Hasselblad 100C and 35XL

[Alkibiades]

Thank you Marco for sticking with it! I've tried your solution on 2 different 35XL images now and it is working really well, better than what I came up with because using overlay isn't introducing noise like divide sometimes does.

On the second image I tried, the lines were still barely visible after applying the first overlay lcc image. I ended up duplicating the overlay lcc image and as far as I can tell it cleaned it up completely.

@Alkibiades , try this with your 28XL and let us know how it works

View attachment 213468

Î will try it next days

 

[anyone]

Oh wow. Hasselblad replied to me finally on the issue:

"For the banding issue you are referring,Our R&D team confirmed that the "horizontal stripes"as you mentioned is caused by PDAF.And it is also known as PDAF banding.When using a wide-angle lens, especially when using a technical camera, it is easy to appear in the edge position, because the CRA of the edge light is large, and the dark Angle problem is more significant. The algorithm will expand the gain of the corners to compensate for the attenuation of brightness, and because the sensitivity of PD pixels is only half of that of ordinary pixels, the difference of PD pixels is further amplified.hence It forms a regular horizontal black line.

Here is the method we provide to mitigate this kind of issue(see attached)."

... and the method they attached, is a word document, I provide a screenshot, there are some elements that may seem familiar to some of you:



Right. It's a copy of the steps AND the images Warren (@diggles) provided. Did they reach out to you, Warren, and ask for permission?

Needless to say that this is not the answer I hoped for. I replied and asked for a native fix, either in the firmware or in Phocus.

 

[Paul Spinnler]

OMG. I'd be furious if they'd take my makeshift solution from a forum post.

Firstly its not ok w/o permission and secondly its not a REAL solution.

Looks like there won't be a fix for this back.

 

[diggles]

Oh wow. Hasselblad replied to me finally on the issue:

"For the banding issue you are referring,Our R&D team confirmed that the "horizontal stripes"as you mentioned is caused by PDAF.And it is also known as PDAF banding.When using a wide-angle lens, especially when using a technical camera, it is easy to appear in the edge position, because the CRA of the edge light is large, and the dark Angle problem is more significant. The algorithm will expand the gain of the corners to compensate for the attenuation of brightness, and because the sensitivity of PD pixels is only half of that of ordinary pixels, the difference of PD pixels is further amplified.hence It forms a regular horizontal black line.

Here is the method we provide to mitigate this kind of issue(see attached)."

... and the method they attached, is a word document, I provide a screenshot, there are some elements that may seem familiar to some of you:

View attachment 213714

Right. It's a copy of the steps AND the images Warren (@diggles) provided. Did they reach out to you, Warren, and ask for permission?

Needless to say that this is not the answer I hoped for. I replied and asked for a native fix, either in the firmware or in Phocus.

Shortly after this thread started, I reached out to Hasselblad support to bring the issue to their attention and shared this solution with them. They have permission, and I’m happy they find it useful enough to share with others.

Perhaps the fix by @mristuccia should be shared with them as well, but I'll leave that up to Marco.

Ideally, they will develop a method within Phocus. If RawTherapee can do it, then they should be able to do it too. I wonder if they know about it…

 

[akaru]

Hasselblad Support Document #13492: “One hell of a mess”

 

[Vilpo]

Hasselblad Support Document #13492: “One hell of a mess”

May I have that also in Swedish please?

 

[Doppler9000]

May I have that also in Swedish please?

Helvete-Smörgåsbord?

 

[Bill_Evans]

I just bought into this ”mess” . I confess that I’m quite a bit disappointed that I wasn’t able to bargain basement shop this “crapy” digital back in the buy / sell forum . But luckily our fabulous forum sponsor had a couple in stock. I’m looking forward to easing it into my various workflows and am grateful for all the useful info this site provides.

 

[anyone]

I just bought into this ”mess” . I confess that I’m quite a bit disappointed that I wasn’t able to bargain basement shop this “crapy” digital back in the buy / sell forum . But luckily our fabulous forum sponsor had a couple in stock. I’m looking forward to easing it into my various workflows and am grateful for all the useful info this site provides.

Congratulations! I think this is an awesome back in many situations. Just this thread is about the fix of a specific problem. If Hasselblad manages to get this fixed, it will be 100% awesome. Now it's "just" 85%.

 

[mristuccia]

Shortly after this thread started, I reached out to Hasselblad support to bring the issue to their attention and shared this solution with them. They have permission, and I’m happy they find it useful enough to share with others.

Perhaps the fix by @mristuccia should be shared with them as well, but I'll leave that up to Marco.

Ideally, they will develop a method within Phocus. If RawTherapee can do it, then they should be able to do it too. I wonder if they know about it…

Any of you feel free to share my fix with HB, if you find this helpful.

I for one am not going to waste my time with HB support again in my life, unless it is really, really necessary. Last time I reported an issue on Phocus for macOS Intel machines, I did not see any fix in any of the next three versions of the software, despite thanks and the promise of a fix (keystone correction was not applied to exported TIFFs/JPEGs). Then I gave up, moved to Apple Silicon and lost track. But I bet the issue is still there.

 

[anyone]

Another answer of the support, to my reply:
"As the technical support,your request is understood by us.We will report your request to our R&D team and currently we kindly recommend you to stay tuned with our official website for the latest news on this."

 

[lookbook]

Another answer of the support, to my reply:
"As the technical support,your request is understood by us.We will report your request to our R&D team and currently we kindly recommend you to stay tuned with our official website for the latest news on this."

nice answer -
be glad that they don't take the camera away from you : ) ..

 

[rdeloe]

The thread is long and I didn't want to read through it all to confirm... but a colleague on another forum who works with image processing wanted to have a look at some files. He used RawTherapee to fully remove the PFAD banding using the LCC as a flat-field file. All it required was setting the blur radius to 0. This may already be a documented solution in this thread! But just in case it isn't, I thought it worth updating.

Here are some before and after files he posted:

Before:

http://imgur.com/Qcja0ib

After:

http://imgur.com/bkbSDVu

Has Hasselblad done any work on this, or are people still expected to struggle along outside of Phocus with whatever solution they can find using other tools?

 

[mristuccia]

The thread is long and I didn't want to read through it all to confirm... but a colleague on another forum who works with image processing wanted to have a look at some files. He used RawTherapee to fully remove the PFAD banding using the LCC as a flat-field file. All it required was setting the blur radius to 0. This may already be a documented solution in this thread! But just in case it isn't, I thought it worth updating.

Here are some before and after files he posted:

Before:

http://imgur.com/Qcja0ib

After:

http://imgur.com/bkbSDVu

Has Hasselblad done any work on this, or are people still expected to struggle along outside of Phocus with whatever solution they can find using other tools?

As far as I understand the only drawback of this solution is that we can't leverage HNCS as Phocus is not in the PP chain anymore. Right?

 

[anyone]

Has Hasselblad done any work on this, or are people still expected to struggle along outside of Phocus with whatever solution they can find using other tools?

Hasselblad has not solved this issue until now.

Edit: It was a good reminder to get in contact with them again.

 

[rdeloe]

As far as I understand the only drawback of this solution is that we can't leverage HNCS as Phocus is not in the PP chain anymore. Right?

Exactly. The person who did this test happens to not like using manufacturer profiles so it's all good for him! But if you bought Hasselblad for the HNCS, then this is not the solution you would want.

 

[f8orbust]

The thread is long and I didn't want to read through it all to confirm...

Indeed - I posted that solution on Page 4 of this thread.

Can't believe that over a year later and H/B haven't baked in a similar solution to Phocus. Pretty poor.

I don't know if the developers of RT have added the H/B 100MP chip to their PDAF filter - since their last update was August. 2024, probably not - but that would be a 'one click' solution (in RT at least).

 

[SrMphoto]

Indeed - I posted that solution on Page 4 of this thread.

Can't believe that over a year later and H/B haven't baked in a similar solution to Phocus. Pretty poor.

I don't know if the developers of RT have added the H/B 100MP chip to their PDAF filter - since their last update was August. 2024, probably not - but that would be a 'one click' solution (in RT at least).

The solution is to fix it in the camera, not in Phocus.

 

[mristuccia]

I may be wrong, but I don't think we will ever see a solution any time soon. As far as I understand, the displacement of the bandings is not evenly spread and it seems to depend on the lens model and the amount of shift. The firmware would not have the related metadata to implement a correction transparently...

 

[SrMphoto]

I may be wrong, but I don't think we will ever see a solution any time soon. As far as I understand, the displacement of the bandings is not evenly spread and it seems to depend on the lens model and the amount of shift. The firmware would not have the related metadata to implement a correction transparently...

I assume GFX shows no banding (they fixed it in GFX). If that is true, then Hasselblad should be able to fix it too.