Jump to content
UltravioletPhotography

Baader U and Invisible Vision 308nm filter comparison for skin imaging


Recommended Posts

Ok, a very preliminary test (in the 5 mins between getting home and the sun going in). I recently got a 308nm band pass filter on loan from Invisible Vision in the UK (http://www.invisible...ts/accessories/). This filter is not aimed at skin imaging, but used for monitoring combustion processes, and the chap at Invisible Imaging said he'd never heard of anyone using it for skin imaging. Given I have the UV-Vis-IR Eos 5DSR, I did a comparison of this 308nm filter and the Baader U - just to see whether I could see anything with it, to be honest.

 

I started with a quick transmission measurement - both filters, on UV Vis at work. I ran them on a Perkin Elmer Lambda 650S UV-Vis spectrometer (150mm integrating sphere) to measure transmission between 250nm and 800nm (1s collection time per nm). The graphs are shown below.

post-148-0-15034600-1502893293.jpg

 

Quick photos of my lovely and long suffering wife. She was wearing some sunscreen on her face at the time, and the photos were done about 4:30PM in direct sunshine, facing the sun.

 

Firstly, with the Baader U. ISO 800, f8, 1/10th second exposure. Image has not been modified since coming from the camera as a JPEG.

post-148-0-31471100-1502894481.jpg

 

Focus not brilliant I accept, but I had limited time to optimise everything before the sun went in.

 

Second picture, with the Invisible Vision 308nm filter. Very different settings, ISO800, f8, 1.3 seconds. To be honest this was further tweaked in Photoshop, and I think the actual exposure to get this brightness would have been nearer 2 or perhaps even 3 seconds. So about 4 stops different to the Baader U.

post-148-0-06920700-1502894490.jpg

 

For imaging these lower down wavelengths I wasn't sure what to expect, although I thought I'd see similar to the Baader U image but perhaps with more contrast. However I'm wondering here whether I am seeing a product of very little sensor sensitivity at these low wavelengths, at which I am starting to pickup contamination from other wavelengths (the Invisible Vision filter is classified as OD4). Perhaps if I'm being optimistic the sunscreen looks darker through the 308nm filter than the Baader U, but I think it's far to early to be certain of that.

 

The camera sensor is covered by a WG280 Schott glass, so even the transmission of that is starting to tail off at these low wavelengths.

 

Has anyone done photography of people at these low wavelengths, or know what to expect?

Link to comment

I'm fascinated and wanting one of these 308 filters. I'm not surprised about the stop difference due to the lower UV transmission capability of the 308 and due to the smaller amount of that 308 peak wavelength range in the solar spectrum as compared to the 350-360 peak range of the BU.

 

But I'm wondering about IR contamination with that 308nm filter? Do you have any IR filters you could stack over it to check for that? Although I do see your mention of OD4 for the 308, so there is probably very little IR "leak". Still with longer exposures necessary for the 308, any IR leak becomes a concern.

 

I have had very little success photographing under 320 nm so far. A large part of that is due to my not having not owning really good filters in that range. And you need a lens which can transmit well between 300-330 nm. I think our camera sensors give up once you get below 300nm, give or take a small range, maybe 10 nm or so. And the illumination strength factor.

 

Going to go look up the UAT transmission range.

 

ADDED LATER: down to 200 nm.

Link to comment
I think the UAT should be fine down there. I'm going to do some more shots blocking the UV when I can, to see whether it's picking up visible and/or IR.
Link to comment

Yep, the UAT is a champ. Goes down to 200nm.

 

Try shooting at high noon on a cloudless day for more 308 ??? B)

 

Is the 308 shiny or mirrored on any side?

Link to comment
John, I'd rather not, as I have an issue with the baseline of the device. Even after calibration, I get values for transmission of about -0.1% in parts which should be 0, which makes me very nervous about trying to do this as a log scale and giving misleading data. I'm comfortable with the overall shapes of the curves to within about 0.1% based on other work I've done here, but for now would rather keep these as linear plots, so am sharing this in the spirit of complete transparency (as it were).
Link to comment

Spirit of transparency, that's a good one! :lol: I do however, appreciate not sharing data in which one lacks confidence.

 

Even with the sphere attachment I would have thought your double beam double monochromator would have a dynamic range of nearly 4 OD. My old PE Lambda 6 could stretch beyond 3 OD with a sphere if I opened up the bandwidth some.

 

Something you might consider doing to empirically evaluate the dynamic range is to run a transmittance calibration and then scan an empty (100%) path and then a totally blocked (zero%) beam. Plot both on a log transmittance scale. The noise in the zero% scan will sometimes not be smooth and may yield negative values depending on how the software processes the dark current.

Link to comment
Yeah I'm surprised by it too John, will try those when I get chance, so thanks for the advice. One thing that isn't ideal at that the temperature in the room where the device is does fluctuate by +/- 2C, seemingly at random throughout the day, which I'm guessing isn't ideal for something like this. Also the device is normally off, so when I need to use it it doesn't get much warm up time. Ideally it would be left on for longer, but work doesn't like that.
Link to comment

Andrea, the 308nm is shiny on the side which faces the lens, not the side what faces the subject.

 

Right I have some follow up images (no beautiful model this time, just part of my garden). I took a series of images with the 308nm, 308nm plus 092 deep red, 308nm plus the B_W 486 UV/IR blocker. These were then repeated with the Baader U. All images with Monochrome UV/Vis/IR EOS 5DSR, and Asahi UAT lens. ISO800 and f8 for all images. Exposure times 2s for the 308nm filter, and 1/13th second for the Baader U.

 

First how do all these filters looked for transmission when plotted together.

post-148-0-09370500-1502979407.jpg

 

So, the B+W 486 does overlap slightly with the Baader U, but absolutely no overlap with the 308nm.

 

Results 1 - Deep red 092 filter

So, 2 pairs of images. First the 308nm filter alone and with the 092 filter.

post-148-0-65042900-1502979499.jpgpost-148-0-00533400-1502979506.jpg

 

Second the Baader U, alone and with the 092 filter.

post-148-0-43508600-1502979566.jpgpost-148-0-13037700-1502979575.jpg

 

As expected the Baader U image is showing no IR effect at the exposure time used here (1/13th second). However the 308nm filter shows a significant image when the 092 filter is present (keep in mind this is 2s exposure now). To me that says there is a significant contribution of the IR to the overall image prodcued with my camera/lens combination. I'm presuming this is due to the much higher sensitivity of the sensor to IR compared to the 308nm region. Obviously the exposure times are different, but this is a real world assessment at the times needed to make an image with the respective filters.

 

Results 2 - B+W 486 UV/IR cut

The aim here was to see whether visible light was contributing to the 308nm filter image. Again, 2 pairs of images. First the 308nm filter alone and with the 486 filter.

post-148-0-14486400-1502980177.jpgpost-148-0-33393000-1502980213.jpg

 

Second the Baader U, alone and with the 486 filter.

post-148-0-58344700-1502980258.jpgpost-148-0-35988000-1502980265.jpg

 

The 308nm shows no evidence of visible light contamination. The Baader U does show a slight image in combination with the B+W486 filter. There is an overlap in the transmision curves for the Baader U and B+W 486, so I presume that is the reason for the dark but visible image in this case.

 

Overall

Looks to me that there is a significant IR contribution to the image from the 308nm filter using my setup. I'm presuming this is due to the relative sensitivity of the 308nm and IR regions of the sensor (much much higher sensitivity for the IR compared with 300-320nm UV). For my system tested here, I would need more IR suppression before I could call this a true '308nm' UV image.

 

The Baader U did not exhibit this behaviour with IR at the exposure needed to create an image with it. However there was a dark image formed from using the Baader U from when the UV/IR block filter 486 was was used, most likely due to the overlap of the transmission characteristics around 400nm.

Link to comment

Well, that was interesting. Now let's think what to do.

  • If a UV-pass filter leaks some red or IR, then that does not necessarily preclude its use as a UV-pass filter. However, the filter must be tested against some kind of known target to determine how much the IR leak is "washing out" the UV. As always, we suggest looking for a flower which has a known light/dark UV-signature such as a Dandelion (Taraxacum), Rudbeckia or Sunflower (Helianthus). Something like a color checker passport is also useful because that does exhibit a light/dark UV-signature in its patches. Then you shoot a series at a fixed aperture and different times. I'll remind everyone that the much used BaaderU can leak some IR under strong illumination and long exposure. But in the typical UV-flash lit short-exposure photos it is not observable. (There are some BU tests here in the Filter section somewhere.)

  • Shiny-sided 308 filter: If you use a lens hood to cut some of the light entering the lens off-axis, then the chance of inner reflections are lessened. It might be a good idea to try to rule out any inner reflectivity problems.

  • B+W 092 Red+IR-pass filter: It would be nice to rule out whether there is "high red" leak ( 700 nm). Here we remind ourselves that there is not a clear demarcation between what is considered high red and low IR. Anyway, testing this would require some IR-longpass filters at 715, 730, etc.

  • Blocking IR in a 308: Whewie! This is a tough one. Either S8612 or any BG glass is just going to kill the 308 transmission and move the peak. No good.

  • What does Invisible Glass have to offer about the IR leakage? Like, where and how much. Also could a 308 be given a dichroic coating to block IR by Invisible Glass or by some third party optical company?

Link to comment
Looks to me that there is a significant IR contribution to the image from the 308nm filter using my setup. I'm presuming this is due to the relative sensitivity of the 308nm and IR regions of the sensor (much much higher sensitivity for the IR compared with 300-320nm UV). For my system tested here, I would need more IR suppression before I could call this a true '308nm' UV image.

Jonathan, don't forget that a big part of the problem is that sunshine contains very little UVB relative to UVA+vis+IR also! That is the third issue. If you change the light source, that would probably help.

Link to comment

All good questions and things to think about Andrea. It must be kept in mind that Invisible Vision don't aim their products at SLR users for general UV photography, so my first thought would be that maybe the industrial cameras their customers are using are more UV sensitive and less IR sensitive, or the light they're imaging has less IR component. As such the filter may be adequate for the purpose it was designed for. They were good enough to loan me one to try out, so I don't want to go rubbishing a product that is being asked to something that may be well outside the original design brief. I hope to follow up with them next week if possible.

 

The filter is thick, and I seem to remember when chatting to the chap there, he said it was 2 filters sandwiched together. So perhaps there is scope for modifying the construction to block more IR.

 

One thing I didn't do today was look at the IR blocking of the Baader U and 308nm filter, at the same exposure time. If the Baader U and 308nm filter have similar IR blocking capability then further improvements will be hard. If it's not, then perhaps further improvements could be made. I'll try and do that if we get any more sun :)

Link to comment

 

Jonathan, don't forget that a big part of the problem is that sunshine contains very little UVB relative to UVA+vis+IR also! That is the third issue. If you change the light source, that would probably help.

Absolutely Andy, it is a very big ask for the camera to create an image at these wavelengths in sunlight. However I would have big concerns about intentionally irradiating my lovely wife with high intensity UVB to get her photo. Perhaps a Psoriasis lamp which operates at 311nm would be good for other, non living subjects though.

Link to comment
Oh definitely not for skin! Is that your only interest, though? I would love to see what the flowers and so forth do!
Link to comment
Hi Andy, yes that's my primary interest. I work in skin research, mainly ways of measuring and visualising it, so for me it'll always come back to 'how can I safely use XYZ for looking at skin'.
Link to comment

Ok, results of a few more experiments. Same camera and setup (Monochrome EOS 5DSR, UAT 85mm lens, ISO800, f8 for all shots, but now 2s for everything). The aim here was firstly to look at the 308nm filter and Baader U, under the same exposure times, with and without the B+W092 filter to directly compare how much IR they block. Secondly, to turn the 308nm filter round and see whether having the shiny gold/pink side facing the subject made a difference compared to the black side (which is the 'as supplied' setup).

 

Sunny day 11AM, no clouds during the experiment.

 

Experiment 1 - 308nm and Baader U IR blocking comparison

2 pairs of images for each filter - filter alone and in combination with the B+W092. Firstly the 308nm one.

post-148-0-12348100-1503051100.jpgpost-148-0-10235800-1503051107.jpg

 

Secondly the Baader U - alone and with the B+W092 filter.

post-148-0-13113300-1503051113.jpgpost-148-0-58108100-1503051118.jpg

 

As expected the Baader U image is much brighter under the same 2s exposure as the 308nm filter. However it also blocks more IR. There is in image in the Baader U plus 092 shot, but it is very dark and may not show in the online version. So fairly confident saying that the Baader U blocks more IR than the 308nm filter in this test.

 

Experiment 2 - turning the 308nm filter round (metallic side facing the subject)

The 308nm filter as supplied had the black side facing the subject and the shiny, metallic (pink/gold) side facing the camera. Swapping this wound I repeated the filter alone, and in combination with the 092 images.

post-148-0-24393800-1503051126.jpgpost-148-0-61986100-1503051131.jpg

 

Turning the 308nm filter round improved the contrast, but to my eyes did not lessen the IR transmission.

 

Overall

Looks like me as though the IR transmission of the 308nm filter would have scope to be reduced, as it is higher than the Baader U. Whether this is feasible or not remains to be seen. Plus it must be kept in mind that the demands on the filter here may be very different to its intended usage, so this may not be a problem for what it is designed for.

 

Also, in this experiment contrast was improved by having the metallic side facing the subject, although IR transmission did not appear to be significantly impacted.

Link to comment
I think dichroic filters are intended to be used with metallic side facing subject! Sometimes they will have a small arrow on the side of the filter indicating which way they are meant to go, but I think that's the typical way.
Link to comment

I think dichroic filters are intended to be used with metallic side facing subject! Sometimes they will have a small arrow on the side of the filter indicating which way they are meant to go, but I think that's the typical way.

Yep that's what I thought and why I tried turning it around, and it looks to have improved the contrast.

Link to comment

The problem with shiny-sided filters is that they can induce reflections and flare within the lens barrel and/or camera chamber from any entering light, but most particularly from off-axis light. The reflections can cause a general lack of contrast or create interesting annular patterns in the photo as well as the typical aperture-shaped light blobs on a diagonal. Informal observations indicate that transmission of the desired wavelengths is not affected by which direction the shiny side is facing.

 

Ideally with such shiny, the photographer remembers not to shoot contra jour. Turning the shiny filter side outward to face the photographic subject and using a lens hood help mitigate potential reflection and flare problems. With a filter like the BaaderU, just as one example, 98 times out of 100, you won't see any problem from its shiny side. But I have a very expensive hard-coated Edmund Optics narrowband 340/10 UV-pass filter which is unusable due to its two mirrored surfaces. (Not Edmund's fault in any way. I knew what I was getting into!)

 

As UVP member OlDinyo reminded us recently in another post, a dichroic filter also can create some artifact problems due to the angle of shooting. Dichroic filters can induce ringing and also false color changes near the edges of a photo. Again, using a lens hood, repositioning of subject or photographer as needed and shooting "in the middle" of the frame perpendicular to the subject mitigate dichroic artificacts. Wavelength transmission is slightly affected by the angle at which dichroic filters are aimed at the subject but I've never found it to be noticeable in my BaaderU botanic photography.

 

UVP member Andy Perrin reminds us that rear mounting a small version of a dichroic or shiny-sided filter can help mitigate possible flare and reflections from off-axis light.

Link to comment
One can also rear-mount the filter. I think the rays are more parallel as they leave the back of the lens, so the angular artifacts are reduced. Also you can use a smaller (and therefore cheaper) filter.
Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...