Jump to content
UltravioletPhotography

UV, Visible and IR filters test using Monochrome camera


Recommended Posts

I was recently at NASA Kennedy Space Center, and took my monochrome converted EOS 5dsr with my along with a load of filters. My aim was to get a set of photos with different filters, under good lighting, to get a feel for a couple of things. Firstly, typical exposure times for UV vs Visible vs IR under sunlight. Secondly to try out a set of filters I recently got from Cadmium (U340 in 2mm and 4mm, and S8612 in 1mm) and see how they compared to the Baader U.

 

I'm sharing this as a pdf document, in case it's of interest. There are a lot of pictures in there, but it does give the idea of relative exposures with the different filter setups.

 

Visible UV and IR filter test Nov 2017.pdf

 

The images I took were with the following filters;

No filter at all (UV, visible and IR).

Baader UV/IR cut

Baader U

U340 2mm

U340 4mm

U340 6mm

U340 2mm + S8612 1mm

U340 4mm + S8612 1mm

U340 6mm + S8612 1mm

Heliopan 715

Heliopan 780

 

Camera details are in the word document, but everything was done with my monochrome EOS 5dsr and Asahi UAT 85mm f4.5 lens (ISO 400, f8).

 

Some of the pictures were not without issues unfortunately. I had some guys who were in the bottom left of my image who came as I was about half way through my shots, so some of them have arms and or hat/head in them. After about 5mins of them getting more and more in the way I had to politely ask them if I could just finish my photos. Thankfully they were happy to do so (well, they didn't grumble too much). Given I wanted to get everything done as quickly as possible I didn't have time to redo the shots they were in.

 

EDIT 18th Nov, added in optimum exposures for each set of images to the PDF. These are based on aesthetic and appearance rather than anything technical.

Link to comment

Jonathan, thanks for the test results!

 

Do you have some way of indicating what you think is the best exposure for each filter? The photos are small so I was not easily able to decide which exposure in each row might be considered optimal for comparison purposes. (Of course, some are clearly not optimal.) Like, maybe you could mark the optimal versions with an X or a "best" or something?

 

It does appear that the monochrome conversion is faster in UV, but I would need to go look at some similar photos made with a Bayered conversion. Do you think it is faster?

 

May I ask -- who performed the monochrome conversion? We like to keep track of who is (or is not) making good conversions. Are you happy with this monochrome conversion?

 

 

Something you might want to consider in the future is using a Solar meter to get a measurement of the amount of UV-A for your current shooting conditions? It is surprising just how quickly the amount of UV-A in sunlight can drop, or (sometimes) increase depending on time-of-day, especially in late afternoon during winter.

 

I too have had people and animals and boats wander through my test scenes. :rolleyes:

 


 

Added a moment later:

It occurs to me that we do not have any working definition of what constitutes a "good exposure" in UV. In any photo file, I usually want to see that the raw histogram has not piled up on either the dark or the bright end. But that is difficult to achieve in UV because our cameras do not always have the exposure range ("dynamic range") to manage both ends of the raw UV histogram well. So I usually go with not having a pile up on the left histogram side to prevent solid black shadows in in a UV photo.

 

Of course, the particular UV subject can and does change what the definition of good UV exposure is. Might be that it is impossible to define "good UV exposure", but don't we usually know them when we see them? :lol:

 

I've often thought that a good UV exposure would require a combination of one shot optimized for the brights and one optimized for the darks - that is, some kind of mini-hdr effort.

Link to comment
Thanks for these! Do you have the transmission spectrum for that lens you used here? Cadmium has pointed out that the lens is often our limiting factor, so it would be nice to know how this lens compares to ones like the EL Nikkor 80mm/5.6 that I have transmission curves for.
Link to comment

The Asahi Ultra-Achromatic-Takumar 85/4.5 is achromatically corrected for UV/Vis/IR between 220-1000 nm. I think it's more like 250-950 because at the edges there things are a little off, IIRC.

 

It transmits at about 80% 75% from just past 220nm.

EDIT: I think I got that too high. It is probably more like 70+%. So I'll put in 75% until someone comes up with a chart. B)

 

(Can't find my chart? Thought I had something somewhere.)

Link to comment
Wow, that is a fancy lens! So it is not the limiting factor here then. Sounds like the sunshine spectrum, sensor, and the filters will limit it, mainly?
Link to comment

Thanks for the comments.

 

Andrea. I've redone the PDF file and included what I think of as optimum exposure for each of the sets of the images. Nothing technical here, just based on what I thought looked right.

 

It is faster with the Bayer filter removed. Compared to my ACS converted cameras which have UV filters over sensors, but retain the Bayer filter, the monochrome camera is about 2 stops faster in the UV. Of course this isn't quite an 'apples for apples' comparison, as I'm using the Baader U on the monochrome camera, and the ACS cameras have their own filters over the sensor, but it is a real world test. If I had a multispectral EOS 5dsr with the Bayer filter retained that would be ideal, but that may have to wait for 2018 to do - they're not cheap cameras to play around with conversions with. Mine was done by Dan at MaxMax. Overall I'm very happy with it. With my work it was vital to get as much UV sensitivity as possible and this gave me a couple of stops advantage. Dan claims x6 improvement in UV sensitivity (about 2.5 stops), and I'm seeing 2 stops so the claim seems pretty accurate. At NASA I tried to get these done as quickly as possible as to minimise changes in lighting. Overall, it took me 15 mins to get all these pictures. I do have a USB spectrometer, but it would have meant taking laptop and spectrometer with me to the US as well, and given I was photographing something a long way away wouldn't have told me a lot as lighting could easily have been different where I was compared to the subject. So I relied on the mk1 eyeball, to tell me whether the building in the distance was still in direct sunlight for all the shots.

 

Andy. As Andrea mentioned the lens transmits much lower than the camera is seeing here, so I am assuming that the transmission of the lens is pretty constant from UV up to IR. I do not have a transmission chart to verify that however (again, another job I hope to be able to attack in 2018). So yes, this should come down to sunshine, the filters themselves and the sensor. The other thing to consider though is the glass in front of the sensor. When the Bayer filter was removed, it was replaced with Schott WG280 by MaxMax. The choice of that did surprise me slightly, but their reason was that the sensors are not really picking anything up where this cuts off. It does mean though that the shortest wavelengths of the U340 are being severely attenuated before reaching the sensor (obviously with the S8612 used in combination with the U340 this would also be attenuating the shorter wavelengths).

Link to comment
Yeah, I guess the sensor glass would matter, although sunshine has so little UVB that I wonder if it makes much difference to the overall exposure here. Obviously you could try to force the issue with a UVB-heavy light source (and I know I would be tempted with that setup!). Taking due care not to expose skin/eyes obviously.
Link to comment
Absolutely, Andy. I suppose that's why it's difficult to look at filter properties in isolation when thinking about final effect on the image. We need to consider the wavelengths in the light source, the sensitivity of the sensor, and the transmission of the filter and lens. Nice complex area :)
Link to comment

the monochrome camera is about 2 stops faster in the UV

 

That's a good healthy jump. We always need all the UV help we can get for sensitivity and light. :D

 

a real world test

 

The best kind of test, methinks. The only place most of us can use our gear is in the real world. B) For your UV/sunscreen/dermatology studies you may be able to set up more rigorous lab-like conditions. But knowing how the gear works out in the sunlight, you will already have a good feel for how it will perform during experiments.

 

*****

 

Sensor sensitivity is good but it also helps to have a sensor that can capture a wide dynamic range for UV work.

 

*****

 

so I am assuming that the transmission of the lens is pretty constant from UV up to IR.

 

Yes. Just look up the transmission charts for quartz and fluorite glass and go from there. The UAT has 5 elements. I dug out a diagram for it, which I'll go retrieve and post here.

 

From the website of Marco Cavina: http://www.marcocavi...etri/00_pag.htm

The diagram was originally made by Pentax. Marco has relabeled it.

UAT_Elements.gif

Link to comment

Jonathan, Very cool! :-)

 

It would be interesting is to have some ###PB10 filters with lower peaks than 325nm, like with my Sparticle bandpass array, then you can see what the UV transmission range is between the Bayer and Non-Bayer sensor,

if you also have a copy of the same camera that is full spectrum converted with Bayer.

I agree (with MaxMax) that anything 'better' than WG280 will not change anything given the limitations of the sensor.

I think Spectrosil is feature fluff for our use. I think WG280 is the best choice for a full spectrum sensor.

Such a Sparticle with a wider BP range would not show colors of course, however it would show range depth difference.

 

Update with U-340 graph in a few...

Link to comment
Alex thought the Spectrosil was really soft, also. Thankfully I haven't had any problems in that regard yet (assuming my conversion really used Spectrosil...I don't trust that guy anymore!).
Link to comment
It is only my opinion that Spectrosil for our use is mostly a selling point. I don't see anything wrong with it, and I could be wrong, but I just don't think the sensor is going to record anything below 280nmm, however, with a Bayer-less sensor, it might be worth comparing WG280 with Spectrosil to know if there is any difference. I don't know who has both glass types available for conversions. Then you would need 3 conversions, or would you need four?! ;)
Link to comment

Spectrosil = fused quartz or is it fused silica? Fused silica.

Transmits around 90% from 200 - 4000 nm, give or take.

 

Schott WG 280 is 90% from 340 onward. But not to worry, it is already 80% at 300 nm.

I'm looking for its Moh. Should be tougher than Spectrosil?

 

Cannot find any accurate hardness data. Sorry.

Found in Wikip: fused silica 5.3-6.5 Mohs

 

hmmmm......fused silica is non-crystalline while fused quartz is crystalline silica.

Fused quartz is transparent. Fused silica is translucent.

Fused quartz is prone to phosphorescence.

Spectrosil is very pure fused silica so is "transparent deeper into the uv". (whatever that means.)

 

fused quartz vs. fused silica what is the difference?

Link to comment

What thickness of Spectrosil is your data for?

What thickness do they use in cameras for Spectrosil and/or WG280?

I have never seen any demonstration comparing the two types of glass for UVIR cameras. Certainly Spectrosil transmits lower.

 

Here is 1mm thick WG280 in T diabatic. For what it's worth.

post-87-0-49208500-1511061354.jpg

Link to comment

Here are some comparison graphs of the U-340, 2mm, 4mm, 6mm, and even 8mm,

and also stacked with S8612 1mm.

Note how the S8612 stacks move the UV peak and upper width lower as the U-340 is thicker. This is true of any U-glass stacks using a same S8612 (suppression glass) thickness.

On the other hand increasing the U-340 alone (when not stacked) does not change the stack peak. Sorry, that is off topic, but something I always like to point out about U-glass stacking... how you can move the peak with thickness when stacked.

 

post-87-0-13577700-1511061727.jpg

 

post-87-0-11750200-1511061748.jpg

Link to comment
I'm glad I posted this now, as it's prompted an interesting discussion. I would love to know whether the lack of sensitivity at and below 300nm is driven by the Bayer filter being highly absorbing of light, or whether the sensor itself has simply run out of sensitivity. If I ever get another one made, I might ask for something other than WG280 just to check. When I borrowed the 308nm filter from Invisible Vision, I did indeed see a UV image, but with a lot of IR bleed, likely as I had to lengthen the exposure so much just to see anything. At some point I'll get round to stacking 2 of those 308nm filters together to see what happens.
Link to comment

Yes that would be very interesting to know. I would assume that the Bayer plays at least some role in UV depth sensitivity, but you would probably need two camera conversions (of the same camera model) to compare,

and those two should be using the same type of glass. Then if you wanted to compare the two types of glass... well, you see what I mean... more conversions.

Of course you could always sell the cameras after you compare them, but...

Link to comment

hmmmm......fused silica is non-crystalline while fused quartz is crystalline silica.

 

 

Both, "fused silica" and "fused quartz" are vitreous, you could name them both vitreous "SiO2 "

They are made ("fused") from different starting material (both very clean, but quartz a little bit less).

The processing seems to be different as well, the fused silicon process seems to be done in a different atmosphere and also the melting might be done in a tank (a continuous process) or in a pot/crucible (a batch process). (The "fused silica" might be based on a CVD process (chemical vapour disposition, which needs a different atmosphere than e.g. heating with a oxidizing flame in a crucible).

 

The higher the purity of the SiO2 is, the better is the tetrahedral structure of the vitreous SiO2, nearly all O-atoms have two bondings to Si-atoms, there are only few "non bridging oxygen" atoms (NBO).

The impurities deform the structure of the SiO2 teraheder and cause even more NBO (some ppm of other oxides can do).

The NBO can be stimulated by UV-light and thus the UV is "lost" in the glass.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...