The GetDPI Photography Forum
Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!
Hasselblad 100C and 35XL
- Thread starter diggles
- Start date
rdeloe
Well-known member
Here you go... I started a thread over at DPReview to get some of those folks on the case. In the starting post I provide a link to four files, two that Warren gave me permission to put up, and two of mine. Mine are with the GFX 100S, and Warren's are with the CFV 100C. Both are 15mm rise in landscape, and both have LCC.
www.dpreview.com
Hasselblad CFV 100C banding during shift with a symmetrical lens: Medium Format Talk Forum: Digital Photography Review
Expert news, reviews and videos of the latest digital cameras, lenses, accessories, and phones. Get answers to your questions in our photography forums.
www.dpreview.com
usm
Well-known member
I hoped that the "Remove checkerboard patterns" does the trick, but it doesn't. Also C1 can't do anything against.
I was in contact with HB regarding some Phocus issues with the "Remove checkerboard patterns" and thy did some updates in the last 2 Phocus versions. There is an update coming. Maybe they already know this Problem.
@diggles
As I don't have a CFV100c I can not make my own files. I can't send HB your files without permission.
I was in contact with HB regarding some Phocus issues with the "Remove checkerboard patterns" and thy did some updates in the last 2 Phocus versions. There is an update coming. Maybe they already know this Problem.
@diggles
As I don't have a CFV100c I can not make my own files. I can't send HB your files without permission.
diggles
Well-known member
The camera in the article has a SK lens mounted to it. I'm not sure which one though. At the end of the article, they provide an email address to send your questions to. I reached out with our findings in this thread. I also asked what lenses they have tested and if they have seen any similar issues.Alpa did a blog write up but I haven't had time to read it, idk if they said anything about it or not:
NEW Hasselblad CFV 100C on ALPA cameras | ALPA Log
A Trifecta of Imaging Possibilities and a Revolutionary Digital Back for Technical/View Camera; Publication: Feb 09, 2024; Update: Apr 05, 2024www.alpa.swiss
Since Alpa no longer sells the SK XL lenses I doubt they would've tested them. At the end of the day I'm not sure it's fair to expect HB to have tested the SK lenses either since they are no longer made, but idk if the same issues pop up with Rodenstock lenses. We need to keep in mind, at least regarding most SK lenses, that they are legacy and no longer produced so there is a risk of lack of manufacturer support even with continued use and as technology develops. It's definitely worth contacting HB support and letting them know. Rodenstock should be a different story.
I couldn't get a tech cam demo from HB when I was interested in the H6D-100c (even after I bought the accessories needed to pair an H5D back with a tech cam out of my own pocket), and I'm not sure they are overly concerned with that market based on my experience and interaction with them regarding tech cams. From my own experience using tech cams in the field, I'm pretty convinced at this point that GetDPI is a large part of the tiny population of photographers and the only corner of the internet that cares about tech cams, or even knows what a tech cam is....aside from the folks at Arca, Cambo, Alpa, etc., maybe the Chinese market.
All that said it's still a bummer though. Seems it can be corrected in software so maybe they'll update Phocus, but just be careful with extreme movements in the meantime.
The more people they hear from, the more likely they are to look into it.
I chose to not cold-call, but deliver the message via my dealer. She's going to address this with Hasselblad. Maybe that's also an option for you?
In the end, it was for me the main purchase argument of the CFV100c to be able to use my Schneider wide angles.
i just rechecked a test shot i did a few weeks back and can confirm that it is an issue with the SK 43XL as well (unfortunately!!!)
you can see for yourself here (LCC included):
www.dropbox.com
it's most evident in the upper left corner, visible as dark horizontal stripes.
this was shot on an Rm3di with CVF100c and SK 43XL at f11, focused at infinity (decentered on the camera -5mm vertically and 8mm to the right)
i also did test shots with a Rodenstock Sironar 90 (symetrical design) and pushed it to the max shift position of the Rm3di and have zero "stripes" to report
you can see for yourself here (LCC included):
Dropbox - File Deleted - Simplify your life
www.dropbox.com
it's most evident in the upper left corner, visible as dark horizontal stripes.
this was shot on an Rm3di with CVF100c and SK 43XL at f11, focused at infinity (decentered on the camera -5mm vertically and 8mm to the right)
i also did test shots with a Rodenstock Sironar 90 (symetrical design) and pushed it to the max shift position of the Rm3di and have zero "stripes" to report
rdeloe
Well-known member
That is unfortunate. The lines are a bit harder to see than on Warren's samples with the 35 XL, but I see them.i just rechecked a test shot i did a few weeks back and can confirm that it is an issue with the SK 43XL as well (unfortunately!!!)
you can see for yourself here (LCC included):
Dropbox - File Deleted - Simplify your life
it's most evident in the upper left corner, visible as dark horizontal stripes.
this was shot on an Rm3di with CVF100c and SK 43XL at f11, focused at infinity (decentered on the camera -5mm vertically and 8mm to the right)
i also did test shots with a Rodenstock Sironar 90 (symetrical design) and pushed it to the max shift position of the Rm3di and have zero "stripes" to report
Paul Spinnler
Well-known member
It seems that it has to do with the angle at which the light hits the sensor plane (as with the tiling) because most likely Hasselblad has PDAF circruitry above the sensor leading to small height differences on the surface above the sensor which becomes evident the more steep the angle is at which the light hits the sensor.
Three things will affect this:
1) How wide is the lens in general
2) Retrofocus or symmetrical
3) How much shift
Long symmetric lenses are ok, short symmetric ones are the most challenging ones, especially when shifted a lot.
Rodenstock wides shifted a lot would be interesting to see, namely the 23, 32, 40 HR at 15mm+ shifts.
I wowuld assume anything on the longer side, 60mm+, even symmetric, won't be a problem.
Looks like a post processing solution needs to come along ... and that P1 is still the way to go for SK.
Three things will affect this:
1) How wide is the lens in general
2) Retrofocus or symmetrical
3) How much shift
Long symmetric lenses are ok, short symmetric ones are the most challenging ones, especially when shifted a lot.
Rodenstock wides shifted a lot would be interesting to see, namely the 23, 32, 40 HR at 15mm+ shifts.
I wowuld assume anything on the longer side, 60mm+, even symmetric, won't be a problem.
Looks like a post processing solution needs to come along ... and that P1 is still the way to go for SK.
If the phenomena is related to the PDAF array (and it probably is), it's a surprisingly common problem. Under the right (or wrong) circumstance it appears that most sensors with PDAF circuitry can be made to exhibit linear artefacts. It's exacerbated with a digital back like the CFV-100c because you can stress it a lot pretty easily, e.g. with shifted W/A symmetrical lenses. In some ways, it's impressive just how unobtrusive the lines (in the examples above) are given that H/B's main focus was probably on its XCD line of lenses rather than making the DB work well on tech cams with god knows what stuck on the front in the way of optics. It's not all doom and gloom though, as seeing how well something like Raw Therapee can deal with the problem for sensors for which it has the PDAF array data (if anyone reading this wants to send them a few RAWs maybe they'll include the CFV-100c in their next update), it's reasonable to assume that H/B can mitigate the issue in software pretty well (if so inclined). Maybe not 100%, but good enough that we all don't have to run out and drop $50k on a 150MP back (if only!)
rdeloe
Well-known member
Have a look at the post by user Bokesan here: https://www.dpreview.com/forums/thread/4748591#forum-post-67549207 It's the 11th one down. He processed a file in Affinity with HDR. This reveals the problem in black and white, literally. There's the added problem of a vertical tiling line.
diggles
Well-known member
The vertical line like looks like 'normal' sensor tiling.
rdeloe
Well-known member
Normal is good!
Chris (bokesan) had to work the file hard to make it look that bad. His comment to me on the side was that he hadn't noticed the pattern until I put the post up that linked to your sample.
Yeah, if you crank everything up you can see it, but it's not an issue in 'normal' use.
Back to the matter in hand ... opening Warren's files in Phocus and Camera Raw, the horizontal banding (which appears to be around every 210px or so) can easily be rendered visible by cranking stuff like clarity to the max. Doesn't seem any way to mitigate the banding in Phocus (that I could find), even by using the LCC.
Onto Raw Therapee. Right off the bat I could see that the banding was less obtrusive at all the default import settings. And there are a lot of settings in RT ! I've only used it occasionally - definitely no expert. It would be great if RTs 'PDAF' filter worked on HB CFV-100c files straight out of the can, but my understanding is that it doesn't ... yet.
Regardless, after a lot of playing around I found that I could get the banding to disappear (to my weary eyes) via Raw > Flat Field > File (select the LCC) > Blur Type: Horizontal > Blur Radius: 0
Jim
BTW I saw Jim Kasson - who knows infinitely more about this stuff than I do - post over on DPReview that the banding (@ 210px spacing) is unlikely to be related to a PDAF array (quote: 'Still not the PDAF row spacing'). Curious. I wonder what could be causing it ? Maybe it can be fixed in firmware ? If so, that would be ideal.
corvus
Active member
Yesterday I had a very good conversation with a German dealer who was officially named by Schneider Kreuznach for support at my request (Jos.Schneider GmbH itself no longer provides information on the lenses for view cameras that are no longer produced).
My enquiry initially related to a completely different topic. At the end of our conversation, I briefly told him about the problem with the CFV100 with 35XL, which the professionals here at getdpi are discussing. He explained to me in great detail that he was already familiar with such discrepancies from earlier cameras and that it depends a lot on the combination of lens-shift-sensor and so on. He doesn't see it as an unsolvable problem and assumes that it will probably be solved with a firmware update ...
I'm keeping my fingers crossed for all CFV users!
My enquiry initially related to a completely different topic. At the end of our conversation, I briefly told him about the problem with the CFV100 with 35XL, which the professionals here at getdpi are discussing. He explained to me in great detail that he was already familiar with such discrepancies from earlier cameras and that it depends a lot on the combination of lens-shift-sensor and so on. He doesn't see it as an unsolvable problem and assumes that it will probably be solved with a firmware update ...
I'm keeping my fingers crossed for all CFV users!
One curious thing. When applying the H/B LCC (supplied by Warren) to itself in Raw Therapee, you get the expected uniform image - basically something that looks like the sheet of white plastic used to capture the LCC in camera (as you'd expect). You also see this in C1 when applying an LCC to itself (but not a H/B LCC 
). But in Phocus when I try the same thing (and I'm new to Phocus, so forgive me if I've missed something), while the image is generally uniform, it does nothing with the banding. It's like it doesn't see it. Oddly, going back to RT, after selecting 'horizontal' as the blur type, you can switch back to 'area', 'vertical', or 'vertical + horizontal' and the banding seems to remain hidden; it's like RT finds the file a bit confusing as well. I wonder if this issue ultimately has to do with the way the file is being written out in camera. Hmm. (BTW, FYI I've been using the AMaZE demosaicing method in RT throughout).
). But in Phocus when I try the same thing (and I'm new to Phocus, so forgive me if I've missed something), while the image is generally uniform, it does nothing with the banding. It's like it doesn't see it. Oddly, going back to RT, after selecting 'horizontal' as the blur type, you can switch back to 'area', 'vertical', or 'vertical + horizontal' and the banding seems to remain hidden; it's like RT finds the file a bit confusing as well. I wonder if this issue ultimately has to do with the way the file is being written out in camera. Hmm. (BTW, FYI I've been using the AMaZE demosaicing method in RT throughout).I played with your picture for while in RawTherapee 5.9.i just rechecked a test shot i did a few weeks back and can confirm that it is an issue with the SK 43XL as well (unfortunately!!!)
you can see for yourself here (LCC included):
Dropbox - File Deleted - Simplify your life
it's most evident in the upper left corner, visible as dark horizontal stripes.
this was shot on an Rm3di with CVF100c and SK 43XL at f11, focused at infinity (decentered on the camera -5mm vertically and 8mm to the right)
i also did test shots with a Rodenstock Sironar 90 (symetrical design) and pushed it to the max shift position of the Rm3di and have zero "stripes" to report
Good news is that I could get rid of those stripes if I can see. Bad news is that I cannot get acceptable colors or tones with LCC. I get better colors without LCC, still not very good as I guess it should be possible. Also almost invisible stripes are left, very weak ones. I think my version (5.9.xx) of RT does not support CVF100c yet. Version 5.10 is already published.
You can meet RT ( and ART ) developers at https://discuss.pixls.us/c/software/rawtherapee/14 discussion forum and ask them for help.
Last edited: