Hasselblad 100C and 35XL

[Paul Spinnler]

Is this banding effect being caused by minute differences in the height of the sensor stack (from AF circuitry?) being highlighted by the angled light emerging from the lens?

This is my theory. The unevenness induced by the layer of AF circuitry impacts the amount of light hitting the sensor the more angled it comes in, ie with big shifts and especially with SK glass.

One would almost need to run a compatibility test for each lens and take note where the images start to display banding with lines.

Removing banding in Photoshop is a no-go workflow-wise... It is not cool to have to re-work every single image in a tedious multi step process because a) you need to first get distortion corrected in C1 and then b) banding in PS.

On the Rodie side you have the issue that you need to pass your images through two raw converters or PS with the Alpa plugin in legacy mode somehow to get proper distortion correction.

Surprising to see this problem.

 

[f8orbust]

The DB has a 'sensor calibration' option for use with the 907X and XCD lens - has anyone tried using it with something like the S/K 35mm prior to shooting (assuming you can) ?

 

[Paul Spinnler]

The DB has a 'sensor calibration' option for use with the 907X and XCD lens - has anyone tried using it with something like the S/K 35mm prior to shooting (assuming you can) ?

That would be amazing if it solved the problem!

But you would need to create calibrations for set shift positions ... and apertures ...

 

[f8orbust]

Does using the correct CF make any difference ?

(I mean not to the calibration, just in general).

 

[f8orbust]

If the locations of the actual pixels that are affected are the same each time (which they should be if related to something like the AF array), then I don't see why a firmware update couldn't resolve the issue.

 

[Paul Spinnler]

If the locations of the actual pixels that are affected are the same each time (which they should be if related to something like the AF array), then I don't see why a firmware update couldn't resolve the issue.

The effect varies by lens. You can’t fix that via firmware. There is a physical difference on the sensor which creates uneven light absorption that is dependent on the lens, the shift, etc.

CF also doesn’t solve it. Has nothing to do with it.

I think only better LCC algorithms can solve this in post processing. Somehow the C1 LCC algo doesn't catch sharp lines of uneven lighting which appear when you shift extreme with SK symmetrics (e.g. the famous tiling problem); since it can be corrected manually in PS there's nothing speaking against better LCC implementations getting also rid of sharp horizontal lines, tiling, etc.

Since this is a CFV100c problem, I wouldn't bet on C1 re-working their LCC algos to fix this. This would need be remedied natively in Phocus.

 

[vjbelle]

A quick test with my 60XL and 100 II with 30mm rise shows no lines at 200% magnification. I can't go any wider but thought I would post this info.

Victor B.

 

[f8orbust]

The effect varies by lens. You can’t fix that via firmware. There is a physical difference on the sensor which creates uneven light absorption that is dependent on the lens, the shift, etc.

CF also doesn’t solve it. Has nothing to do with it.

Let's not forget that the jury is still out on what's actually at the root of all this. If it's a physical obstruction, then no, firmware or using CFs won't help, but only time will tell what's actually causing this problem. Maybe it's just a fabrication issue with a certain production run? Who knows? BTW, when Sony build the chips they mask areas for PDAF - it would be interesting to see where those are exactly.

Whatever, if it can't be fixed via firmware, then LCC + an update to Phocus should be able to deal with it. If H/B get their act together and jump on this before it 'catches fire' on the internet, then I don't see this as a deal breaker at all. Even now, I doubt anyone would even notice it other than via pixel peeping large shifts in areas of broadly continuous tone.

 

[Paul Spinnler]

Let's not forget that the jury is still out on what's actually at the root of all this. If it's a physical obstruction, then no, firmware or using CFs won't help, but only time will tell what's actually causing this problem. Maybe it's just a fabrication issue with a certain production run? Who knows? BTW, when Sony build the chips they mask areas for PDAF - it would be interesting to see where those are exactly.

Whatever, if it can't be fixed via firmware, then LCC + an update to Phocus should be able to deal with it. If H/B get their act together and jump on this before it 'catches fire' on the internet, then I don't see this as a deal breaker at all. Even now, I doubt anyone would even notice it other than via pixel peeping large shifts in areas of broadly continuous tone.

It warrants investigation, exactly.

For now we just have one report: 35 XL creates horizontal lines. We'd need more data points. Time wil tell.

 

[f8orbust]

Raw Therapee has a PDAF Lines Filter since it's a known issue with a number of sensors. Maybe someone who has a CFV 100c could send them some RAW files so that they could take a look, comment, and add support if needs be.

 

[f8orbust]

It would be a big help if anyone who has a RAW file showing the problem could dump it on WeTransfer, Google Drive or whatever and post the link here (in this thread).

 

[Alkibiades]

could somebody report this problem to hasselblad?
This is a really important issue for all who want to use this back on technical cameras.

 

[rdeloe]

Raw Therapee has a PDAF Lines Filter since it's a known issue with a number of sensors. Maybe someone who has a CFV 100c could send them some RAW files so that they could take a look, comment, and add support if needs be.

On the off chance... I tried the PDAF filter (Raw Therapee v. 5.10) on someone else's sample file that I have, and it has no effect.

 

[vjbelle]

Another quick test with my 35XL. Granted I can't get anywhere near infinity but that shouldn't matter for the lines to show. I shot a wall at about 2 feet away with 16mm of rise and center filter with my Fuji 100 II. I can't go any further than that as I start to get an obstruction on the leading edge. I see no lines.

Victor B.

 

[anyone]

could somebody report this problem to hasselblad?
This is a really important issue for all who want to use this back on technical cameras.

Even better: multiple people.

For now I treat this as an annoyance. It occurs visibly only in extremes where I wouldn’t use the lens and an LCC shot is anyway required. I’ll keep an eye on the issue. Thanks to Warren’s fix it also doesn’t compromise the image quality.

The LCC by the way looks usually quite clean, a lot better than what could be expected from non-BSI CMOS backs.

 

[ruebe]

could somebody report this problem to hasselblad?
This is a really important issue for all who want to use this back on technical cameras.

Even better: multiple people.

i gladly do once my Arca is back from service. in the meantime i still wait for a reply from hasselblad about the dead pixels i get in live view (three weeks in with no com at all other than a „we received your email“ note so far …)

 

[Alkibiades]

I tested Fuji100S with a lot of lenses and did tests with 30 mm ( it was above that what the camera allows, you get the black edges of the camera mount in the picture) but never get any lines. So never seen also such lines on 150 MP phase back so it is not a problem of this sensor ( Sony problem) but a problem Made By Hasselblad.

 

[diggles]

Is it possible to fix this issue with LCC in Phocus? Would the banding appear also with SK 43mm or SK 47mm lenses?

LCC doesn't fix it in Phocus, but it is necessary to export the LCC as a tif if you want to fix it in Photoshop. Based on my experience with the 28XL, 43XL, and 60XL on the CFV II 50C and IQ4150– it wouldn't surprise me to see a similar issue with the 28XL and 43XL, but I would be very surprised to see this with the 60XL. Of course, there is no way to know for sure without trying it.

I am not sure if with roddy the 100c will work fine. color correction is color correction, dont matter what lens you use. So why you use 23HR, 32 HR with large movements for example the CC will be also visible on BSI and needs correction. Why it should work with Roddy glass and not with schneider? I am sure this problem will be with roddy also when big movements are done and the CC will be visible.

I don't think this is Color Cast since the LCC doesn't fix it. It is behaving just like sensor tiling, but lines parallel to the long side of the sensor instead of lines that split the sensor into quadrants. Like this…



The Roddie 28HR had this same sensor tiling problem for me, but I never saw it with the 23HR or 32HR when I had those lenses. It would be interesting to find out for sure by testing it though.

@Steve Hendrix if CI is willing to comp a rental of some Roddie glass I'd be happy to test it

it is really a joke that the most suitable back for technical cameras has issues that no other back has (even these that are absolutly not fully usable on technical cameras ).

To quote Rumsfeld, 'There are unknown unknowns'…

 

[f8orbust]

On the off chance... I tried the PDAF filter (Raw Therapee v. 5.10) on someone else's sample file that I have, and it has no effect.

My understanding is that the camera / sensor has to be supported by Raw Therapee's PDAF filter in order for it to work. At the time of writing the CFV 100c isn't supported, but I'm sure the developer would support it if sent the appropriate RAW file(s).

 

[f8orbust]

Here are some PDAF line artefacts on a Sony sensor: https://rawpedia.rawtherapee.com/images/8/8f/Pdaf_lines_filter_sony.png