yellow hue in UV (Cadmium verified the general range with a diffraction grating, which cannot be fooled by anything out-of-band.) I nominate these items as authentic examples.

[Cadmium]

Clark, Yes, those do, but I have never seen reflected green, or even illuminated green before in UV.

Have none of us gone out to take photos of stop lights and tail lights yet to test this our selves yet?

 

Andrea, referring to your link, are you absolutely sure you have U-360 that is 1.75mm thick? Exactly where did you get that thickness?

 

Also, I kind of disagree about testing for leaks. I mean stacking a UV-only filter with an IR longpass filter is something that will show where a UV-only filter might be leaking,

but it is 'pushing' it when you use a longer exposure.

If you want a real leak scenario to judge a filter by, take one shot normally, exposed optimally, then take another shot with the IR longpass of your choice stacked with the UV-only filter,

use the same settings, same exposure time, this will show you if there is any reason to be concerned about a leak.

If you see a leak, in the stacked shot using the same exposure as the normal shot, then you can be concerned.

Other longer exposures only show us a more accentuated example of a leak, and where it leaks, and may be good for comparison to other UV-only filters at the same exposure time,

but it is not a real scenario, it is only good for an example.

 

In Hornblende's green light shots, these are normal exposures in daylight, everything is exposed normally and optimally, we don't need to 'push' the exposure to test for a leak.

If there is a leak, you should only need to take the same shots with longpass filters stacked on the Baader U using the same exposure time.

From my experience and testing of my Baader U, I see no leaks from it, unless 'pushed' using a longer exposure than would ever be used for normal lighting situations,

and then only in the 900nm range, which looks monochrome/gray/white compared to everything else, no color, so I rather doubt that the green is an out of band leak.

[Hornblende]

Have none of us gone out to take photos of stop lights and tail lights yet to test this our selves yet?

Ok, here is something strange: not every stop lights and tail lights are green. In this image (from wich came the tail lights crop I posted earlier) the stop lights were red when I took the picture, but they don't show any green. And some tail lights are green, some aren't..

Maybe it is a WB issue, I don't now..

Andrea, I should try using the Raw digger app.

 

and then only in the 900nm range, which looks monochrome/gray/white compared to everything else, no color, so I rather doubt that the green is an out of band leak.

 

Yes, I am pretty sure my green color can't be caused by an IR leak through the Baader-U. However, theoricaly, since the UV WB boosts the green channel relatively to the red and blue channel, it would makes sense that a "monochrome 900nm IR leak" appears green, right?

 

Bye the way here is another green thing: the sun's reflexion on these sunglasses. For this shot I used a Carl Zeiss Jena Tessar 50mm 2.8 which does not seems to transmit very deep in the UV spectrum (compared to my EL-Nikkor 80). White balance was done on my PTFE piece.

[Andrea B.]

Andrea, referring to your link, are you absolutely sure you have U-360 that is 1.75mm thick? Exactly where did you get that thickness?

 

In that post I have U-360 (2.00mm) + S8612 (1.75mm) stated in the text. Those are the correct thicknesses.

I seem to have mislabeled a couple of the photos! So I will add a comment to correct that.

Thanks for catching it. :)

[Andrea B.]

I believe that the 340BP10 and 325BP10 filters show a green-yellow hue in UV

 

Not with my Edmund 340/10. It white balances to the same old blue/yellow, although with more yellow than usual. No green (see link below). In the Raw Digger raw composite of the 340/10 photo, note that the photograph appears orange because most of the light has been recorded in the red and green channels.

LINK: Trying again with the 340/10

 

 

In the composite photo (linked below) the 4th flower from the left shows some green. That photo was made with a 4.00mm thick U-340 which was not given an IR blocker (usually S8612). Thus a bit of Red+IR leaked through to help produce a false green after a click-white. But once the Red+IR is blocked, the remaining flowers do not show green after click-white -- only blue and yellow.

LINK: [Filter Test] Looping around with U-340

[Cadmium]

Andrea, referring to your link, are you absolutely sure you have U-360 that is 1.75mm thick? Exactly where did you get that thickness?

 

In that post I have U-360 (2.00mm) + S8612 (1.75mm) stated in the text. Those are the correct thicknesses.

I seem to have mislabeled a couple of the photos! So I will add a comment to correct that.

Thanks for catching it. :)

 

OK, so the U-360 1.75 means 2mm?

[Cadmium]

I believe that the 340BP10 and 325BP10 filters show a green-yellow hue in UV

 

Not with my Edmund 340/10. It white balances to the same old blue/yellow, although with more yellow than usual. No green (see link below). In the Raw Digger raw composite of the 340/10 photo, note that the photograph appears orange because most of the light has been recorded in the red and green channels.

LINK: Trying again with the 340/10

 

 

In the composite photo (linked below) the 4th flower from the left shows some green. That photo was made with a 4.00mm thick U-340 which was not given an IR blocker (usually S8612). Thus a bit of Red+IR leaked through to help produce a false green after a click-white. But once the Red+IR is blocked, the remaining flowers do not show green after click-white -- only blue and yellow.

LINK: [Filter Test] Looping around with U-340

 

 

Andrea, he is not talking about shooting through a bandpass filter, he is talking about a diffraction grating. same idea s a Sparticle bandpass array, just not as delineated as a Sparticle array.

Shooting through a bandpass filter is a completely different thing, because you are white balancing within that narrow band.

In a Sparticle array scenario, you are white balancing with with PTFE or whatever overall sample you want, not with one narrow bandpass range/color.

 

It is true with every bandpass array I have seen anyone use, 340BP10 = green. Look at Colin's version, or Enrico's.

In an array, you are not white balancing with the green, or any other bandpass color, you are white balancing on the white of the full transmission range of the lens.

You are shooting through your UV-pass filter (Baader U... Moon U, or whatever UV- Pass filter/stack you choose).

 

You could try this with you U-filter on your lens, and your 340nm filter in a scene against the sky, maybe put some PTFE in there also, or white balance the entire frame.

That should make your 340nm filter green, if it is indeed a 340BP10.

 

With my Sparticle, I have one filter position that has a very thin piece of PTFE in it, that light comes through from behind, which I sometimes use for white balance.

Even thicker PTFE will show some light through it. Or whatever you like.

[colinbm]

Just a thought....1000nm leak is double the wave length of green, ie half of 1000nm = 500nm or green ?