lukaszgryglicki Posted August 23, 2021 Share Posted August 23, 2021 Hi, my name is Łukasz Gryglicki. I'm from Poland (Minsk Mazowiecki), 39 yo. I'm interested in multispectral (UV in particular) photography for years. I have a Nikon D600 camera converted to full-spectrum black & white - Bayer filter and anti-aliasing removed. This makes this camera 2 stops more sensitive than a normal camera.I also have Nikon D3200 (APS-C) full spectrum converted. Recently ordered another D600 with just full-spectrum conversion (so I can play with unreal colors). I was mainly working on the IR part of the spectrum, so I have a lot of IR filters like 720 760 850 950...I also have a bunch of old Nikon film cameras FA, F4 plus one medium format one - Pentax 67. Additionally, I have Nikon D3, Df for normal shooting. I've collected all this gear over about 10 years.I'm not a photographer at all - just a hobby, I work for the Linux Foundation/CNCF (remotely): https://www.cncf.io/people/staff/ (search for "Lukasz Gryglicki"), FB profile: https://pl-pl.facebook.com/lukasz.gryglicki. After my last trip to Spitsbergen with Nikon D3200 full-spectrum, I decided to go serious in UV. 9 years back when I was on a polar expedition (Polish Polar Station 34th expedition - wintering) I had point and shoot camera converted to full spectrum at the end of the expedition - this was my first time trying anything other than visible light - conversion was awful and the camera was unable to focus at infinity, but even because of that and the uncorrectable pink hue (white balance setting was impossible - full spectrum was just too much for that particular model) - I knew that I'll go that way. Recently I decided to give UV a try, I have Hoya U-340 55mm and tried it on a few lenses like (all Nikkors): 20/3.5, 55/1.2, 50/1.4 (the best so far, but still ugly). I have a lot of other Nikon-F lenses, but they are "too complex" for UV (multicoating, many elements like 135/2 ai-s, 14/2.8 AF-D, Irix 11/4, 85/1.4, 28/1.4, Macro 105/2.8, AF, etc.)So I did some extra steps, bought (none of them received yet): - "Kolari Vision UV Photography Filter" 58mm - I'm not related to Kolari at all.- Nikkor 50/1.8 AF-D (reported in many sites as a good UV performer - will see).- Hoya IR/UV cut 77mm.- UV8040BK2 + T-NI lens (especially interested in this one https://www.universeoptics.com/product/uv8040bk2-ultraviolet-quartz-lens-assemblies/, reported as transmitting from 200nm and my sensor has no Bayer filter, no AA filter, and no IR/UV filter - it's a monochrome CMOS without anything after it). Not technically ordered yet, because I asked them some customs-related questions and waiting for feedback. Link to comment
lukaszgryglicki Posted August 23, 2021 Author Share Posted August 23, 2021 One example full-spectrum photo from my recent trip (Sveabreen, Svalbard): Link to comment
colinbm Posted August 23, 2021 Share Posted August 23, 2021 Welcome Łukasz,Looking forwards to seeing more of your photos. Link to comment
lukaszgryglicki Posted August 23, 2021 Author Share Posted August 23, 2021 Hi, thanks, I'll share more when I have my gear completed. BTW: forgot to mention that I also have an IR camera 8-12 microns (FLIR E70) but these are totally different wavelengths... 25 times longer than UVI've actually written a tool that is able to get RAW frames from FLIR E75 camera and convert them to H264 video file, applying many possible filters/algorithms on the way, effects are really interesting IMHO.Water is black in opaque IR (sorry for the quality but this IR camera is only 320x240): - - - Germanium metal is transparent to thermal IR:- - Glass, Ice, drinks, etc:- Swimming pool:- Car, machinery:- Link to comment
Andy Perrin Posted August 23, 2021 Share Posted August 23, 2021 How are you getting the RAW video frames from the camera? I know how to obtain RAW still frames, but for video the only method I have is to digitize the analog output from my E60. Link to comment
lukaszgryglicki Posted August 23, 2021 Author Share Posted August 23, 2021 There is an option to save radiometric data - CSQ files. They basically contain raw data in 3-dimensional format: width x height x time (30 frames/second).Their format was reverse engineered so e time ago, I'm just reading those files and generate videos myself - adding any kind of processing in the meantime (even user applied functions typed as strings). I'm a programmer, all details are here: https://github.com/lukaszgryglicki/csqconvProcessing CSQ frames happens here: https://github.com/lukaszgryglicki/csqconv/blob/master/csqconv.go#L206-L273 Link to comment
lukaszgryglicki Posted August 23, 2021 Author Share Posted August 23, 2021 Say 1 minute CSQ file contain 320(width) x 240(height) x 30(fps) x 60(s) frames - each is a 16-bit 0-65535 signal level from the sensor.And BTW: no idea why YT is lowering the quality. I'm saving files as MOV with no compression artifacts like those seen on YT. Files are low res 320x240 but then crystal clear. Link to comment
Andy Perrin Posted August 23, 2021 Share Posted August 23, 2021 I know how to get radiometric JPG off the camera but I have never seen CSQ files. My FLIR E60 is rather old (bought used so I am not sure how old…) so maybe it does not use this format because it hadn’t been introduced yet? Anyway, I’m a programmer also, but I mostly work in MATLAB because of their excellent image processing toolboxes and the many researchers who post their code in that language. Link to comment
lukaszgryglicki Posted August 23, 2021 Author Share Posted August 23, 2021 I think it does, search in settings for video options - you can have it record normal video - but it will contain all other info from the screen and you can't get "real" data then - just compresses video stream. You should alos have an optiin to choose "radiometric data *.csq files" - those files can be processed by my program - you can apply a lot of own functions while doing so - I've written it to allow myself to display this data in a very different way than just a 6 or 8 possible color modes that come from the camera. See https://www.thermalimaging.co.uk/flir-e60/ - it says E60 has "Radiometric IR-video streaming" Link to comment
Andy Perrin Posted August 23, 2021 Share Posted August 23, 2021 The camera called the E60 has gone through many revisions by FLIR over time. The most recent specs are NOT the specs for older versions. Hardware and firmware have both changed with time. I have no MSX for example either. Link to comment
lukaszgryglicki Posted August 23, 2021 Author Share Posted August 23, 2021 OK, so doesn't have CSQ option?It is in"settings -> video compression -> radiometric data (*.csq files) my E75. Link to comment
Andy Perrin Posted August 23, 2021 Share Posted August 23, 2021 I do not have a video compression setting. Link to comment
lukaszgryglicki Posted October 17, 2021 Author Share Posted October 17, 2021 Not sure where to put some photos, decided to put them here. Some first shots with EL Nikkor 80/5.6 (afternoon 3-5 PM, autumn Oct 16th, overcast skies - this was my first day when I was able to assemble all of this). Nikon D600 full spectrum (no debarring) + EL Nikkor 80/5.6 (old metal version) + Kolari UV-Pass (mostly shoot at f=5.6, f=8) - all handheld, using live view (which was barely able to allow focusing) - f=5.6 is so dark in UV, mostly compensated by high ISO. EXIF will always say 80/5.6 because the operation is 100% manual, camera "thinks" there is no lens attached. Camera wasn't able to white-balance (Nikon) so it was post-processed and I see light leak - when there is a lot light just above the photographed frame - it leaks - when I use a hand "as a shade" just above then lens, light leak is gone light also passes through the eyepiece but this can be covered by the thumb). And here autobalance from JPG's instead of NEFs: Link to comment
lukaszgryglicki Posted October 17, 2021 Author Share Posted October 17, 2021 And here Hoya U-340 4mm thick - also overcast day (even worse than yesterday), so shutter times were 1/5s or so (handheld) and ISO=6400, aperture was always 5.6. Link to comment
ulf Posted October 17, 2021 Share Posted October 17, 2021 These last images are clearly an IR-contaminate. U-340 4mm has an OD of between 2 and 3. From an artistic point of view they are still nice, but not pure UV-images Link to comment
lukaszgryglicki Posted October 17, 2021 Author Share Posted October 17, 2021 Yep, I know, disussed this many times, my S8612 2mm didn't arrived yet. BTW: the real last one is pure FS (no filters at all) Nikkor Ai 300/2.8... (indoor, no flash, 5 meters away) Link to comment
lukaszgryglicki Posted October 18, 2021 Author Share Posted October 18, 2021 Now a few "landscape" UV photos (almost all of them). Filters used: both Hoya U-340 4mm and Kolari UV-Pass (actually to surpass IR). This needed times like 1/15s, ISO=6400 etc. Skies were almost clear, but in UV they're white - this is expected IMHO. Link to comment
Andy Perrin Posted October 20, 2021 Share Posted October 20, 2021 Yes, light colored skies in UV. Why so red? Was this an artistic intent? Link to comment
lukaszgryglicki Posted October 20, 2021 Author Share Posted October 20, 2021 White balance stuff I guess, I didn't install any RAW processing programs yet. Link to comment
dabateman Posted October 20, 2021 Share Posted October 20, 2021 I thought you were going for the Red rum look as Halloween is coming to North America: https://m.dpreview.com/news/3677728197/cinestill-launches-redrum-an-iso-200-redscale-120-film-stock-with-spooky-autumn-tones Link to comment
lukaszgryglicki Posted October 20, 2021 Author Share Posted October 20, 2021 :) indeed looks a bit similar, but this is just a conincidence. BTW: I was suprised today that I can get really, really good images, hand-held by using brute-force on Nikkor 55/1.2 Ai. By brute force I mean - Nikon D600 fs mono + Kolari UV-Pass + shoot wide open at f=1.2, minimum 1/50s and let "A" mode choose ISO - usually just 300-1600 range... and today is not a sunny day, just the average... note that this is a very complex lens, not close to 4 elements design and it has coatings... after http://www.photosynthesis.co.nz/nikon/lenses.html I see (Nikon 55/1.2 Ai sn: 421667) it is 7/5 elements design from 1977 or 1978. With that setup I can shoot just like in visible light, handheld, very very easy Lv focus, even movies! So I guess sometimes the raw speed of f=1.2 is quite good, comparint to slow lenses like EL Nikkor 80/5.6 - I guess transmission drops way higher (say 360 nm vs 310nm) but still, I can make UV photos very very very easy with f=1.2. Here they are: Link to comment
nfoto Posted October 20, 2021 Share Posted October 20, 2021 I'm not the least surprised. What is lost in non-optimal glass is offset by sheer lens "speed". I have similar experience with other fast lenses, for example the Nikkor-N 5 cm f/1.1, The issue however with some of these lenses in UV is that they are horribly unsharp abused in that manner, but you seem to have found a nice exception. Just don't expect it to go deep down in UV, that would be asking too much. Link to comment
lukaszgryglicki Posted October 20, 2021 Author Share Posted October 20, 2021 Yep, this one can easy be focused in live view and seems to be sharp wide open at f=1.2, tomorrow I will also check other fast lenses, why not? Nikkor 28/1.4 AF-D, 85/1.4 AF-D, 135/2 Ai-S etc. Link to comment
Andy Perrin Posted October 20, 2021 Share Posted October 20, 2021 Yeah, if you are making them monochrome anyhow by not white balancing, then there is no issue of “reach” for the most part and you may as well pick the fastest lens that gets you the look you want. (when people on here are saying it’s hard to shoot handheld, it’s usually understood that they also want to have nice false yellows and things, but if you give that up, it’s much easier) Link to comment
dabateman Posted October 20, 2021 Share Posted October 20, 2021 Yes with a ZWB3 or Ug5 filter and your monochrome camera, you might get the best UV look with fastest shutter speeds. Link to comment
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now