Like I said, the banding was likely caused by electronic first curtain or electronic shutter. Seeing that they were shooting long exposure Astro at high ISO, this potentially introduces another set of problems for that use case.
Most/all sensors starts showing banding when you take a lot successive long-exposure shots after each other. So be careful doing fast conclusions from the given dpr samples featuring lot of Neowise shots.
Stig Nygaard wrote:
Most/all sensors starts showing banding when you take a lot successive long-exposure shots after each other. So be careful doing fast conclusions from the given dpr samples featuring lot of Neowise shots.
Then load them in Photoshop/Lightroom. Dynamic range looks pretty darn good to my eye. Maybe a touch more luminance noise than Sony A7rIII At base ISO but a bit early to tell.
These are all the same photo taken at ISO12800, 1/25 second, RF 50mm @1.2. All edited RAW photos in lightroom classic using the adobe beta DNG converter.
First photo at baseline exposure with noise reduction
Second photo boosted 1 stop with noise reduction
Third photo boosted 2 stops with noise reduction,
Fourth photo boosted 3.5 stops with noise reduction,
Fifth photo boosted 3.5 stops without noise reduction.
I can see a small amount of banding at 3.5 stops overexposed in post using the BETA adobe DNG converter (not sure if that is contributing), which improves with noise reduction. This is no where near as bad as the 5D2 low light areas at high ISOs and the R5 is totally liveable. That said, if you are screwing up your photos by 3.5 stops at night then perhaps you should look into spending lessons on photography instead of buying a $4k camera.
TeamSpeed wrote:
The banding is there in all 3 images actually. Even just a mild post processing can show it.
You and i have very different definitions of mild post processing if +2 EV is mild. If I stare at the +2 ev for a while I can convince myself there is some mild banding. I just updated it with +1 and I don't see any there.
In any case, hoping the banding that is there is potentially sorted out when adobe releases a version of PS/LRC that supports the native files. If not there, then hopefully canon can fix this in a firmware update. Even if that fails, I lived with the significantly worse banding on the 5D2 for years and sold plenty of performance photos taken at high ISOs without a customer complaining. I think I'll be ok.
I can see the banding in the dark sections on the +2 as posted, and I mean that with just a bit of post processing that image to get rid of noise and increase clarity, etc, you can really start to see the banded blotchy noise in what appears to be a speaker or whatever in the image.
I am sure others can see it too, just depends on their type of monitor and whether it is calibrated or not, etc, but I can make it out as is (section of green and also orange).
It's not bad, probably better than the 5D4 at this point, considering the 45Mpx sensor, it is easily mitigated. But also as you point out, if you are shooting at 12800 and actually do miss 2-3 stops and try to correct for it later, you shouldn't expect great things.
Firmware fixed the banding on my EOS R. My 5D Mark IV never got a firmware to fix it, but the R did.
TeamSpeed wrote:
No firmware cannot fix banding. NR can clean some of it up, but there is always a residual effect that can still often be seen after post processing.
Unfortunately, I can confirm that the camera has visible banding at 12,800 and higher. It's certainly disappointing. It may also be in relation to the sensor building heat because at the point I took my shots the camera had already been on and shooting for about an hour and I was in the menus the entire time fiddling with stuff.
I clearly see quite offensive banding at 6400. Maybe I'm just sensitive, but the R doesn't do that. It's not as bad as my 1DS was at any ISO, but it's certainly off-putting. This is when shooting "my neighborhood owls", which is pretty much always at 1/30s, 6400 or 12800, f2.
I hope the owls are around again today, otherwise I guess I still can shoot the trees.
Stan, was that owl photo mechanical or electronic shutter? Have you compared both in similar conditions and are the results the same? FWIW, I too can clearly see the banding in that photo.
rscheffler wrote:
Stan, was that owl photo mechanical or electronic shutter? Have you compared both in similar conditions and are the results the same? FWIW, I too can clearly see the banding in that photo.
The ones from yesterday were electronic first curtain, mechanical shutter. Today I will go there with pure mechanical and deflicker turned on, just to be sure. Stay tuned, owls will start flying in 30 minutes
Visceral Image wrote:
Firmware fixed the banding on my EOS R. My 5D Mark IV never got a firmware to fix it, but the R did.
Yes I guess some types of banding might be fixed by software removing it for you, but some consider blotchy color sections to be a form of banding and that can be caused by hardware. I had a 5d2 and 1d4 with that and try as Canon may, they couldn't fix it after I sent it in.
rscheffler wrote:
Stan, was that owl photo mechanical or electronic shutter? Have you compared both in similar conditions and are the results the same? FWIW, I too can clearly see the banding in that photo.
Well, since the chicks grew up it's been getting pretty hard to get the owls in front of the lens, so I just took a few comparison shots of the same tree. In a double blind test I really don't think that I could tell full mechanical shutter vs. electronic 1st curtain apart. Overall disappointing, but you can't win them all.
While I doubt it will make a difference, I do wonder if full electronic shutter produces the same results. I.e. since it's a 12-bit vs. 14-bit file, whether that might mask or enhance the banding?
rscheffler wrote:
While I doubt it will make a difference, I do wonder if full electronic shutter produces the same results. I.e. since it's a 12-bit vs. 14-bit file, whether that might mask or enhance the banding?
Hmmm....come to think of it, I need to see if I was using high-speed shooting when I tested ISO. If the camera is forcing itself into 13-bit in that mode it could impact the way it handles increased sensitivities. I need to learn the ins/outs of the camera's limitations.
But also as you point out, if you are shooting at 12800 and actually do miss 2-3 stops and try to correct for it later, you shouldn't expect great things.