arbitrage wrote:
Agreed, I just threw that jpeg into LR and cranked up exposure 3 stops and can't see any.
May be related to what shutter mode was used??
Although I also downloaded the DPR 6400 shots and cranked its exposure, played around with dehaze to extremes and don't really see the banding in that file either.
So maybe it really is only showing itself at the real high-ISO??
The black level in JPGs would be clipped relative to the raws so any banding is likely obscured/clipped as well when raising the shadows of JPGs.
roanjohnnyc wrote:
.........is it me or Canon sensors have been plagued with banding on high ISO? I remember this issue since the 20D days
There has been little-to-no banding in Canon's High ISO images for several generations of their cameras. Any banding was limited to low ISO and even that was mostly resolved.
arbitrage wrote:
Agreed, I just threw that jpeg into LR and cranked up exposure 3 stops and can't see any.
May be related to what shutter mode was used??
Although I also downloaded the DPR 6400 shots and cranked its exposure, played around with dehaze to extremes and don't really see the banding in that file either.
So maybe it really is only showing itself at the real high-ISO??
---------------------------------------------
Ralph Conway wrote:
This is 6400 only. In ISO 6400 the DPR files doen´t show banding, too.
When I had the 5D IV I had a similar issue but it was visible only when shooting full frame dark scenes (astro). It was showing even at ISO 100 when the file was pushed above 2 stops. It was hardly visible if the shot contained highly lit spots in comparison to the rest of the frame.
I still hope that whatever we're seeing right now it's just a firmware maturity issue and the production model is ok.
This is 6400 only. In ISO 6400 the DPR files doen´t show banding, too.
Maybe, you weren't as brutal
This is what I got:
LEFT: DPR and RIGHT: Cameralabs
https://i.imgur.com/cll1d3j.jpg
__________
EDIT:
__________
OK, so I'm not really sure what's going on...
Both files DPR and Cameralabs from the above, are shot with cameras running the initial version of the firmware.
The only variable that separates these two shots are RAW conversion (probably) and exposure time.
Hopefully, I receive mine in a week or so and I'll can do some extensive testing...
AmbientMike wrote:
A7r4 had banding, granted at a rather extreme setting, the A7r3 didn't have.
I just checked my near pitchdark shots i took from neowise @ ISO 6400 on the a7r4 and i found zero banding even after pushing 5 stops and playing with blacks and clarity and texture all the way up which would reveal the slightest hint of banding. Maybe you shot in hot conditions?
When I had the 5D IV I had a similar issue but it was visible only when shooting full frame dark scenes (astro). It was showing even at ISO 100 when the file was pushed above 2 stops. It was hardly visible if the shot contained highly lit spots in comparison to the rest of the frame.
I still hope that whatever we're seeing right now it's just a firmware maturity issue and the production model is ok.
Maybe, you weren't as brutal
This is what I got:
LEFT: DPR and RIGHT: Cameralabs
https://i.imgur.com/cll1d3j.jpg
__________
EDIT:
__________
OK, so I'm not really sure what's going on...
Both files DPR and Cameralabs from the above, are shot with cameras running the initial version of the firmware.
The only variable that separates these two shots are RAW conversion (probably) and exposure time.
Hopefully, I receive mine in a week or so and I'll can do some extensive testing...
I still can detect stronger banding in the sky at the DPR pic. Maybe it is just "enhanced" through the raw conversion. Or might be depending on exposure time/noise reduction settings?
Raw processors have to catch up with a new release. Those that didn't use DPP almost always saw lackluster results in the beginning after a release, we saw it with a number of APSC and FF offerings in the past.
If DPP has been updated to accept R5 raw files, then that is probably the best you can do in analyzing files right now until Adobe, etc all add any necessary support. Just a slight change in the file/sectional structures in the raw file could cause all kinds of issues with raw processors.
I just checked the (new?) DPR (the purchase and sale shop image No. 111) shot with ISO 12.800. In opposite to the night shot again I can not find any banding. The difference is again the exposure time (1/20th instead of 4 seconds). They are all out of camera jpegs. Maybe "automatic noise reduction for longer exposures" causes those results.
therealthings wrote:
I just checked my near pitchdark shots i took from neowise @ ISO 6400 on the a7r4 and i found zero banding even after pushing 5 stops and playing with blacks and clarity and texture all the way up which would reveal the slightest hint of banding. Maybe you shot in hot conditions?
Like I mentioned, I believe it was on the DPReview image comparison tool. Picking up shadows quite a bit.
I did not check the last days but this morning I saw that the R6 is implemented now in DPRs Studio shot comparison. I tried to set the link, but it did not work. Please look for it yourself.
I could not see any banding in any jpeg file up to Iso 102.400.
sorry ... in normal use.... i have yet to see banding in night or dark images from weddings.
I don't have to push that much because of exposure being pretty decent.
or i just messed up the shot completely.
i can make any camera do that above say 2400/3200 iso.
I am using ISO 6.400 for event shootings regulary. With my 6D I can go up to 12.800. Sometimes I used 25.600. I had to get the shot in dim light and at 1/250th. It is possible because you have to and can work on the noise only. But If banding occurs one just can delete the shot. And not trust those ISO settings.
This is why I am interested in the background, why some DPR R5 pictures show banding, but an other one and Gordon Laings sample pics at 12.800 do not. Concerning R6: No banding up to ISO 100K. And noise looks on par with Z6 and even better than A9 IIs to me at a first view.
It's very likely that these images were captured using Electronic First Curtain or Electronic Shutter - which can commonly create horizontal banding. It is my understanding that EFC can cause this a lot. I look forward to testing mine when it arrives.
Fingers crossed that once I get my hands on this thing, that there won't be banding issues.
I often shoot at 12800 or even 25600 with f/1.2 and 1.4 glass wide open (in bars etc), hence why the 28-70 f/2 is also on the order. I'm hoping for an extra stop of sensitivity at the same ISO, when downsampled to ~20 megapixel of my 1DX II. If that happens, we're in business. If it doesn't, then I'm going to be feeling serious remorse. If it's worse, then it's going to be returned.
Dpreview just added the R5 to its studio-scene comparison images and no banding is visible. Here's an ISO 25,600 low-light link I generated: https://tinyurl.com/y4brdrak
snapsy wrote:
Dpreview just added the R5 to its studio-scene comparison images and no banding is visible. Here's an ISO 25,600 low-light link I generated: https://tinyurl.com/y4brdrak
Thank you. Looks on par with Z7 and 7R IV to me, better than the S1R. And like expected the 20MP R6 beats them all easily.
Ralph Conway wrote:
Thank you. Looks on par with Z7 and 7R IV to me, better than the S1R. And like expected the 20MP R6 beats them all easily.
Yep, but when using comparison view in the tool so all camera samples are sized the same I don’t see any real difference. Would rather take the larger 45mpix files and work from there
.
Please look for it yourself.