Hasselblad 100C and 35XL

[diggles]

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.

It is fixable in Photoshop so it is just another annoyance to add to the list of wide angle lens annoyances … center filters, color cast, sensor tiling, lcc, etc…

 

[buildbot]

I computed the change in brightness per row in the 2048x2048 llc crop @diggles posted, averaging all the columns together:

Clearly there is banding there! (I was having a hard time seeing it). About every 173 pixels there is a darker band of pixels. Anyone now the spacing between pdaf pixels on sony sensors?

 

[Godfrey]

Just like a lot of other people on the forum, I am very interested in using wide angle Schneider lenses on the Hasselblad 100C. While testing it out over the last couple of weeks I found an issue that is important to share with everyone here.
{... snip ...}

I must admit, I have a great deal of difficulty even seeing the lines you speak of. But then, it must be said, I have no technical camera nor any interest in using SK lenses. I use my CFVII 50c with Hasselblad X and V (and other , adapted, mostly Leica M- and R-mount) lenses on the 907x and V-system lenses on 500CM bodies.

Someday I'll try the CFV 100c, I'm sure...

G

 

[rdeloe]

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).

<Slaps forehead> I used a DNG exported from LR.

However, Raw Therapee will read the fff file. I tried that. I don't understand what I'm seeing. It seems to be a partial demoasicing, if at all. If the lines are present, I'm unable to see them in what it's showing me, but don't read too much into that.

 

[rdeloe]

Another data point... With a rise of 15mm and 20mm on a GFX 100S (out of focus so I can clear the mount), I see no lines. I don't see why the image would need to be in focus for the lines to appear, but I'm not certain.

 

[daz7]

I must admit, I have a great deal of difficulty even seeing the lines you speak of.

It is quite easily visible if you try to scroll the image up and down - the darker lines become obvious even on a low resolution screen.

 

[Paul Spinnler]

... its Hassy specific and related to the PDAF ... the IQ4 has no such lines

 

[Godfrey]

It is quite easily visible if you try to scroll the image up and down - the darker lines become obvious even on a low resolution screen.

Nope. Can't see them at all, even on my 5K Mac Studio Display. I've scrolled the images, all of them, up and down and side to side, after downloading and looking at them in four different resolutions. Just about impossible to see.

I really debate that it's worth fussing over any further. If I can't see them, and if they don't show up in any print I can make, who cares?

G

- I clipped a 250x250 pixel patch out of one of the sample files, converted it to B&W, and rendered the tone curve down to a very very narrow delimited band based on the graph buildbot presented. Then I could finally see the bands at 800% view. Sheesh. This is what keeps you up at night... ? Can you see them in any normal print of the image?

 

[Alan]

...who cares?

Professionals who produce images for discerning clients for one.

If you don't, great - carry on.

 

[tcdeveau]

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: Jul 18, 2024

www.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.

 

[Alan]

Anyone have wide (23-35mm) Rodie HRs to test on the CFV100C?

I'll second the recommendation for everyone to report this to Hasselblad - also the need to correct distortion on shifted Rodie lenses.

 

[Geoff]

Do I have this right: we are talking about the 35mmxl on the 100C back at exteme shifts.
What happens if we give up the extreme shifts, and stay more reasonably at 15 or even 10mm? Still lines?
If not, then it seems its an extreme case, not the norm?

 

[diggles]

Do I have this right: we are talking about the 35mmxl on the 100C back at exteme shifts.
What happens if we give up the extreme shifts, and stay more reasonably at 15 or even 10mm? Still lines?
If not, then it seems its an extreme case, not the norm?

The example I posted to open the thread is about 15mm of camera fall, which is extreme for this lens in my opinion. The lines are more faint, but still noticeable at smaller shifts of 5mm camera fall + 5mm shift.

 

[Paul Spinnler]

It should be fixable in post with algorithms - if one can LCC the CC away and remove the banding via PS there's no resson why advanced processing taking into account the PDAF array shouldn't be able to get rid of the effects at the level of raw processing already.

The C1 LCC algo also was never further developed to deal with tiling although it most likely could be done with some R&D.

Problem is that the development of these features costs real money and if the demand for it is "niche" we won't see anything coming.

So the strategy to make Hasselblad aware of this seems reasonable to me. They might update LCC correction within Phocus.

Back in the day P1 also updated the LCC correction algo from a gen I version to a gen II which is used today - it brought a lot of improvements in the handling of the CC side.

 

[daz7]

Nope. Can't see them at all, even on my 5K Mac Studio Display.

well, I can see it quite clearly on my mobile phone and on my LCD laptop display. Not sure why you cannot see it. Horizontal banding is evident.

 

[hcubell]

Just like a lot of other people on the forum, I am very interested in using wide angle Schneider lenses on the Hasselblad 100C. While testing it out over the last couple of weeks I found an issue that is important to share with everyone here.

Here is a JPG made from the RAW image file.
View attachment 210951

Here is a JPG made from the RAW LCC file.
View attachment 210952

Here is a 2400x2400 crop of the top left corner of the LCC file.
View attachment 210953

As you can see, there are lines that run parallel to the sensor. When I make an image in portrait orientation the lines are vertical. The bad news is that the lines are there, the good news is that it can be fixed using the same process that is used to fix sensor tiling.

How to Fix Sensor Tiling With Photoshop

My first choice for photo editing is ACR/Photoshop. While looking for a way to apply LCC corrections without using C1/Phocus/Lightroom I ran across this article: https://pduncan.com/2015/11/20/vignetting-4/ The steps outlined in the article do a good job of fixing vignetting, but didn't really...

www.getdpi.com

The above images are shifted about 15mm which is extreme for this lens. Even with 5mm camera fall + 5mm shift the lines are present, but not as noticeable. Extreme shifts also introduce noticeable noise in the darker areas of the image. With smaller shifts the noise is not very noticeable.

[Edited: download links removed]

Send me a PM if you are interested in downloading the RAW files and exported Tiffs for a closer look. In order to do the sensor tiling fix you'll need to export the corrected LCC as a tiff file along with the corrected image. Phocus can do this, but I'm not sure if Lightroom can.

Hope this helps!

NOTE: This is just one person's experience with this combo. If there are any others that have the 35XL and 100C, please let everyone know what your experience is.

How big would a full size print be based upon your 2400x2400 ppi crop?

 

[diggles]

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

Those PDAF line artifacts sure look similar.

 

[diggles]

How big would a full size print be based upon your 2400x2400 ppi crop?

It is a 100% crop of the top left corner. The full image size is 11656x8742 pixels. A 300dpi print of the complete digital file would be about 38" wide.

 

[Paul Spinnler]

I think we should NOT write off SK for the CFV100c - it may very well be dealt with by Phocus in post in the future ... as Buildbot shows, the pattern is predictable and identifiable, especially with an LCC file on top.

 

[hcubell]

It is a 100% crop of the top left corner. The full image size is 11656x8742 pixels. A 300dpi print of the complete digital file would be about 38" wide.

Thanks.