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

[Cadmium]

Birna, All those pics are nice. I especially like the flower/tire/car pic, and the shoes pic.

[dabateman]

The shoe shot is nice. Reminds me of the Wizard of oz.

[nfoto]

The plot thickens.

 

Today I shot some footage with an orange filter (O56) that allows a pseudo-IE (false-color Infred Ektachrome) 'look' to the outcome. I did channel adjusts directly inside Photo Ninja, as this approach resulted in more pleasant colours than doing the traditional channel swap in Photoshop. I have updated Z6 firmware to v.2 and at least with this "false colour" approach no banding was observed even at ISO 100??

 

Curiouser and curiouser.

 

 

The 50/1.8 Nikkor S-Line performs very well with the O56 filter and delivered crisp and clear images with no hot spots. At least, I didn't see any during the shooting session even though the lens was stopped well down (f/8-f/11). The red Norwegian Mail carrier became bright green as I would expect from a good IE emulation.So far, all went well.

 

 

Later I did a second attempt with the 35mm f/1.8 Nikkor S-Line. Images were still very sharp, but now a bad hot spot marred the outcome. I "rescued" some shots by judicious massage of the image characteristics, but this approach is untenable in the longer run. I didn't try the 24-70/4 zoom lens, but as this hot-spots badly in IR it likely would not work with the O56 either (have to test, though).

[nfoto]

The modification of the Z6 continues to puzzle me. All my F-mount wide-angle lenses are way off the focusing scale to indicate the register distance is a tad too short. When using CRC-equipped lenses this leads to the wrong optical correction being applied as a consequence of infinity focus now is believed to occur at approx. 0.5m instead leading to massive overcorrection and thus, very unsharp corners of the frame. With longer focal lengths the issue is negligible or non-existing.

 

I'd say this is an error committed by Life Pixel.

[Andy Perrin]

Will you send it in for correction?

[Cadmium]

The modification of the Z6 continues to puzzle me. All my F-mount wide-angle lenses are way off the focusing scale to indicate the register distance is a tad too short. When using CRC-equipped lenses this leads to the wrong optical correction being applied as a consequence of infinity focus now is believed to occur at approx. 0.5m instead leading to massive overcorrection and thus, very unsharp corners of the frame. With longer focal lengths the issue is negligible or non-existing.

 

I'd say this is an error committed by Life Pixel.

 

Birna, I think you might be onto something there. I have a D7200 they did that does the same thing, at least with out or visual band, and Andy had similar problems I think.

Something is amiss.

[Andy Perrin]

It does sound similar to what happened to me. In my case they told me they had simply put the wrong replacement glass in for the camera.

[Cadmium]

Well, too many bad things happening with their cameras these days.

[Alaun]

This week, I had time to pick up my Z6 from Optic-Makario (He uses a borosilicate glas with AR coating on both sides, he said it transmits from about 280 to 2000nm).

 

Some quick first remarks:

 

As expected, I get the banding (Firmware 1.0), but I think, it is not a big issue, as I can remove it with Topaz Denoize 6 (not the AI version).

Some pictures taken with a flash of a dried flower no banding was visible. A picture taken outside of a flower, some very very minor banding was visible looking at 400% magnification.

 

I updated to Firmware 2. today and did some IR-shots with my Nikon 8-15 and the Wratten 88 A, these pictures were fine (only little hot spot from the lens on some pictures).

[Alaun]

Some pictures taken with the Nikon 105 UV and the Baader U; original, original demonised, crop and crop denoised. WB was done in PSCC (double use/two step use of the WB in RAW editor).

[Alaun]

The only little flower I found (same setting):

[Alaun]

And some of the IR pictures

[nfoto]

Werner's examples indicate that UV and banding are interconnected "features" of the Z6.

 

When some visible light is added, banding apparently disappear ± completely. I have shot "false-colour emulated IE" (through a deep orange O56 filter) and observed no signs of any striping whatsoever.

 

 

Taken with the 105mm f/2.5 Nikkor AI. The black phyllaries on the dandelion heads indicate true IR is also registered.

[Cadmium]

Werner, Very nice examples.

[nfoto]

I probably keep the Z6 despite its documented illnesses. The camera simply invites to experimentation, a trait I really like. Just built a rear filter holder combined with F-mount for it to allow all my various Nikkors etc. to be used for UV, IR, or multispectral purposes.

[Andy Perrin]

The camera simply invites to experimentation, a trait I really like.

Me too. This is why it will always be mirrorless cameras for me (even if I someday switch brand).

[dabateman]

Birna,

Maybe you should get a Fuji 50S or 50R converted. Trolling the web, there seems to be at least two that have been done. One for IR only with 850nm filter and one full spectrum. I found that out here:

https://www.getdpi.com/forum/medium-format-systems-and-digital-backs/65409-gfx-infrared.html

 

I have become interested in the Fuji cameras as they seem to have outstanding high ISO performance, all though a cheat since only shoot to 1600 then push 5 stops to go higher. Also I like the control layout. The flange is 26.7mm, so would allow for almost any lens as well, with a wide mouth to accept many adapters.

 

 

Both of these seemed to have been done by Kolari.

[nfoto]

I have the Fuji S3Pro UVIR and the S5Pro (full spectrum). Those are as much Fuji as I can stand. I wish to utilise F-mount lenses natively as most of the manual ones are CPU-modified.

[Alaun]

Birna or Andrea, when you do the conversion (WB of Z6-files) with Photo Ninja, can you open the raw file directly?

I only get something like this (PN 1.3.7a on MAC and raw uncompressed 14 bit):

 

 

 

 

Though it looks nice, it is not really what I want :-)

Off course, I can first create a TIFF, but ...

[nfoto]

No problem opening the NEFs directly in any version of Photo Ninja I have tried so far, including the latest 1.37a.

 

Fancy banding pattern, though :)

[Alaun]

Ok, thanks.

Maybe I have to deinstall and reinstall it. I still have an older version also installed, that might be the problem. A project for the next weekend. :)

[Andy Perrin]

Alaun, you should make some modern art with that!

[nfoto]

Unfortunately, Photo Ninja can choke on some NEFs -- never had this occur to me earlier, but today I also got "modern art" courtesy PictureCode without any efforts on my side. This occurred with the first test shot done with the Kolari "IR Chrome" filter. A more normal version obtained through RawTherapee 5.6 to save my day.

 

Photo Ninja 1.37a

 

Raw Therapee 5.6 and some PS adjustments to contrast and red colours, which initially came out yellowish.

 

 

As parts of the visible range is included using the Kolari filter, the banding issue is not manifested. However, I have to say that this filter is not a real substitute to emulating "Infrared Ektachrome". It is true green vegetation becomes reddish, but red paint does not turn into yellow or green, and the garishness of the reds reminds me more of EIR in E-6 than IE in E-4. Which is not what I had hoped for, in particular considering the really steep price of this filter.

[Alaun]

So it's not my faulty installation. Just tested to convert raws first to dng's, then it works fine.

Here are some converted and tweaked with PS. Zoom Nikkor 35-200/3.5-4.5 with Hoya B-410:

 

(Yes, there are some spots)

 

 

 

 

 

 

[Alaun]

And some more with the little new fish zoom from Nikon and Kodak Wratten 88 A gel filter: