• Ultraviolet Photography
  •  

UV performance of IS Pro vs full spectrum D600

UV Camera UV Lens UV Lighting
15 replies to this topic

#1 Ultrapurple

    Guardian of the Aardvark

  • Members
  • 27 posts
  • Location: Bedford, UK

Posted 09 October 2020 - 19:37

An unscientific comparison of Near-UV sensitivity between Fuji IS Pro and full spectrum converted Nikon D600

The Fuji IS Pro DSLR was marketed as a 'forensic' camera with extended IR and UV sensitivity. It was based on a Nikon D200 body and is compatible with Nikon F-mount lenses. The Nikon D600 is a full-frame DSLR that I understand had some teething troubles and was quickly superseded by the improved but visually very similar D610. The teething troubles do not affect the camera to any significant extent when it is converted for full-spectrum use (done in the usual way, by replacing the IR/UV cut filter with a material having wideband transmission).

A test scene was created using an X-rite Colorchecker Passport, a pair of glass spectacles with non-specific UV-protection photochromic lenses, and a small teddy bear called Dave.

A SB-140 flash was mounted at about the same level, positioned about 1m from the scene, offset slightly toward the 'Dave' side to avoid obscuring the view of the camera. The SB-140 was used unfiltered and with the SW-5 UV clip-on filter.

An interchangeable head tripod was arranged about 1.3m from the scene, straight-on and at approximately the same level. The distance was arbitrarily chosen so the scene did not exceed the image area of the Fuji camera crop sensor.

Two cameras were to be compared: a Fujifilm Finepix IS Pro (2007-2010) and a full spectrum converted Nikon D600 (2012). An original UV-Nikkor 105mm f/4.5 lens was equipped with a 2" Baader U filter on the hinged mount. Only one lens was available for the test so it was swapped between cameras as required, re-focusing as necessary using Live View. The focal reference was the white cross on the coloured side of the colorchecker.

Both cameras were set to ISO 400, 1/60s, and to record both full resolution JPG and maximum quality RAW files. For simplicity, only the JPG files have been assessed here.

The test environment was a domestic room lit by a COB-type LED ceiling lamp. Tests showed that the level of illumination was sufficiently low compared to the flash intensity that it would not materially affect results.

Other than swapping the lens between camera bodies and adding/removing filters as described, no other changes were made between tests. The cameras and flash had well-charged batteries and sufficient time elapsed between exposures to ensure the SB-140 delivered full power for each exposure.

Three sets of tests were performed with each camera, during which the lens aperture was varied from f/32 to f/4.5 in 7 steps.

Test 1: Baader U filter mounted on lens, no filter on flash.
Test 2: Baader U filter mounted on lens, SW-5 UV filter on flash.
Test 3: No filter on lens, SW-5 UV filter on flash.

The results, summarised in the attached image, showed that the full-spectrum converted Nikon D600 is significantly more sensitive to UV in the spectrum passed by the Baader U filter. Presence or absence of the SW-5 UV filter made only a minor difference to the exposures and this small effect was subjectively similar on both cameras.

Test 3 (filtered flash, unfiltered lens) was intended to give a general indication of IR+UV sensitivity. The SW-5 UV filter is known to leak substantial amounts of IR (as evidenced by the transparency of the spectacle lenses) but the test did indicate that the UV+IR (mainly IR) sensitivity of both cameras was broadly comparable.

Copies of the RAW and JPG files will be available for a reasonable time on request via a PM, but please be aware that the zipped files total some 2.6GB.

Conclusion: 380-320nm UV sensitivity of the IS Pro is seen to be markedly inferior to that of the full spectrum converted D600.

Further research might include checking the spectral response of each camera in more detail, perhaps using a wide band UV light source and a monochromator, however such equipment is not presently to hand.

Attached Images

  • Attached Image: Group A-B-C comparison final.png
  • Attached Image: Scene_vis_reference_GR4_0893.JPG

Edited by Ultrapurple, 09 October 2020 - 19:46.


#2 nfoto

    Former Fierce Bear of the North

  • Owner-Administrator
  • 2,806 posts
  • Location: Sørumsand, Norway

Posted 09 October 2020 - 21:22

I'm not surprised. As far as my memory serves, Fuji claimed the IS camera did UV down to pprox. 380nm, vs. approx. 350 for the modified S3Pro version sold as S3 Pro UV/IR.

I have the latter Fuji and it certainly levels out way earlier than the D200 modified (both using UV-Nikkor lenses and Baader U filters) when it comes to UV response. The "full spectrum" modified S5Pro still cannot entirely match the D200 (modified), but responds much better than the S3Pro UV/IR (and by induction, the Fuji IS).

Thus, for anyone keen on having a full-spectrum Fuji, the S5Pro gets my vote.

For IR and false-colour emulated IR, though, both Fuji models (when modified) do the job admirably.

#3 Cadmium

    Member

  • Members+G
  • 3,821 posts

Posted 10 October 2020 - 00:37

I am wondering if this difference is due to the glass they are using over the sensor, or if it is the sensor itself that lacks the range?
A newer forensic model is the Fuji X-T1 IR, which I think is rated at 380/350 also. Perhaps the forensic people aren't looking for the deeper range, but just wondering if it is the glass.

#4 dabateman

    Da Bateman

  • Members+G
  • 2,118 posts
  • Location: Maryland

Posted 10 October 2020 - 04:19

I had thought that Fuji just discovered that their cameras were leaky down to at least 380nm. Just like an off the shelf Olympus camera. I wonder if there is any difference between an off the shelf Fuji and a forensic one.

My first converted camera was an Olympus E510 that I bought used. It was only 1 stop more sensitive to UV than my Olympus E3. But my E3 had a better high ISO, about a stop better. So it was like they were the same.
Recently I took the E510 apart to see what was going on. It had the blocking filter removed from the sensor and nothing was placed back. This explains why my IR sucked as it wasn't focusing properly to infinity with native lenses. But it had the dust shaker intact. So that piece of glass was responsible for cutting UV off at 350nm and below, matching my E3. Basically indicating to me that only an IR blocking filter is on the sensor and the dust shakers are blocking the UV.

When I had it open, I cut off the dust shaker glass leaving nothing there. Yes it will not focus to infinity with native lenses. But it didn't properly before. Now I can see to 313nm with the camera. But not well at 300nm. Its too noisy there for my liking.



#5 Cadmium

    Member

  • Members+G
  • 3,821 posts

Posted 10 October 2020 - 04:48

Any difference, you mean as far as the UV end. They do IR up to about 1100? something like that, the forensic ones, Fuji X-T1 IR.
Just wondering if changing the glass in them would make the UV deeper.

Edited by Cadmium, 10 October 2020 - 04:50.


#6 Ultrapurple

    Guardian of the Aardvark

  • Members
  • 27 posts
  • Location: Bedford, UK

Posted 11 October 2020 - 17:36

It's probably worth me clarifying that although the Fuji IS Pro is built on a Nikon D200 chassis, the sensor and image processing is purely Fuji.


In fact, it's a bit of a Frankencmera. The metering, autofocus, shutter and so on is achieved using purely Nikon software: in that respect it behaves just like other Nikons. But when you go to the menus you discover the Fuji software, which is rather different from the Nikon equivalent. I'm fortunate that I have a couple of Fuji bridge cameras that use a very similar software setup so it wasn't too unfamiliar. But the (Nikon) buttons on the camera body don't quite line up with the Fuji functions, although they have clearly done their best to make it as logical as possible. However, a generous person might describe the image review function as "clunky".

One item of interest: Fuji claim the IS Pro was the first DSLR with Live View. Yes, it has it; yes, it works - but you're limited to a maximum view of 30 seconds. I am not sure why.

On a completely unrelated note, I was playing with my laser cutter this afternoon, making a small box from 3mm plywood as a test. At the same time I tried a long exposure with the converted D600, UV-Nikkor 105/4.5 and Baader U filter. I think the result is intriguing.

Attached Images

  • Attached Image: _UVA9447_PerfectlyClear_cutting in progress-UV-small.jpg
  • Attached Image: _UVA9480_PerfectlyClear crop small.jpg

Edited by Ultrapurple, 11 October 2020 - 17:39.


#7 nfoto

    Former Fierce Bear of the North

  • Owner-Administrator
  • 2,806 posts
  • Location: Sørumsand, Norway

Posted 11 October 2020 - 17:41

Intriguing, yes.

By the way, the S3Pro UV/IR had liveview -- 30 secs.

#8 Stefano

    Member

  • Members(+)
  • 1,074 posts
  • Location: Italy

Posted 11 October 2020 - 18:12

View PostUltrapurple, on 11 October 2020 - 17:36, said:

On a completely unrelated note, I was playing with my laser cutter this afternoon, making a small box from 3mm plywood as a test. At the same time I tried a long exposure with the converted D600, UV-Nikkor 105/4.5 and Baader U filter. I think the result is intriguing.
I think you saw UV, maybe with a bit of IR leaked through the Baader U filter (the "setup" was far from optimal, a lot of IR and a little bit of UV, that's where at least OD 5 or even 6 is required in my opinion). That's not the main discussion in this topic, but it is quite intriguing, I agree.

#9 Andy Perrin

    Member

  • Members+G
  • 3,397 posts
  • Location: United States

Posted 11 October 2020 - 18:52

Stefano, how are you drawing the conclusion that there was a lot of IR and a little UV? Is the laser cutter an IR laser?

#10 Stefano

    Member

  • Members(+)
  • 1,074 posts
  • Location: Italy

Posted 11 October 2020 - 19:43

View PostAndy Perrin, on 11 October 2020 - 18:52, said:

Stefano, how are you drawing the conclusion that there was a lot of IR and a little UV? Is the laser cutter an IR laser?
Not necessairly an IR laser, I don’t know the laser being used, but to cut wood with a laser you need to burn it, probably it becomes red-orange-hot. Objects at that temperature (I know by experience) do emit enough UV to be seem by a camera, especially at long exposures. The amount of UV is anyway very little, so IR radiation due to blackbody radiation probably contaminated the image.

If the laser emits IR light, it could be that in the image above you see IR leaking through the Baader U.

#11 dabateman

    Da Bateman

  • Members+G
  • 2,118 posts
  • Location: Maryland

Posted 11 October 2020 - 23:35

Ues the Fuji IS pro was an interesting oddball camera. Its basically a S5, with the cool hexagonal pixels. I remember wanting an S5 for the great dynamic range. But I think its only 6Mpixel.
The D200 was a 10Mpixels sony Sensor that Sony placed in one of its first cameras, at a fraction of the cost.

The Fuji live view was first, but only black and white and max 30 seconds. Olympus made live view fun to use. I didn't buy a E330, but still have my E510 and E3 cameras.

Sorry Cadmium, I get too fixated on UV and forgot about the IR. The camera was reported to have signal from 380nm to 1000nm with special glass. I think it used any old Glass in hindsight as that response isn't that great. If you can locate a cheap Fuji S5 you could convert it to test out. I remember almost buying one for $100 6 years ago.
I think I read that someone somewhere used one converted and saw significant increase in UV sensitivity.

#12 nfoto

    Former Fierce Bear of the North

  • Owner-Administrator
  • 2,806 posts
  • Location: Sørumsand, Norway

Posted 12 October 2020 - 05:22

View Postdabateman, on 11 October 2020 - 23:35, said:

.... I think I read that someone somewhere used one converted [S5Pro] and saw significant increase in UV sensitivity.

That someone would be me? The UV response of a modified S5Pro does improve, but other brands are superior. However, the S5Pro has a very nice rendition of false-colour IR and it's mainly for this purpose I'm using my Fujis.

#13 dabateman

    Da Bateman

  • Members+G
  • 2,118 posts
  • Location: Maryland

Posted 13 October 2020 - 12:38

View Postnfoto, on 12 October 2020 - 05:22, said:



That someone would be me? The UV response of a modified S5Pro does improve, but other brands are superior. However, the S5Pro has a very nice rendition of false-colour IR and it's mainly for this purpose I'm using my Fujis.

Just might have been. My memory has become a little foggy over the last 3 months. I can't remember exact dates and pages like I used to.
Good to know the Fuji's are best for IR than UV. I was always curious about the hexagonal pixels. I might just stick to the Sigma for its fun IR though. I do hope Sigma can work the bugs out of their 135 format Foveon sensor camera. Seem like the technology didn't scale well. But if they can fix it, it should be really good as they may work out an underling problem.


#14 nfoto

    Former Fierce Bear of the North

  • Owner-Administrator
  • 2,806 posts
  • Location: Sørumsand, Norway

Posted 14 October 2020 - 08:48

The main drawback to these Fuji models is the peculiar RAF format which few programs can deal with in a satisfying manner. This difficulty stems from the dual-size pixel layout. Most RAW converters simply ignore the small pixels leading to a loss of resolution so the output is effective 6 MPix. Ideally the outcome should be 12 MPix, but even the better converters can only make the final result approx. 9-10MPix. They do, however, reclaim much of the added dynamics provided by the smaller, less light-sensitive pixels embedded in the sensor.

RawTherapee (to some extent), S7Raw (which is not actively delevoped anymore, but hey that applies to the S3/S5Pro as well), Fuji-s own Hyper-Utility3 (worst UI interface ever), and SilkyPix Pro can hande the RAFs. ACR claims to do, but fails due to the mishandling of the small pixels.

#15 dabateman

    Da Bateman

  • Members+G
  • 2,118 posts
  • Location: Maryland

Posted 15 October 2020 - 04:21

Just spun around the web. Could find the S7raw program. All web links were dead. I was able to get a manual oddly enough.
Raw therapee since 2017 doesn't handle the pixels correctly. Seems to drop something out.
Dcraw seems to be best, but need the fujiturn.c code which is still up on Dave Coffin site.
Fuji hyper utility seems to also range best to second after dcraw.
Photoacute has profiles for the S5, but I am not sure if its built in raw converter is being used or if you need to find an old Adobe dng converter like 8.45.

Oddly just looked and a minty S5 is asking for $750 on the bay.
I recently got a mint Em1 with 900 shots on the bay for $250. I might have done better.



#16 nfoto

    Former Fierce Bear of the North

  • Owner-Administrator
  • 2,806 posts
  • Location: Sørumsand, Norway

Posted 15 October 2020 - 09:08

I found the zipped file for S7RAW. It works on 64-bit Win10. In addition there is an install for 32-bit systems. Anyone interested can shoot me a PM to get the file.

The UI is a bit unpolished yet entirely functional and well worth the try if one struggles with processing of S3Pro/S5Pro RAFs.