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
mristuccia
Well-known member
Today I've written this email to Hasselblad.
As suggested by some of you, I'll also try to contact them via LinkedIn.
I won't hold my breath however...
As suggested by some of you, I'll also try to contact them via LinkedIn.
I won't hold my breath however...
Dears,
My name is Marco Ristuccia, I'm a computer engineer based in Berlin.
I'm also a photographer, and I happen to own a beautiful 907x + CFV-100c.
This is a message that should be firstly addressed to the Hasselblad/Phocus Technical Team, and in
second instance to the Hasselblad Commercial Team.
Since I frequent various photography forums, including the GetDPI and DPReiview medium format ones,
I know that you're well aware of a serious technical issue which makes the use of the CFV-100c as digital
back on technical cameras a challenge, and which gives the Phase One backs a huge advantage in respect
to your product.
I'm talking about the PDAF banding issue, which affects images taken with the CFV-100c on a technical camera
when using wide angle technical lenses, mostly Schneider-Kreuznach rectilinear (symmetrical) lenses, still widely
used, but also some of the common Rodenstock ones.
This is proven by the many who tried personally the various combinations.
Your technical team has already answered to the ones who complained about this, but the answer is vague and
does not seem to give concrete hopes on the possibility to have a fix, either in firmware or in Phocus.
A member of the GetDPI community already shared with you a workaround, which you recommend as a fix
when asked about this issue.
Now, I myself have spent a big time working on potential solutions.
At first, I've shared (still on GetDPI) a simpler solution, which makes use of an inverted and normalized flat-field
image (LCC / Scene Calibration) put in overlay over the real image in Photoshop. This works pretty much well, and
is less noise-prone than the one you're currently recommending, but it is still tricky to fine-tune as the first one is.
Then, I worked on a real tool, which I've just finished to develop, which simply applies the flat-field (scene-calibration)
without blurring the LCC first. If applied properly, like it is documented by the ample literature on the subject, it
works great and it has the advantage of not only removing PDAF banding, but also dust spots, sensor tiling, etc...
I know that not pre-blurring the scene-calibration image may induce some slight noise in the final image, so you could
just provide it as an option that can be turned on or off depending on the situation, and/or apply a slight noise reduction
after the fact.
Now, I'm going to offer my tool to the community, and there are already other softwares that do it right as well,
like RawTherapee to name one. But I sincerely hope that you could implement such enhancement to the Phocus'
Scene Calibration algorithm. The technical cameras' community, of which I myself am a part, could then benefit in
full from your fantastic digital back and, maybe, you'll gain some market share among the technical camera users
like the high-end landscape photographers and architects over Phase One due to the better cost/quality ratio.
So, I've shared some concrete solutions with you, and I've a tool which I'm going to sell until you'll come up with a
real fix. My tool is also able to perform frame-averaging and frame-median on 3FR RAW images, keeping the final
combined image still in the original RAW format. This is another feature you should consider to implement, and
which is already offered "in-camera" by Phase One for example.
Now, this should not be my business, I'm putting a small price tag to my tool only to recover some of the time I've
spent on such topic. But it really should be your business, and I sincerely hope that sooner or later my tool will not
be necessary anymore. You have way more resources and better technical know-how than what I alone could ever have.
I hope this email of mine will raise your awareness of the issue further.
Right now your tri-fecta is more of a 2.5-fecta.
Feel free to reach me out in case you need further details on how I've fixed the issue.
--
Marco Ristuccia
Photographer & Computer Engineer
Last edited:
Your tool is also very useful for Adobe users as it allows applying AI NR after LCC and banding fix have been applied.
harshdreams
Member
Hi Marco, Firstly, kudos to you for coming out with this solution/interface for cleaning up files created using Hassy backs . Is there any way your software can be used to clean up sensor tiling issues I have hit upon with certain extreme shifts using the Schneider Digitar 28Xl and the IQ4150? Thank you
In his long letter to Hasselbad Marco says that his tool can fix also tiling issues. That's promising. But Marco's tool is made for Hasselblad only and only for in-camera correction.
For Phase One user I suggest RawTherapee. I set the LCC blur to zero to fix dust spots. I guess that is the same with tiling. If there's no blur when assigning the flat-field file it's very important to keep all settings, tilt, shift, focus, aperture exactly the same in the LLC and in the picture. Only change the exposure time to get well exposed LCC.
For help look: https://rawpedia.rawtherapee.com/Flat-Field
mristuccia
Well-known member
Yes, I did not make any miracle here, I'm only applying the LCC without blur, like RawTherapee does, to name one. That should fix most of the structural artifacts that can be captured by the LCC, basically counterbalancing them. However, I've added a small set of fine-tuning parameters that are not usually offered by RAW developers.
If one blurs the LCC before applying it, all small structural anomalies will be smoothed out, leading to a slightly less noisy final image, but loosing the ability to correct small spots since they are not in the LCC anymore.
Thus, having the ability to regulate the blur radius can help in finding the sweet spot.
The only reason for not using RawTherapee for this purpose is that I like to keep the corrected file still as a native Hasselblad RAW file, in order not to loose the ability to develop it in Phocus and take advantage of the Hasselblad Natural Color Solution. Moreover, it all started with the goal of performing frame-averaging, not LCC correction.
As suggested by TimoK, with IQ4 files one can use RawTherapee.
As a side note, all this work makes me think that we could probably benefit of some sort of "LCC Studio", a dedicated software for fine tuning the application of an LCC to an image, with a live preview and a set of fine tuning sliders for fixing residual color dominances, changing the correction factors, LCC blurring, etc... Right now RAW developers don't offer that much on this area, just very few and basic options. But I sincerely don't know if nowadays it would be worth it. Just thinking out loud...
Last edited:
mristuccia
Well-known member
Answer from Hasselblad:
That is a good reply, they really read and understood your email.
mristuccia
Well-known member
Now that I think about what I'm doing here, I might need to explore the potential legal implications in dealing with proprietary RAW files.
So I'll be very transparent with them with two goals:
So I'll be very transparent with them with two goals:
- for my frame-averaging tool to even exist without risking legal infringements
- for them to clearly understand and maybe implement at least the PDAF banding fix directly in their Phocus' scene-calibration implementation.
If not a miracle you made a good job with your tool. Thanks for it!Yes, I did not make any miracle here, I'm only applying the LCC without blur, like RawTherapee does, to name one. That should fix most of the structural artifacts that can be captured by the LCC, basically counterbalancing them. However, I've added a small set of fine-tuning parameters that are not usually offered by RAW developers.
If one blurs the LCC before applying it, all small structural anomalies will be smoothed out, leading to a slightly less noisy final image, but loosing the ability to correct small spots since they are not in the LCC anymore.
Thus, having the ability to regulate the blur radius can help in finding the sweet spot.
The only reason for not using RawTherapee for this purpose is that I like to keep the corrected file still as a native Hasselblad RAW file, in order not to loose the ability to develop it in Phocus and take advantage of the Hasselblad Natural Color Solution. Moreover, it all started with the goal of performing frame-averaging, not LCC correction.
As suggested by TimoK, with IQ4 files one can use RawTherapee.
As a side note, all this work makes me think that we could probably benefit of some sort of "LCC Studio", a dedicated software for fine tuning the application of an LCC to an image, with a live preview and a set of fine tuning sliders for fixing residual color dominances, changing the correction factors, LCC blurring, etc... Right now RAW developers don't offer that much on this area, just very few and basic options. But I sincerely don't know if nowadays it would be worth it. Just thinking out loud...
Keeping HNCS alive is a big advance. Also doing the fix in the camera is a very nice thing. It saves much work in post.
So far I know any of the main stream raw developers can't do that kind of fix. So there's no way but RT to correct the banding issue in post. ( I guess very soon Phocus will be able to do it.
) But CFV-100c is not ( yet? ) supported by RawTherapee, so it's not a real solution.I saw you added some horizontal ( in landscape format) blur in your tool. That helps with noise issues and still fix the banding. I think you can correct dust spots wiht other tools.
Btw. Can you make frame averaging and LCC correction to the same picture in-camera?
And as I said before try RawTherapee with Phase One files. IQ4 150 is supported by RT.
mristuccia
Well-known member
What do you mean by "in the camera"? I cannot do this in-camera, only Hasselblad could do. I'm doing it on the 3FR files produced by the camera.
In fact I tried once and it did not work...So far I know any of the main stream raw developers can't do that kind of fix. So there's no way but RT to correct the banding issue in post. ( I guess very soon Phocus will be able to do it.
Yes, I've added that feature.
The banding however is not fixed anymore if you exceed 1-2 pixels radius when blurring along the long side. But it varies on a use-case basis.
You will of course lose the ability to fix PDAF banding if you blur along the short side.
Blurring can be used with more generosity when using lenses that do not produce banding.
N.B.
I've chosen to use the terms "long side" and "short side" instead of the ambiguous "horizontal" and "vertical" because photos can be taken in landscape or in portrait orientation.
When frame-averaging (or median), you have the ability to choose a single LCC which is applied directly to the final composed RAW file before writing it to disk. So: yes.
Indeed. Moreover, I don't think P1 has the concept of HNCS, so IQ4 files can be processed with RT. But I think that C1 has already the ability to correct dust spots when applying the LCC. This should also fix sensor tiling and other potential structural anomalies. So why using RT with IQ4 files?
Unfortunately I don't have enough resources to venture into the reverse-engineering the IQ4 RAW files. It took already a lot of time doing this with HB 3FR files.
Last edited:
MGrayson
Subscriber and Workshop Member
Marco,
I beg to differ. MANY people talk about how easy it would be for Hasselblad to do X or Y. You have actually done it.
Matt
Last edited:
I misunderstood how your tool is working, sorry.What do you mean by "in the camera"? I cannot do this in-camera, only Hasselblad could do. I'm doing it on the 3FR files produced by the camera.
I also tried it using a file from somebody in getdpi forum. I could correct banding, but did not get colors and tones right. Then I found out that 100c was not supported.
I'm not a native English speaker. I don't always get the right terms, my bad.
That's fine!
I did not know that. So, why not using C1?
I agree with Matt in this case.Unfortunately I don't have enough resources to venture into the reverse-engineering the IQ4 RAW files. It took already a lot of time doing this with HB 3FR files.
mristuccia
Well-known member
Thank you Matt. Yesterday I did send more details to HB. They seem very interested. Today they answered thanking me for the detailed explanation and apologizing for the delay with the justification that their technical team is analyzing what I've sent to them.
Cross fingers...
Last edited:
mristuccia
Well-known member
No worries, and I'm sorry for the misunderstanding. Doing it in-camera would be a dream but I don't have access to their firmware, at least yet...
You can see the tool as a 3FR-to-3FR pre-processor.
The development chain would be something like this
Code:
Camera --> 3FR --> HBComposer (pre-processor) --> 3FR ---> Phocus (import) --> FFF --> Phocus (raw development) --> TIFF --> Photoshop (post-processing) --> ...
\_-> LcR (import & raw development) -----------------------------_/Your understanding here was perfectly correct. I just wanted to explain better the reason why I'm not using the terms horizontal or vertical in my tool.
And, by the way, we are in the same ship as far as English speaking. I'm not native as well.
At least I made you happy here.
I'm not and will not be a fortunate owner of an IQ4, so I can't test myself what C1 is capable of against RT.
Thank you!
Last edited:
Doppler9000
Well-known member
Marco - thank you so much for taking this on.
mristuccia
Well-known member
I appreciate, but no need to thank me. Having this PDAF banding issue fixed is in my own interest as well.
cirofrankschiappa
New member
Grazie
mristuccia
Well-known member
So, while we are waiting for a feedback from Hasselblad headquarters, and since the situation is not 100% clear right now, I've decided to put the new version of the tool online in conjunction with a temporary trial license, which expires on 1 June 2025.
This is the version which contains the advanced Flat-Field (LCC) correction which helps fix the PDAF banding issue, sensor tiling and dust spots.
The tool is still on its early beta version, there may be bugs. If you'd like to test it and give me feedbacks, you can get it at the following URL:
https://photography.marcoristuccia.com/hbcomposer-an-hasselblad-raw-file-composer/
You'll need to download the license file, copy and paste its content into the tool's registration window once you open it. An Internet connection is needed while registering.
The user guide is available online at the following URL (excuse my bad English, there may be errors):
https://photography.marcoristuccia.com/hbcomposer-user-guide/
I've also kept the original CLI and GUI versions, which are fully functional and totally free.
On this point, I'd like to thank everyone who is supporting me with donations. That's really appreciated! Depending on the outcome of the conversation with Hasselblad, I may put a small price tag to the full version. In that case, I will refund all donations coming from the ones who will buy it then.
This is the version which contains the advanced Flat-Field (LCC) correction which helps fix the PDAF banding issue, sensor tiling and dust spots.
The tool is still on its early beta version, there may be bugs. If you'd like to test it and give me feedbacks, you can get it at the following URL:
https://photography.marcoristuccia.com/hbcomposer-an-hasselblad-raw-file-composer/
You'll need to download the license file, copy and paste its content into the tool's registration window once you open it. An Internet connection is needed while registering.
The user guide is available online at the following URL (excuse my bad English, there may be errors):
https://photography.marcoristuccia.com/hbcomposer-user-guide/
I've also kept the original CLI and GUI versions, which are fully functional and totally free.
On this point, I'd like to thank everyone who is supporting me with donations. That's really appreciated! Depending on the outcome of the conversation with Hasselblad, I may put a small price tag to the full version. In that case, I will refund all donations coming from the ones who will buy it then.
Last edited:
Marco,
Thanks for all your software innovations around the CFV100C from Hasselblad. For me the PDAF banding solution you made is great news I just tried it and the first results are better than what I have achieved with other solutions so far. There is a bit more noise and the colors are not quite where I want them but I think I need to experiment a bit more for that. Very cool so far!
Thanks for all your software innovations around the CFV100C from Hasselblad. For me the PDAF banding solution you made is great news
I just tried it and the first results are better than what I have achieved with other solutions so far. There is a bit more noise and the colors are not quite where I want them but I think I need to experiment a bit more for that. Very cool so far!