Nikon Z6/Z7 as a UV/IR Conversion: the Bad News

[nfoto]

I tested more with the Kolari IRChrome filter against a standard O56 (deep orange) on the Z6. Again, Photo Ninja bombed out on the NEFs and I wrote Picturecode Support about this problem. RawTherapee 5.6 again came to assistance.

 

My understanding from using the Kolari filter under various weather situations is that it simply is the most overpriced blue filter one possibly can imagine. Apart from turning greens into a muddy yellowish red, it does nothing in terms of making an "IE" image. The O56 (with channel swapping directly in RawTherapee) comes much closer to what one would expect from a good IE emulation.

 

The examples below are shot with hand-held camera under rain-heavy skies. I selected this scene because it included red and yellow traffic signs. Nikkor 50mm f/1.8 S-Line (Z) on the Z6.

 

On the plus side, no banding was observed :D

 

Kolari "IRChrome"

 

 

Nikon O56

 

[nfoto]

The author of Photo Ninja informs me Nikon changed details in the NEF data structure for Z6/Z7, but apparently only for uncompressed files. So the advice to eliminate "modern art" striping is to use lossless compressed NEFs.

 

He also showed interest in what might be done in the conversion to mitigate striping. I'll forward some of the worst examples.

 

At least, some new insights and improved hopes..

[Andy Perrin]

Perhaps you should ask if something can be done about the UV banding issue during RAW conversion. If it were integrated (rather than having to run it through Topaz) then maybe the camera would be usable for UV also.

[Cadmium]

Werner, That is the typical B-410 look. It was Colin (on here) who first found that filter for use.

Here is Colin's example of B-410 2mm alone (no KG3).

https://www.ultravio...ndpost__p__8233

You have to click on his photo to enlarge it, I don't know why that is, it use to load large in his post without clicking on it, they might have changed something, oh well.

Now, try stacking your B-410 with KG3, and you might like the results a little bit better I might guess.

By the way, I like your fish shots, especially the first one, cool shot.

 

Birna, It was known that it is a Blue + Green + IR filter.

It doesn't work the same way as Aerochrome/EIR. It doesn't transpose visual red to yellow, it doesn't transpose any visual range color, it only replaces visual red with IR.

It basically suppresses the visual red range, and attenuates the IR range, then the visual red range is filled in by only the attenuated IR.

Now the trick becomes to get the orange to look more red, which is mostly all in the white balance... I am told.

You are a master of WB, but you may want to contact Yann about WB with this, maybe he can get it working better for you.

 

Shooting two photos, Visual, and RG9 (or RG715) + KG3 for IR,

pasting the IR stack shot into the red channel of the Visual shot, might result in a similar look, Visual Blue + Visual Green + IR (or what could be called LEWNIR, low end weighted near Infrared).

But then, I guess we have all basically done that before...

 

A more interesting test would be to get a sheet of Lee or Rosco Scuba Blue #729 ($7), and try that, alone and/or stacked with KG3 ($30+).

Since you are the only person I know of who has received one of them there 'Pre-order" gadgets... you would be an excellent person to compare.

You might be surprised, who knows until you try it.

[nfoto]

Perhaps you should ask if something can be done about the UV banding issue during RAW conversion. If it were integrated (rather than having to run it through Topaz) then maybe the camera would be usable for UV also.

 

Exactly what I asked, and I'll send him some test NEFs to demonstrate the issue.

[Andrea B.]

I'm thinking that if a Z6/Z7 works for everything except UV, then it might still be a worthy conversion as shown by many examples above. But I think I need to put something into the Sticky cautioning about using converted Z6/Z7 for UV until the problem has been figured out.

 

---

 

Birna, please also write to Raw Therapee about the white balance issue. They are quite responsive and might be able to provide an improved WB which would cover our needs. I think most raw converter developers do not encounter such extreme WB needs when writing these apps.

 

---

 

I think I read thru most of the preceding long thread. But I might have missed something. Has anyone yet performed a banding comparison using dichroic BaaderU and some non-dichroic filter?

 

---

 

I had never seen that crazy Photo Ninja uncompressed NEF phenom because I have always used the lossless compressed setting. Can you change the NEF saving compression in Capture NX-D or in View NX?

 

 

.

[Andrea B.]

Werner, your first fisheye foto in post #125 is almost Boschian, thematically speaking. I didn't quite catch everything the first time thru, but later some interesting details popped out at me. Love that foto!!

 

---

 

Birna, your foto in post #100 is all Alice-in-Wonderland to me. The fisheye swirl to it makes me think of Alice thinking about falling down the rabbit hole. I would just entitle it "Alice". B)

 

 

.

[Andrea B.]

I am unclear about the focus problem described above.

Life Pixel does not perform focus calibration on mirrorless cameras.*

So isn't the focus problem due to a natural, expected IR shift or UV shift?

 

I apologize in advance, of course, if I am misunderstanding what is going on. I do have a lot going on in the personal life these days with the upcoming move and all.

 

*MIRRORLESS. No focus adjustment (i.e. lens calibration) needed.

[nfoto]

The focus shift is way beyond what can explained with a response to UV or IR. To wit, the UV-Nikkor is badly affected even when no filter is attached to it and the same goes for the Coastal 60. These lenses should be virtually identical focus in UV and visible .... most wide-angle lenses use nearly all their focusing range just to get to infinity and with a Baader U some of them cannot focus at all. Life Pixel has not put the sensor in its correct position, period. I'll bring the camera to my friendly Nikon repair shop later and let them have a go at it -- probably they can shim the sensor to remedy the register error.

 

I have hope that Jim Christian over at Picturecode might improve Photo Ninja to render more UV and less striping :D Will follow advice and contact the RawTherapee crew regarding the failure to set w/b in that software.

[nfoto]

I availed myself of a interlude in the endless rain to shoot a few comparisons of Z6 against the old Fuji S5Pro (full-spectrum). As the Fuji and its precursor, S3UVIR, have been my chosen "false-colour IE" cameras for years, I was interested to see how the Z6 stacked up.

 

The lens was the old Zoom-Nikkor 28-45mm f/4.5 AI, and the filter Nikon O56 (deep orange). The Fuji RAFs were run through Silkypix 8 whilst Photo Ninja 1.37a handled the NEFs. I had set the camera to record NEF in lossless compressed mode and thus the bizarre "modern art" outcome of Photo Ninja did not occur.

 

Z6:

 

 

S5Pro:

 

 

Do note the zoom settings were different. I did post processing following my usual routines for falsecolour IE, so ideally the end results could be similar. Obviously they are not.

 

While the Z6 image obviously is much sharper, my nod goes to the S5Pro because its output is nicer, more pleasing, and better conforms to my notion of a digital IE rendition. More food for thought. The Rolling Stones succinctly stated the dilemma in their classic song "You can't always get what you want"

[Andrea B.]

Birna, thank you for the update about the focusing problem. I think I’m a bit surprised that there would be much room in a mirrorless for the FFD to move back or forward. But then again, it only takes a bitty little shift to foul up infinity focus.

 

*******

 

Just for grins, would you mind putting the Z6 into pinpoint AF-area Mode to engage CDAF and running some UV banding checks? Unless that has already been done above and I missed it.

 

I have that AF-area Mode setting under the i button. But if you don’t find it there, then hit the menu button and go to the Photo Setting Menu. After scrolling thru a few hundred pages, you will eventually find the AF-area Mode setting.

 

*******

 

I’m still all wild to get that Panasonic S1R converted. But I am so tired of being a Alpha/Beta Tester for Full Spec Conversions as used for UV. (OTOH, maybe I’ve saved some folks somewhere a bit of $$ by occasionally discovering certain problems. Hope so.) I’m thinking that it might be tricky to convert an S1R because of its shift-stacking sensor.

 

*****

 

Birna, sometimes the sliders in the Photo Ninja Color Enhancement tool can be used to move the colours around a bit. This example of your Z6 foto doesn't exactly reproduce the S3 output, but it is closer perhaps? This jpg is a bit smushed from the resize, so just look at the colors, ok? "-) I could have moved that one green roof a bit more. Those Photo Ninja color sliders mystified me for a long time. Eventually I figured out some useful tweaks.

[nfoto]

The camera is set up with Pinpoint AF.

[Cadmium]

Quick comparison between Hoya O56 and Schott OG550 and OG570. I got wondering, so here it is if anyone else wondered too.

Yellow (Schott OG550), Red (Hoya O56), and Black (OG570) lines all represent linear T data (Blue line is Hoya O56 linear Ti data).

[nfoto]

Used the new 14-30/4 Nikkor S-Line with a B+W 040 (orange filter today). Files look good and are easily cast into a pretty good facsimile of IE, however I have a feeling the lens is not well centered as there are signs of uneven sharpness one side to the other. More testing is required and of course, the camera itself cannot be ruled out. The sensor obviously is offset from the correct register distance, maybe more than 1mm. My friendly Nikon repair tech will have to take a look at it.

 

[nfoto]

Availed myself of a Nikon S (rangefinder) > Z mount adapter and gave the tiny 3.5cm f/3.5 W-Nikkor a work-out under Baader U filtration. Tis lens from the early '50s is a simple 4/3 Tessar design and should be able to do some UV imaging provided one doesn't demand it to go really deep into UV. It is very small and is almost dwarfed by the camera.

 

 

The lens renders quite sharp UV images and they are low in chromatic artefacts as well, unlike most of my other 35/3.5 lenses used for UV work.

[ulf]

Used the new 14-30/4 Nikkor S-Line with a B+W 040 (orange filter today). Files look good and are easily cast into a pretty good facsimile of IE, however I have a feeling the lens is not well centered as there are signs of uneven sharpness one side to the other. More testing is required and of course, the camera itself cannot be ruled out. The sensor obviously is offset from the correct register distance, maybe more than 1mm. My friendly Nikon repair tech will have to take a look at it.

 

The B+W 040 is my favourite for IE-simulations.

 

Nice picture.

By chance I noticed that the bottom half of it only, as a crop has pictorial potentials of its own.

Then the cloud reflections, plants above and below the surface reminds me of my favourite image by Esher, maned Three worlds:https://en.wikipedia...Worlds_(Escher)

[Cadmium]

B+W 040, and B+W 099 are both made of Schott OG550. All the same glass, 550nm longpass filters.

[dabateman]

Quick comparison between Hoya O56 and Schott OG550 and OG570. I got wondering, so here it is if anyone else wondered too.

Yellow (Schott OG550), Red (Hoya O56), and Black (OG570) lines all represent linear T data (Blue line is Hoya O56 linear Ti data).

 

Yes I got to wondering as well. I think the Nikon O56 filter is the same as a wratten #21.

 

[Cadmium]

If Nikon O56 is the same as Hoya O56 then they are both 560nm, #21 is close, it has 550nm = 29.0 T, and 560nm = 65.0 T.

 

http://www.hoyaoptics.com/eo_pdf/O56.pdf

[nfoto]

The B+W 040 on my Nikkor 105mm f/1.4 @f/1.6 (I compared it to the Rodenstock 100mm f/1.6) delivers excellent imagery on the Z6.

 

[Cadmium]

Pink.

[nfoto]

"Pretty in Pink" :)

 

The 105/1.4 Nikkor is sharper than the 100/1.6 Rodenstock, no surprise there, but the latter has a lot more atmospheric quality due to the unique blend of core sharpness and residual spherical and chromatic aberrations. After all it was never intended to be used as a general-purpose lens in daylight, but who will not avail themselves of its innate character to make those special pictures? On the Z bodies, the Rodenstock can focus easily to infinity, so one is not limited to moody close-ups any more.

[nfoto]

I had already made an adapter for my 55mm f/1.2 CRT-Nikkor to the Z, now found it could do double duty by providing the mount for a Kyoei 35mm f/3.5 lens which already had been converted to 42mm focusing mount. A quick test shooting the darkness outside indicated the lens did focus to infinity and also showed how IR can be reflected off window panes. As we have triple glazing, lots of doubled lines as well in this not entirely planned "selfie".

 

 

No striping for a change (ISO 1600, no filter).

 

I'll do UV, IR, and emulated IE with this lens tomorrow provided the weather isn't as bad as the forecast. I plan to include some of the Nikkor(RF) lenses I already have the proper adapter for. The old 3.5cm f/3.5 W-Nikkor (approx. 1954) apparently does nice in all spectral ranges. The fast 5cm f/1.4 Nikkors (RF) tend to make a hot spot in IR, but the 5cm f/2 seems to be clean. I add the 8.5cm f/2 and 10.5cm f/2.5 Nikkors (RF) as well.

[nfoto]

I decided to keep the Z6 -- despite the banding issue (can be dealt with using the Topaz plugin, awkward but doable) and its offset register distance. For the latter I decided to add shims to the various adapters I either have, or are constructing, for this camera. The advantage is the shims can be easily removed once the camera has had its register distance corrected, but as of now, the local Nikon repair techs neither have the repair manuals nor the required tools for such operations on any Z. Until the situation improves I can use my custom adaptation.

 

The native Z lenses are suitable only for IE emulation or IR, and as of now, only the 14-30 and 50/1.8 are usable candidates to this end. Apparently they have sufficient focal adjustment to cope with the bad register distance of my converted Z6.

 

Today, I commenced the adjust project by adding 0.6mm shims to the Nikon S (RF) - Z adapter and this measure made my old rangefinder lenses focus properly with a little to spare in the distant region. As focus tends to move closer when using filters, this behaviour is about as good as it gets.

 

The old rangefinder lenses do have [simple] coatings yet pass some UV. Probably not very deep into the UV band, but sufficient to give the "UV look" to everyday objects. Filter threads are varying from 40.5 to 52mm thus a lot of step rings are required. Fortunately the legendary 8.5cm f/2 Nikkor (made famous by David D. Duncan in the Korean war) has 48mm threads thus the Baader U fits directly. The lens flares badly unless a very long shade is provided, though.

 

Here is the "new" Z6 with the 8.5cm lens and a makeshift long hood.

 

 

I had a lot of fun using this combination today. It felt enjoyable to escape the pressure of getting "perfect" UV photos for a change :D

 

Lupines run rampant on a road verge. These species are now on the national Black List and it is forbidden to sell stock of them.

 

 

The nearby river is in spring spate these days, thus heed the warning sign and mind your steps. Falling into the current might be fatal.

 

[Andy Perrin]

Love the river there. I use a BadElf for geolocating my photos also.