Jump to content
UltravioletPhotography

[Filter Test] 293BP10 UV-Pass: Still No Luck


Andrea B.

Recommended Posts

I've tried the 293BP10 UV-pass filter more than once with no success, but thought I'd give it a trial on the converted Nikon D610. Still the same results, i.e., not much was recorded. This should come as no surprise. It is generally accepted that our converted cameras cannot record well around 300 nm even when making use of a UV-dedicated lens capable of recording down to 200 nm.

 

 

Equipment [Nikon D610-broadband + Nikon 105/4.5 UV-Nikkor]

 

Visible Reference [f/11 for 1/80" @ ISO-100 with Baader UV/IR-Cut Filter]

A jarful of grocery store Helianthus makes a beautiful targer for testing UV-pass filters.

helianthus_vis_sun_20170907wf_6477pn.jpg

 

Ultraviolet Narrowband [f/4.5 for 6" @ ISO-6400 with Omega 293bp15 (remainder)]

  • The lens was set wide-open and the ISO cranked up high so that the exposure could be as short as possible way down there in the region around 293 nm.
  • I do not know what causes the flare in the lower half of the 293 photo. The upper LCD on the D610 was blocked as it is a known leaker in long exposures. But this flare does not look like that produced by the upper LCD. Maybe this is from the filter holder somehow, but haven't found it yet. Maybe in the filter itself? Not sure.
  • There is some UV-signature recorded for this Helianthus under the 293bp15. But compared to the normal UV version (below), it is very hard to see unless drastic edits are applied.

293_610_2983pn.jpg

 

 

Ultraviolet Narrowband [Raw Composite of UV-pass photo]

The bandpass is only 15 nm, so we wouldn't expect much from the raw composite. However, a little contrast would have been nice. That isn't going to happen because the camera really cannot record well in the 293 nm region.

293_610_2983rawComp.jpg

 

Ultraviolet Narrowband [Raw Composite of UV-pass photo with some PN edits]

The contrast was cranked up very high and the white & black endpoints were adjusted in an attempt to extract some info from the very dull raw composite.

293_610_2983rawComppn.jpg

 

Ultraviolet Narrowband [Histogram of area on flower above flare]

Whatever light is getting through is recorded primarily in the red and green channels. The blue is mostly in the noisy, darker background.

293_610_2983histoFlower.jpg

 

 

Ultraviolet Reference [f/4.5 for 1/40" @ ISO-6400 with Uviroptics Luv U 2 UV-Pass Filter]

The Helianthus is beautiful under the LuvU2. This foto was also made wide-open at high ISO.

helianthus_uvLuvU2_sun_20170907wf_6441pn.jpg

Link to comment

I wonder how much of this is a result of the sensor limitation or the limitation of the Bayer filters.

A Bayer-less camera is not something many of us have, it seems, but they are available, right?

Given the depth of UV transmission with say a UV-Nikkor lens... it sure would be interesting to see this test done with a Bayer-less conversion as well.

Link to comment
Another suggested experiment: Test the 293bp15 for visible light leakage. For IR light leakage. I'm fairly sure there is no IR leakage, but I'm not sure about visible leakage.
Link to comment

It's possible to do a little better using denoising and an Independent Component Analysis method (which tries to recombine the channels to separate out distinct image components, such as removing the flare).

 

post-94-0-11674300-1505331219.jpg

Link to comment
Could you focus the image on any sort of phosphor screen and rephotograph it in visible? Alternatively, could you try b&w film as a sensing medium? If the sensor performs as poorly as is suggested, the digital image could consist largely of a record of out-of-band stray wavelengths, not in the filter's center band at all. It would be interesting to have something with which to compare it.
Link to comment

Could you focus the image on any sort of phosphor screen and rephotograph it in visible? Alternatively, could you try b&w film as a sensing medium? If the sensor performs as poorly as is suggested, the digital image could consist largely of a record of out-of-band stray wavelengths, not in the filter's center band at all. It would be interesting to have something with which to compare it.

Since that technique worked for me with the SWIR, it had occurred to me to try this already. It's likely to be even more effective in UVB than with SWIR because downconversion is more efficient than upconversion! Stay tuned, I do play to try the experiment sometime soon.

Link to comment

Steve, I had the filter taped to a cardboard donut shape (annulus) and placed inside this Edmund filter holder: https://www.edmundop...imaging-lenses/

 

What you can't see well in that link is that the filter holder has a inner retainer tube which unscrews. Look under the lettering on the filter holder to see the ridged ring on the inner tube which unscrews. (The link has a pdf download brochure which better illustrates the two parts of the filter holder.)

 

Anyway -- I'm pretty sure now that what happened was that the little 20mm diameter 293 filter either was not initially or did not stay well taped to the cardboard, and I got the light leak on one "side" of the inner annulus ring. Also I'm not particularly good at cutting cardboard circles to fit around little filters.

Oh well.

 

*****

 

Clark I don't have a phosphor screen or a BW film camera. I do agree that it is possible that I have recorded only stray oob wavelengths!

Link to comment

OK, I understand.

The link states, "Swivel feature on the back end of the holder allows for rotation of the optic while mounted."

I don't know what the 'swivel feature' is, or how it works, and I don't see that in the illustration, but it looks like the filter on the cardboard, as well as the entire ring could be turned 1/4 turn in the threads.

https://www.edmundoptics.com/document/download/396265

It would be more conclusive to get the filter taped up good with tape that will not leak UV, and do a quick shot, and if you see still see the 'leak', then turn it 1/4 turn (or even less) and see if that changes the position of the 'leak'.

Link to comment
Answering a question posted earlier about cameras with a Bayer filter removed down at 293nm. With my monochrome conversion the Bayer filter has been replaced with a Schott WG280 window, so even the transmission of that would be dropping at 293nm. Info from MaxMax who did my conversion was that the window was chosen as a the sensor has much reduced sensitivity in that region. Or in other words, choosing a window which transmitted much lower wavelengths would be pointless given the response of the sensor. When I tried the 308nmm filter from Invisible Vision, I did see something, but with a lot of IR contribution. The IR transmission of the filter should be low (according to a linear transmission graph), but obviously the effects of even very low IR transmission would be amplified if the UV being observed is such a low wavelength and the sensitivity of the sensor being so low there. Unfortunately I do not have a 293nm filter to try (at the moment), but I would suspect the UV sensitivity would be even lower than at 308nm, making the IR effects even more apparent.
Link to comment
I don't think what Andrea recorded was mostly IR, actually, because you can clearly see the characteristic UV bullseye pattern on the flower.
Link to comment

Some quick shots to illustrate what I was getting at. This was with my monochrome 5DSR (no Bayer filter), and Asahi UAT 85mm lens. ISO400, f11 used for all the shots, using sunflowers in the garden (direct sunlight at midday). These are as shot in the camera as jpegs, with no processing other than shrinking them for here.

 

First pair of images - Baader U based

Both photos 0.5s exposure. First image Baader U alone shows the obvious bullseye pattern.

post-148-0-85561600-1505476663.jpg

 

Second image, Baader U and 092 deep red (to block UV and let IR through), again 0.5s exposure. Nothing - so the IR is effectively blocked as we all know.

post-148-0-73941200-1505476738.jpg

 

Second pair of images - 308nm filter based

This is the 308nm filter I have mentioned before. Manufacturer claims >OD4 blocking, but I have no independent way of verifying that. Now, the exposure time has been increased to 8s to get an image in UV. First image, 308nm filter alone, you can still see the bullseye, but much lower contrast to the Baader U.

post-148-0-72566700-1505476867.jpg

 

Second image, 308nm filter plus 092 deep red, still at 8s.

post-148-0-11963100-1505476964.jpg

 

Now with the 092 filter and the 308nm filter there is a very obvious IR image, and the bullesye pattern has been lost, which says to me that the image with the 308nm filter alone has a big IR effect too. This goes back to my earlier point about the effects of IR being emphasized when taking UV photos at these shorter wavelengths and the need for longer exposures because of the reduced sensor sensitivity. So even while the bullseye pattern can be clearly seen there may still be a lot of IR effect present.

 

Yes, at some point I can compare the Baader U and 308nm at the same exposure times to see whether the amount of IR they let through is comparable, but in a bit of a rush today, and to be honest I forgot.

 

EDIT - Additional information

I've gone back out and looked at the Baader U and 308nm filters again, both in combination with the 092 red filter. As above the 308nm filter plus 092 at 8s exposure and ISO400 gave an appreciable IR image. With the Baader U and 092 filter, I had to increase ISO to 800 and exposure time to 30s, to get an image with similar exposure to the 308nm and 092 combination.

 

So the amount of IR getting through for Baader U and 092 at ISO800, and 30s, was similar to the 308nm filter plus 092 at ISO400 and 8s. By my reckoning that's about 3 stops better IR blocking for the Baader U compared to the 308nm filter. Or put another way about 8x better for the Baader U, or if I understand the concept correctly about an extra OD.

Link to comment
Yeah, but your 308nm bullseye was very faded, and Andrea's not nearly so much. Probably there is some contribution from IR in her case also, though. It would be interesting to see how the 293nm and the 308nm do with an artificial light source that isn't so IR-heavy!
Link to comment
The 092 is actually deep red (it can be used with non-IR film with about a 5-stop correction.) So it is possible that the mutual leakage is visible red rather than IR. This is something of a minor point, but one of which we perhaps should be aware.
Link to comment

UPDATE: Infrared Forcing Test on the 293BP10

 

I ran the 293bp10 through a cerium oxide cleaning to remove a crescent of oxidation and some spots of oxidation on the non-mirrored side. Then I tried shooting with it again by simply holding the filter by hand over the front lens of the UV-Nikkor. Kindly note that there is no way to keep a sharp focus when you are trying to hold a tiny 15mm filter to the front of a longish lens. :D

 

Equipment [Nikon D610-broadband + Nikon 105/4.5 UV-Nikkor]

 

Visible Reference

Photo Ninja conversion with white balance only.

610_3194pn01.jpg

 

Infrared Reference

Photo Ninja conversion with white balance only.

610_3196pn01.jpg

 

 

Ultraviolet 293BP10

Photo Ninja conversion with white balance only.

There are 16.3 stops difference between this UV shot and the corresponding Visible shot.

610_3198pn01.jpg

 

 

Ultraviolet 293BP10 Raw Composite

Recording is mostly in the Red and Green channels.

610_3198rawComp01.jpg

 

 

Infrared Forcing Test: RG9 over 293BP10

This is straight-out-of-camera. No conversion or white balance.

610_320501.jpg

 

 

Infrared Forcing Test: Previous Shot Auto-leveled

This is an extremely harsh push on the previous file. It is not at all obvious what might have leaked through the UV+IR stack. Whatever it was, there was not very much of it getting through.

 

In the middle you see a left half ring just where the 15mm filter would have been under the stacked RG9 filter. So there seems to be a bit of reflection revealed in the pushed file. Along with a mystery streak and an also mysterious brighter left half. The 293 filter could have slipped or perhaps tilted a bit under the RG9 which was supposed to be holding the 293 in place.

 

Note that the bisected white blob is the white house with the pole in front of it. It is not in the exact same position as in the original UV or IR photos because I had no way to see through the UV+IR stack in order to aim the lens accurately.

610_3205resizeAutoLevel.jpg

Link to comment

Jonathan and Andrea, very interesting tests.

Andrea, no so sure you should use cerium on a dichroic coated filter surface, but not knowing your filter exactly, I don't know...?

Link to comment
The pushed blob frankly looks like the reflection of the lens? I don't think anything much is getting through in this instance. Jonathan's situation was not the same it seems.
Link to comment
  • 4 weeks later...
enricosavazzi

One thing to check is whether the transmittance of this filter coincides with one of the "solar blind" absorption notches in the UV transmission of the atmosphere. In this case, if you try to use sunlight as a UV source you get little or no solar radiation at the bandpass wavelength interval of the filter, and you are forced to use an artificial UV source.

 

The graph at https://science.nasa.gov/science-news/science-at-nasa/images/sunbathing/sunspectrum does not seem to suggest that this is the case, but it is not a very detailed graph.

Link to comment
  • 6 months later...

@Andrea,

I think you own the super rare Pentax 85mm f4.5 right? If so you could try to get a used Sigma Sd14. The other nikon mount uv lenses will not work on the Sigma. But the M42 mount ones do. Yes focusing is a pain with no live view. But I think you will have better sensitive with the sigma. I recommend the Sd14 as it was the last not to use a metal substrate on the surface, so the uv response is predicted to be good. However, I don't know if any one has done a direct comparison with the newer Quattro. The Quattro does have live view and evf, but reported less contrast in IR.

Although I did see a new Canadian forum member here has the Quattro, so they may know best.

Maybe an option to test out this lower wavelength filter.

Link to comment

Well, having a Sigma Foveon to experiment with would certainly be enjoyable. "-)

 

The problem with imaging at a 293 nm peak is only partly with the camera. The other half of the problem is lack of UVB illumination in sunlight. (That 293 nm wavelength is in the UVB range.) UVB makes up only about 5% of the total amount of UV which reaches earth. So, given that UV in its entirety constitutes only about 3 - 10% of sunlight (depending on which website you read!), if you then take 5% of that 3-10%, I'd say you're getting very close to nuthin' !!!! I suspect good results from a filter like this 293BP15 will probably only happen if a UV-LED torch with a peak near 293nm is used?

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...