Banding in photos (IR converted camera)

Messages
385
Name
Scott
Edit My Images
Yes
I've had a Nikon Z5 converted to fill spectrum (primarily for infrared, but being flexible). I seem to have an issue with horizontal lines appearing in the first proper IR photos I've taken with it (using a Hoya R72 on the Nikon 24-70 F2.8)

The example is post conversion to B&W and a liberal application of contrast (not how I'd normally process - exaggerated a little here to highlight the issue), however I can see it very faintly in the RAW image. (These are both crops from a vertically held pano - the big one at 100%, hence lines are vertical in this image). The issue isn't obvious where there is any texture, however will probably show up clearly on plain images, cloudy skies, etc like this. Not had this on any other IR converted cameras I've had, can't see why the Z5 would be any different to a D7100 or D90 in that regard, and can't see why full spectrum + filter would result in this?

banding example small.jpg

I'm slightly concerned this might be a hardware issue and will be looking back at other photos I've taken with the camera. However, anyone seen this issue before anywhere? Any suggestions?

banding example crop.jpg
 
Not sure if this applies to your camera, but did you try disabling electronic shutter?

I have a Sony a7riii and it produces banding in certain situations when electronic shutter is used.
 
I'm not saying the electronic shutter isn't the cause but the sort of banding that causes is under artificial flickering lighting. The banding is caused by the shutter / readout from the sensor being too slow to capture the scene before the flickering lights cause the lighting to change, thus you see lighter and darker bands across the image. The example above doesn't look anything like that and presumably there's no flickering artificial lighting anyway.

Tried Googling?

I found this site which seems to be saying that some sensors show lines, they don't look exactly like the example above but maybe it's a similar issue showing slightly differently on a slightly different sensor?...

 
Last edited:
Hi
just an opinion...
I have the leica m10, and when I crank the iso up high I also see banding.
Looking at your picture (and not having any info on iso shutter etc) I am guessing its a fairly high iso, 6400 maybe. Perhaps the sensor bands with higher iso.
I'd try taking several pictures with various isos and compare them to see whether it is the iso, and if so, when it becomes distracting...
 
Welcome to PDAF and phase detect striping. It's a consequence of putting the phase detect pixels on the sensor. Superior DSLR camera's do not do it.

I'd ease up the contrast and hope for the best. Or use the superior D850 with the inferior F mount glass.
 
Jim Kasson posted about this here...


He says it's a Nikon Z issue caused by Nikons efforts to prevent striping, it's worth wading through for anyone with these cameras.
 
Welcome to PDAF and phase detect striping. It's a consequence of putting the phase detect pixels on the sensor. Superior DSLR camera's do not do it.

I'd ease up the contrast and hope for the best. Or use the superior D850 with the inferior F mount glass.

Not quite. See the above.
 
My A7III does it too, though not that badly. It's the PDAF sensors as Steve said.
Any high structure, high contrast image will do it with on sensor phase detect camera in theory.

I wonder what the affect of stopping down has. In the same way sensor dirt/marks show at F8 a bit and F22 a lot and F2.8 probably not.

Will wide apertures hide the issue or because PDAF is in the internals of the sensor it won't matter.

Reason I wonder is for portrait/wedding/wild life togs it may not matter a jot but for those who shoot stopped down it may matter heavily. For landscapes where there's regions of blue sky and cloud I think this could be a serious pitfall. In time I imagine it'll improve but imho mirrorless still has a way to go before I'll sink my £10k in

Anyway, as I have a proper camera with a mirror it won't show on mine.
 
Thanks all, will look into that - reassuring that it's not a conversion fault or broken camera, not so reassuring that it's a fundamental issue with the sensor type vs the type of photography I do! :ROFLMAO:

Just for completeness, the photo was taken at ISO100, F7.1, but yep I've pushed the structure and contrast (obviously)

I might have a play around next time I go IR Steve, just to see whether aperture makes a difference. I've still got my D7100 IR, so may do some comparisons. Will also check whether this appears in either full spectrum or UVIR - I've not got any other IR filters yet.



In terms of the camera, I'm still in semi evaluation mode on my new setup. The Z5 conversion is kind of an experiment as I think I've mentioned before, and is why I went for a Z5 for conversion rather than straight into a Z7. It's not that I can afford to lose a grand, it's that I couldn't afford to lose 2-3! This was the Z5s first proper outing in IR - I used it for colour video last week, and it gave as good an account for itself as the Z7 did, bar user error (I don't do video - this was my first time, so maybe I'm easily impressed). But for IR, the weather and locations just haven't been suitable.

For this walk, I had no real intention of shooting IR as the weather was a bit rubbish, so I left the house with the UVIR filter on the front and the R72 in the back pocket (and the Z7 at home...). Might seem odd, but having the ability to do that is still really quite exciting and ticks one of my evaluation criteria perfectly. Seeing IR through the viewfinder, and not having to worry about focus issues is also truly amazing. Sure, I know I could do all that on the back screen, but.... I never did. I'm still getting used to the electronic viewfinder, and the lag is irritating. I expect the viewfinder to be instant and it's not. I'm very much a viewfinder user, and I doubt that'll change much, but the plus side is clearly being able to see exactly what I'm photographing. Other issue I've encountered is I'd bought the Z5 with the 24-200 Z, which it turns out hotspots terribly - worst I've ever seen! Fortunately the 24-70 of the FTZ works perfectly for IR, so the 24-200 is on the Z7. Yes, I know.. That alone is causing me to question my new setup, though the 24-200 isn't a bad lens to be fair.

But in terms of weight and feel, the Z7 / Z5 combo is definitely, obviously lighter and more compact than the D810 / D7100 combo and feels great to use. At the moment, I think the benefits of mirrorless are outweighing the drawbacks - it's whether this PDAF issue is fatal for my type of photography or not. Surprisingly, actually having 2 modern bodies that shoot (or have the ability to shoot) colour / high quality video has proved invaluable over the last 2 weeks (and my D810 came out to play for stills again too!) so maybe there are other reasons to maintain a 2 (or more) body setup. Not sure I want to go back to a big body yet though Steve, sorry! ;) and sounds like the D8X0 line might be getting a new body soon?
 
My A7III does it too, though not that badly. It's the PDAF sensors as Steve said.

The Z series are the worst for it - but it's a potential issue with any camera with on sensor phase detect AF. It's much less of an issue on an A7RiV for instance - but not at all on the Nikon D850

Thanks all, will look into that - reassuring that it's not a conversion fault or broken camera, not so reassuring that it's a fundamental issue with the sensor type vs the type of photography I do! :ROFLMAO:

If JC is correct I think it's more the way Nikon have tried to help and cocked it up.

According to the link, if correct and applicable to this Z camera, it's caused by Nikons in camera processing to limit PDAF striping which is apparently a different issue. He doesn't mention the Z5 so this may not be the case but if it is it does look like Nikon have gone in a different direction to Sony and anyone else using a similar chip at least with the cameras named in that link and perhaps a firmware option to either do in camera corrections or not would be a better way forward.
 
Last edited:
If JC is correct I think it's more the way Nikon have tried to help and cocked it up.

According to the link, if correct and applicable to this Z camera, it's caused by Nikons in camera processing to limit PDAF striping which is apparently a different issue. He doesn't mention the Z5 so this may not be the case but if it is it does look like Nikon have gone in a different direction to Sony and anyone else using a similar chip at least with the cameras named in that link and perhaps a firmware option to either do in camera corrections or not would be a better way forward.

This looks like PDAF striping though on this image.

PDAF banding would only really appear in an image with 3-5 stop shadow recovery and in the area's of shadow at 100% preview. That level of shadow recovery is pretty extreme under exposure and probably not applicable to this image and it's processing.

I suspect the IR conversion, or the processing has managed to reveal the PDAF on the sensor - hence these uniform lines most visible in the least textured part of the image - the sky.

It's just not an issue of the SLR type digital camera and I think this could be a real issue for landscape shooters. I'm glad I've got my 645z ;)
 
Last edited:
Yeh, have read it now. Z5 obviously wasn't out when the original article was written, but seems it could be the issue, and may well be IR / full spectrum conversion has exacerbated the issue. and typically, I'm a landscape type shooter....
 
I had been thinking about getting a Z5 for IR conversion. From this it looks like I may have to stick with the Fuji. Have you tried using Raw Therapee's line filter to remove the banding?
 
I had been thinking about getting a Z5 for IR conversion. From this it looks like I may have to stick with the Fuji. Have you tried using Raw Therapee's line filter to remove the banding?
I haven't - didn't know about it till you mentioned it! :D Will have a look and see how I get on....

EDIT...
Having never used Raw Therapee before, I seem to be able to get rid of the banding using the Line Filter, set to 25 and Horizontal in the Direction drop down. Using the PDAF Line Filter tick box doesn't seem to do anything, and neither does the suggested "Horizontal only on PDAF rows" in the direction drop down. (It may be something is different in the Z5 compared to the Z6 / Z7 that the suggested fix doesn't work, but as Horizontal alone seems to work OK, not sure it matters)

The problem now is whether I can get the corrected RAW files with the PP3 sidecars into Lightroom or Photoshop so I can merge into a panorama. At the moment I'm guessing... no....
 
Last edited:
Double post. See below.
 
Last edited:
This looks like PDAF striping though on this image.

PDAF banding would only really appear in an image with 3-5 stop shadow recovery and in the area's of shadow at 100% preview. That level of shadow recovery is pretty extreme under exposure and probably not applicable to this image and it's processing.

I suspect the IR conversion, or the processing has managed to reveal the PDAF on the sensor - hence these uniform lines most visible in the least textured part of the image - the sky.

It's just not an issue of the SLR type digital camera and I think this could be a real issue for landscape shooters. I'm glad I've got my 645z ;)

If the link above is correct it's not pdaf banding as such but Nikons in camera trickery making the effect worse.

Ideally they'd have a firmware fix and provide an option to turn the in camera processing on or off. That might help.
 
Last edited:
Might have to revisit this.... I seem to have dropped and damaged my Z5. It's off getting fixed, but has prompted me to ask myself if I want to consider converting my Z7 (and upgrading my colour camera to a Z7II or... gasp....Z8...) but didn't know if the PDAF banding on a Z7 would be better, worse or just the same as with the Z5? Anyone know??

FWIW, after 2 years I've stuck with the Z5 (though lost my photographic mojo somewhat so not been using it so much), I've found that Topaz Denoise AI seems to do a pretty good job of minimising the banding. I'm sure I've posted a thread in here somewhere, but it would have been ages ago. Got some example pics on my Flickr account, though Denoise has been updated since I did those so I'd assume it can only have gotten better.
 
Not directly connected, but could just be useful.

One of the easiest to convert (45min - 1 hour) and cheapest is the Panasonic G3.
Very light and easy to carry and great results. Only disadvantage to the above is it is only 16MP, but then it doesn't have the problem being contrast AF.

They are even available ready converted for a little over £100 (can find the seller who is very good if any interest)

It beats by miles any Canon I have converted.
 
As mentioned, on sensor PDAF banding. Liberal use of contrast and/or dehaze accentuate this.
 
Back
Top