Jump to content
UltravioletPhotography

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


Andrea B.

Recommended Posts

I get the same banding with my unmodified Z6 (30s exposure, ISO 200, UV-Nikkor & Baader U).

Here is a crop. The upper part of the picture -about 140 rows of pixel- does not show the banding (seems there are no special pixel)

post-21-0-06602100-1555366538.jpg

Link to comment

When I was in my banding blues with the D7100 I tried a Topaz de-banding plug-in which helped a little.

Dave

Link to comment

As I mentioned earlier, shooting IR seems to eschew the banding issue ± completely.

 

Again with the UV-Nikkor, but this time with the B+W O-93 (Wratten 87c equivalent if memory serves). 1600 ISO and if there are vestiges of banding here, they are at or below my detection level. There are however a few dust bunnies that I might or might not have added to the sensor?

 

DSC_0366_v2.jpg

 

Interesting news about the de-banding plugin from Topaz. I must try it.

Link to comment

There are however a few dust bunnies that I might or might not have added to the sensor?

 

C'mon now, do you really think I'd let a conversion having a dust bunny the size of Texas (upper right-hand corner) pass my inspection? I made 3 of the requisite ceiling shots for close inspection of trapped dust. Saw nothing.

 

Fortunately, it is very easy to blow or swab dust off the Z6 internal filter given that it is only 18mm from the mount.

 

*****

 

My Z6 test shots were no where near as bad as yours for this banding, and I still don't understand why.

 

The Topaz plug-in does look promising, BUT why should we have to use a camera for UV which requires such a repair to the file? I say no. There are so many other good cameras for conversion which do not have PDAF striping or banding.

Link to comment

The Topaz Denoise v. 6.01 plugin/standalone can virtually eliminate the banding issue. At the expense of added complexity to the workflow of course. The license for that software should be paid by the camera maker (?).

 

However, I think banding is just one of the issues with using the Z models now at hand for UV/IR. More troublesome is the colour vignetting that is seen with all filters I have tried so far. There is no need at all for wide-angle lens to generate colour vignetting -- in the previously shown IR image the uneven residual colour is evident. One has to cast the image into pure B/W in order to mitigate these artefacts, I'd guess.

 

A pity that the promises of a superb system for non-visible imaging turned out to be just that - a dream not coming true. The short register distance and the excellent ISO behaviour would be perfect for such a system, if it otherwise delivered.

Link to comment

the colour vignetting that is seen with all filters I have tried so far

 

oh no! another problem? I think this is the proverbial 'nail in the coffin' for the conversion of the Nikon Z6.

 


 

Birna: A pity that the promises of a superb system for non-visible imaging turned out to be just that - a dream not coming true. The short register distance and the excellent ISO behaviour would be perfect for such a system, if it otherwise delivered.

 

Birna, again, I definitely share your disappointment. Perhaps the next generation of Z will be better. You and I are so hooked on Nikon image quality and ergonomics, that is indeed difficult to look elsewhere. But we do have other good choices for a short FFD.

  • For example, as mentioned by David above, the Olympus conversions are really rather good.

  • There are some Sonys which would be OK if one takes the patience to learn the menus and become used to the ergonomics. Just because Birna and I have problems with this doesn't mean it isn't do-able. :lol: :rolleyes: [beware the Sony version of IR-LED shutter monitor in some models.]

  • The new full-frame Panasonic S1R is getting some good reviews from old friend Lloyd over on https://diglloyd.com/. I think I'll go see what he has to say about PDAF banding/striping, if any.

Link to comment

Update to preceding mention of Lumix S1R: It does NOT have PDAF, therefore no problems with banding.

 

That.sounds.good.to.me!|1!!!11|!!

Link to comment
I am definitively not keen on any Olympus. Was hardly convinced by the image quality of the E M1.2 I currently use in an underwater housing and really wish for a better system.
Link to comment

By the way, I found the source for the light leak that occasionally can be seen with an F-mount lens on the FTZ adapter. Turns out light might enter through the aperture ring around the point where the coupling prong ("rabbit ears") is located. A bit of black tape judiciously placed eliminates any unwanted light from entering. As this light would be in the visible range it has a disproportional and adverse effect on the UV image.

 

I have only observed the problem sporadically for UV shooting, not for IR (so far). The FTZ itself appears to be entirely leak-free.

Link to comment

I used a retrieval app and dug out two photos from my test.

 

Nikon Z6 + Coastal Optics 60/4.0 + BaaderU UV-Pass Filter + Sunlight

 

Subject: Jonquil (Narcissus)

 

As shot.

No edits, JPG extracted in Photo Mechanic.

f/8 for 1/2.5" @ ISO-3200

There's not much UV out there just yet. The Z6 was on tripod, but it was windy.

Clearly, downsizing for posting here supresses quite a lot of the banding.

This really doesn't look so bad.

(Well, other than the fact that there's some motion blur and the composition hardly is stellar.

But my excuse is that I was thinking about testing. And was busy fiddling with settings.)

1_jonquil_asShot.jpg

 

 

As shot, unresized crop.

The banding is not extreme like in Birna's photos,

but banding can be seen behind the flower petals.

I don't have any idea why I didn't get strong banding like Birna did.

2_jonquil_100crop_asShot.jpg

 

 

Unresized crop with added contrast.

Bringing out the banding a bit. Not that we would normally do that, of course. :lol:

3_jonquil_100crop_midtone.jpg

 

 

Unresized crop with added noise.

So, yep, this does almost obscure the banding but at a rather large cost. Yuk!

When I added any less noise than this, the banding was still apparent.

If you know what to look for, the banding can still be detected here too.

4_jonquil_100crop_addNoise.jpg

 

 

Base file as seen by Photo Ninja.

No white balance or profiles added by PN. Very typical raw (in PN style) colors.

5_jonquil_pnFromCamNoProfile.jpg

 

 

White balance made in PN on background behind flower/leaves.

The circular dichroic effects are easily seen on the periphery.

6_jonquil_pnAveWbBkg.jpg

 

 

Previous WB photo with Red, Orange and Magenta turned OFF in PN.

One of the little tricks for dealing with peripheral dichroic effects.

However, this little trick is not always useable for cleaning up such effects.

My next example (later this eve) could not be cleaned up this way.

And the flowers in the lower left corner still need a slight push towards blue.

7_jonquil_pnAveWbBkgRedOrMgOff.jpg

 


 

I have another example to post later this evening.

 


 

Birna, FWIW, on my Retina screens, I can see some banding in the IR shot in your Post #29.

JPG conversion and downsizing obscures a lot of it, but it is there.

birnaIRcrop3.jpg

birnaIRcrop2.jpg

 

I am now wondering why Life Pixel, Kolari Vision and MaxMax are converting any Z6 or Z7 at all.

Why are we here on UVP always the ones who find this stuff?

 


Link to comment

I been trying mirrorless conversions since 2009.

But I've never yet gotten the image quality I seek from one of my mirrorless conversions. Noise, PDAF artifacts, shutter shock, lossy compression, flimsy mounts, poor dynamic range, EVF blackout, badly implemented focus peaking, poor Live View, banding/striping. Been through it all. Wondering when are they ever going to get it all right?

 

It is safe to say here that I am definitely too much of a "purist" on the topic of image quality. :D :D :D

 

Hi, Andrea, long time no see.

 

So, the next bet seems to be Panasonic S1/S1R...at least cheaper than Leica SL..

Link to comment

There is indeed some minor banding in IR, but nothing objectionable like in UV. Downsizing and changing colour space to sRGB make gradients less smooth.

 

Akira, the Panasonic is bigger and heavier. Not my ideal camera when arthritis strikes.

 

I'll give the Z6 the benefit of doubt yet another day, then decide whether it should be returned and reconverted and a refund sought.

Link to comment

Hello, Akira !!!

 

Indeed, it has been awhile since we have talked. I hope all is well with you,

and that you still occasionally pursue a bit of IR or UV shooting. :)

 

I am indeed intrigued by this new S1/S1R line from Panasonic. I've always though my old (very) Panasonic G and my converted GH1 were so easy to learn and use. But I was always hoping for a better sensor. Will the S1 or S1R provide? It is looking good. And how nice it would be not to have PDAF banding. So we shall see, we shall see.

Link to comment

Another example showing that detail "hides" PDAF banding. No surprise, really.

But it does sometimes make it more difficult to detect banding problems.

 

Nikon Z6 + Coastal Optics 60/4.0 + BaaderU UV-Pass Filter + Sunlight

 

Subject: Old, Dried Leaf on an Outdoor Carpet with Lots of Fibers

 

As Shot with Minor Edits.

The JPG was extracted from the NEF file and given a very tiny bit of highlight control,

some gentle sharpening along the curve of the leaf, and a bit of burn on the

central hotspot which was quite mild in this particular photo.

I was shooting this to refresh my knowledge of the hotspot problem of the CO 60/4.0,

so I was as close as possible to the leaf, but with no extension or lens shading.

The "lines" of the fibrous, waterproof outdoor carpet slope downward

from left to right so I don't think anyone will possibly confuse them with PDAF banding.

f/8 for 1.3" @ ISO-3200

10_leaf_hiLite_htSptFix.jpg

 

 

Raw Composite from Raw Digger

No white balance has been applied to this version.

No edits were made either.

The hotspot is now faintly visible in the center of the shot.

The raw false colours look quite normal. That is one of

my long, grey hairs which has floated into the photo on the left.

I felt obliged to clone it out in the first photo.

20_leaf_rawComp.jpg

 

 

Raw Digger RGB Render

This shows how Raw Digger might apply a white balance to the file.

Saturation is low in Raw Digger as it is not intended to be a converter.

But you can see there is dichroic (or other??) discolouration on the periphery.

30_leaf_rgbRender.jpg

 

 

Raw Digger RGB Render - Saturation Pushed

I brought the preceding photo into Capture NX2 and pumped

up the saturation way beyond what we would ordinarily do

in order to illustrate the discolouration. This pumped saturation also happens to

illustrate that the white balance step failed, but that's no fault of Raw Digger.

This mess happened because the file is badly discoloured and Raw Digger could

not "decide" what should be considered a neutral area. (I am oversimplifying.)

We would expect this old leaf to have almost no false colour in UV except for

perhaps some bits of blue-gray. Instead we see blue, blue-green, green

and some greenish-yellow on the leaf with pinkish-orange on the edges of the frame.

What a mess! Pretty I suppose, if you like a rainbow effect!

For the record, I had less success removing the discolouration from this

file than I did in my preceding example (the Jonquil, above).

40_leaf_rgbRender_satPush.jpg

 

 

Unresized crop from leaf photo showing faint banding in undetailed background.

Again I have to wonder why I did not get such strong banding as Birna got?

50_leaf_asShot_100crop1.jpg

 

 

Unresized crop from leaf photo showing that high-frequency detail masks banding.

60_leaf_asShot_100crop2.jpg

Link to comment

I am definitively not keen on any Olympus. Was hardly convinced by the image quality of the E M1.2 I currently use in an underwater housing and really wish for a better system.

 

Do you still have the Em1 mk2? If so can you mount a uv lens and a Baader venus filter on it and see what exposure settings you get?

 

The Em1 mk2 has the sony 20Mp sensor, which is the first for Olympus to have an AR coating on it. I wonder if it cuts down on the stock camera abilities to see into UV.

 

I got a Em5mk2, it has the 16Mp Sony sensor. I am keeping it stock, but it is 3 stops less sensitive using my 390bp25 filter and 6 stops less using my 370bp15, than my full spectrum converted Em1mk1. This isn't as good as my Em1 when I first got it, as it was more UV sensitive. Of course this difference may be due to different dust shakers and on sensor uv/ir stacks. But I now have a feeling the Panasonic sensors are more sensitive to UV than the Sony sensor. I also get this from the Astrophotography, indications looking at the ZWO 1600mm with Panasonic sensor verses othe ZWO with Sony sensors. They complain about the lack of AR coating on the Panasonic sensor though.

 

So the Panasonic S1R may be an interesting camera, as it seems to have the newest Panasonic sensor. The S1 may have the older 24mp Sony sensor. But your correct, they are about 1000g, which is over my comfort 800g limit for a camera. So they will feel heavy.

Link to comment

Bright sunshine, UV-Nikkor + Baader U on the Olympus EM1.2: f/4.5, 400 ISO, 1.6 sec. The output is pale and low in contrast, with washed-our colours to indicate there isn't that much UV going through the stock sensor. Compare to the corresponding UV exposure with the D3200: f/11, 100 ISO, 1-2 sec. Thus the Oly response is down approx. 5 EV compared to my Nikon workhorse using the same filter and lens. The full-spectrum D600 is within 0.5 EV of the D3200. Both Nikons have ~1 EV less response than my old D40X (internal Baader U), but I haven't got the latter camera any longer. Its lack of LiveView and the 10MPix files became too restrictive. Still miss it on occasion though.

 

Perhaps the Mk.1 would be a better choice? but frankly not that interested in building yet another system.

 

(The EM1.2 is kept for its dedicated underwater use. Will keep for the time of the project)

Link to comment

Thank you Birna,

Yes that doesn't sound good. I am new to Sony sensors on Olympus cameras, as my E3, E510 and Em1 are all Panasonic sensors. The Em5mk2 is my first Olympus with a Sony sensor.

I will have to do a direct comparison between my E3 and Em5mk2. The E1mk1 had similar UV response as my E3 does when I got it. I may just be remembered wrong. But it was a hard decision to make to get it converted as I was able to hand hold UV shots with it, using the BaaderU and its IS of course.

 

Also some settings that I use with Olympus cameras that may help. In the menu I set "noise reduction" to on. This always collects a dark frame after the exposure and subtracts it in single shot mode. If you don't want it use multi shot mode as only works in single shot mode. I also set the "noise filter" to off or lowest. This kills detail and doesn't do a good job of noise reduction.

Link to comment

At this point in time, it might be prudent to examine what the modified Z6 actually can deliver -- after necessary mitigating measures are taken in the workflow.

 

All examples are taken with a UV-Nikkor 105 mm f/4.5. I availed myself of two of them, one with and the other without Baader-U mounted in front. My full-spectrum D600 and the Z6 were used with the first and the D3200 with the last one (that camera has the Baader-U internally). Using two lenses allowed for fast swapping of the cameras on my tripod. As the setting was broad daylight, the camera exposures were set to an average for all three of them and the captures taken within a few minutes.

 

The UV-Nikkor was set to f/8, exposure time 1.5 sec. (100 ISO) which incidentally was a tad rich for the D3200 and D600, but within acceptable limits. The lens on the FTZ adapter on occasion indicated a light leak, which later was found to be related to the aperture ring around the aperture coupling prong ('rabbit ears' in Nikonese). For now, just ignore the signs of green flare caused by this leak of visible light. The F-mount cameras were unaffected by this issue.

 

First, Z6 (modified) 100 ISO. Run through Topaz Denoise 6.01.

 

T201904150372_100ISO_Z6_Topas_Edit.jpg

 

Towards the very high ISO settings, UV false colours became more muted. Here an example with ISO 25600. Also noise-reduced through Topaz although the inherent noise of the high ISO more or less covered the signs of banding on its own.

 

T201904150382_ISO25600_Z6.jpg

 

My workhorse UV camera, Nikon D3200 with internal Baader U (tripod was stationary thus the angle of view is smaller for this DX camera).

 

T201904153754.jpg

 

Finally, the broad-band D600.

 

T201904156895.jpg

 

It is obvious that for all cameras we arrive at very near the same UV false-colour rendition, although for the Z6 this was only achieved after some additional data massage. This camera also exhibits some signs of colour vignetting across the frame, not related to the occasional light leak described earlier.

 

Despite the very obvious attractions of having an FX camera with good EVF easily focusable in UV (outdoors at least) and cutting-edge high-ISO performance, I have decided to return the camera for a refund.

Link to comment

Birna,

I think your making the correct decision here. However, I do wonder if this is a Z6 problem or a Lifepixel problem. Maybe you should message Dan Kuespert, since he got his Z6 from Kolari and see how it compares.

Link to comment

It seems appropriate to mention that Topaz have just released a new de-noise AI (whatever that means).

 

Dave

Link to comment

Lest we forget the high-ISO performance of the Z6, here is the 51200 ISO version of the previously posted test shot (with Tussilago). At this elevated ISO setting, the banding is gone as in being 'drowned' by residual image noise. Thus a de-banding operation is not required.

 

T201904150383_ISO51200_Z6.jpg

 

Do note that the light leak through the aperture ring of the UV-Nikkor lens is much more visible at this high ISO rating. The fix is easy, just put a piece of black tape over the aperture ring where the coupling prong ('rabbit ears') is located. Food for thought though I could not trigger a similar flare with the D3200 and D600, even with the UV-Nikkor attached via the PN-11 extension tube.

Link to comment

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...