PrimaLuce Nikon Z5 Full-spectrum camera review

[Stefano]

The problem would be separating the two green subpixels in a Bayer filter.

[Andy Perrin]

Stefano, I don’t know why that’s a problem. They are not combined in the raw. You don’t have to separate anything.

[Stefano]

I thought they were combined into a single green subpixel, to have the red, green and blue color information. If they are stored separately then there's no problem.

[dabateman]

Couldn't you just process it as a monochrome image in RawTherapee?

I will have to test that with a deep IR filter.

[Cadmium]

>850nm IR tends to look sharper than expected due to the transparency of the Bayer allowing each pixel to effectively act as 4 subpixels. The sharpening effect is visible with a normal raw converter (as demonstrated here) but I think more could be done to effectively quadruple the pixel count by writing a converter to deliberately make use of the subpixels as if they were pixels. Another rainy day project for me.

 

Where is the 850nm example?

[Andy Perrin]

Where is the 850nm example?

Basic concept and discussion here, although it's been proposed multiple times in the history of the board. But nobody ever get's around to it...

 

https://www.ultravio...me-conversions/

 

You can see at the link how much sharper the IR is before demosaicing, which just blurs it all together in the case where the Bayer is transparent. You still get some apparent improvement though, even *after* demosaicing.

[Cadmium]

I only meant in this topic so far. I looked, didn't see any 850nm.

[Andy Perrin]

Unless she made it monochrome, the IR image above must be mostly 850nm+.

[nfoto]

When I can shoot a dandelion it's technically *spring*. That time finally arrived in my part of the world. A little odd though that the grass is still not green, so not the truest of springs but it'll have to do for now I surmise.

 

There is the infamous "UV striping" to battle, yet in today's shooting I didn't feel too bothered about it. Probably helped that I had to shoot at ISO speeds 800 and higher as the sun refused to come out from its hiding place behind dense clouds.

 

While the UV-Nikkor 105mm f/4.5 never will win a bokeh contest, it is wickedly sharp and the PrimaLuce Z5 allows it to portray the most intricate details of these composite flower heads.

 

A 100% crop ISO 1000 to illustrate what the camera-lens combination can deliver.

 

 

Original shot at 1:2 magnification. The NEF was run straight through Photo Ninja without any consideration for the PDAF artefacts, which weren't bothersome in this case as far as I'm concerned.

[nfoto]

When ISO drops to 800 (or below), the striping begins to make its presence more visible. Still not too bad and a quick "spring cleaning" in say Topaz Denoise or processing through Silkypix would sort the annoyance pretty well.

 

100% crop, again f/11, 0.6 sec. Baader U.

 

[Shane]

Thanks for a great review Birna. I was in search of information for a Z5 full spectrum conversion and this has pretty much covered it.

[KhanhDam]

nice review. Z5 uses a front illuminated sensor and z6 uses a BSI sensor, (which are about 1 stop more sensitive for UV work , at least for expensive mono conversions where bayer layer is removed)  However as your work shows seems like BSI suffers from PDAF.

Isnt' ghosting in most lenses caused by anodized aluminum reflecting light and also read that black felt materials used might actually be white in the IR spectrum. These 2 factors cause stray light and ghosting. I dont' think the sensor affects if a lens is good or bad for IR. 

[Andy Perrin]

10 hours ago, KhanhDam said:

seems like BSI suffers from PDAF

That issue is not in every camera. Only some Nikons and a few others. Sony uses BSI sensors and (AFAIK) does not have this banding issue. I think you are conflating two separate things here. BSI just means the wires are not blocking the light.