rfcurry Posted November 21, 2014 Share Posted November 21, 2014 Despite several years designing and modifying UV- bandpass filters, I have yet to answer the question: "What is the optimum Optical Density for unwanted wavelengths?" Most conventional UV-Bandpass filters, such as the BaaderU and the PrecisionU, set >3OD as acceptable. That is a compromise to allow maximum UV transmission. However, in many instances we experience a leak in the NIR even when the UV is adequate. Below are photos all taken with a Lumix GF1 modified for full spectrum, a Meyer-Optiks Primagon 35/4.5 wide-open, indoors with solar light through a window. The only post-processing was to reduce file size to 900 pixel width. #1 - Full spectrum, in-camera WB, ISO 200, 1/3200shttp://uvroptics.com/images/Full%20Spectrum%20WB%20900px%20961.jpg Now the same with a PrecisionU filter, ISO 800, 4.5, 10/13th shttp://uvroptics.com/images/PrecisionU%2010%2013ths%20900px%20964.jpg The next uses the same settings above, but with a Bower IR72 (near-IR pass) filter on top of the PrecisionU.http://uvroptics.com/images/PrecisionU%20NIR%2010%2013ths%20900px%20965.jpg The NIR is best seen as its negativehttp://uvroptics.com/images/PrecisionU%20NIR%2010%2013ths%20Neg%20900px%20965.jpg Obviously, the NIR is making a contribution to the UV image. Do we want a greater blocking to cover those conditions? What is the minimum OD in your opinion, if you could have a filter with high transmission and high blocking of unwanted wavelengths? Thanks. Link to comment
nfoto Posted November 21, 2014 Share Posted November 21, 2014 Was the window open? That affects the actual UV transmitted significantly. Link to comment
rfcurry Posted November 21, 2014 Author Share Posted November 21, 2014 No, it is cold here now, the window was shut. However, as you can see, there was plenty of UV, though less than one would want for a crisp photo with, for example, ISO 200 and f8.0. What would you want in terms of an OD? >4OD? >5OD?Would the scenario above be better if there were no leakage for 20 seconds? Would you trade transmission in the UV for blocking of NIR? Link to comment
nfoto Posted November 21, 2014 Share Posted November 21, 2014 I think we need to see the entire waveform of the filter not just the attenuation of contamination bands such as NIR and borderline red. The false colours of our UV captures are very sensitive to what happens around 400 nm say 390 to 420 or so. Too much going through here will enhance UV 'response' and if you do b/w UV cause little harm; however, for false-colour renditions they add colour casts that can be pretty tricky to recover from. Thus I'll gladly sacrifice response in this border region for better NIR suppression, even when it leads to a slight lowering of the peak transmittance in the deeper UV range. If one does UV video the situation might be different so having access to various specialised filters is beneficial. Link to comment
Andrea B. Posted November 22, 2014 Share Posted November 22, 2014 I want NO BLUE and attenuated violet as much as possible. Hard to say about the IR attenuation.>3OD has been fine so far. But if you made a >4OD filter, it would probably get some traction as being a "smidge" better. Link to comment
baffe Posted November 22, 2014 Share Posted November 22, 2014 Please help me: What is 1OD or 3OD ? Link to comment
nfoto Posted November 22, 2014 Share Posted November 22, 2014 OD = Optical density. Transmission is 10-OD, thus OD 3 == 10-3 or 0.1% transmission. Same as say a pH scale. Makes dealing with minute numbers easier for us humans I guess. Link to comment
baffe Posted November 22, 2014 Share Posted November 22, 2014 Thank you Bjørn! I'm used to that but didn't know that ODx refers to it. Link to comment
rfcurry Posted November 22, 2014 Author Share Posted November 22, 2014 Thanks for the excellent input!For the CopperU I have reduced the path length and resultant CA as requested by Enrico. I have decreased the blocking from 30s to five seconds for the same transmission of NIR as the PrecisionU and BaaderU leak in 2 seconds. See below: The PrecisionU in solar light through double-paned window. Lumix GF1, Primagon 35/4.5 @ 5.6, ISO 800, 2s ... No post-processing except cropped and width reduced to 900pxhttp://uvroptics.com/images/PrecisionU%202s%20crop%20900px%20008.jpg With the CopperU as specs above, except 1 second exposure, not two. [Obviously I could have used a shorter shutter time.]http://uvroptics.com/images/CopperU%201s%20crop%20900px%20010.jpg PrecisionU with Bower IR72 IR pass filter, 1 second exposure (this is the contribution NIR is making to the UV)http://uvroptics.com/images/PrecisionU%20NIR%202s%20crop%20900px%20009.jpg and the CopperU with Bower IR72. It took 5 seconds exposure to capture almost as much NIR as abovehttp://uvroptics.com/images/CopperU%20NIR%205s%20crop%20900px%20019.jpg The transmission of the CopperU will improve slightly when I get the quartz disks, UV-antireflection coated. And the visible light imagehttp://uvroptics.com/images/Vis%20crop%20900px%20024.jpg Thanks. Link to comment
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now