Fuji Sub-um technology will bring pixel-shift and 400mpx on GFX100

[mristuccia]

I was very excited by "Enhance Detail" when it first came out.

But my testing showed it basically brought LR on par-ish with Capture One's native raw processing, but with the added cost of workflow speed, workflow complexity, larger storage requirements, and the requirement to use LightRoom.

Really much better (in my highly biased opinion) to just use Capture One.

Just to close this analysis, even if it is a little bit out of topic (apologise), I tried Capture One as well (don't ask me how I managed to load the Hasselblad RAW file into C1... )
You were right in saying that C1 manages the demosaicing great, no color artefacts in the fence and in the water (maximise your browser window to better see details).

The image URL: http://www.marcoristuccia.com/shared/ComboTest.jpg

Kudos to C1!

Best,
marco

 

[MGrayson]

Marco,

Of course, I have no idea if this is what you did, but comparing different RAW converters at their default settings is useless.

C1's defaults are usually better tuned, and its images look better out of the box. LR doesn't bother, and requires you to adjust their default settings to get the best (or even decent) results. I have not done this experiment, but what happens if you increase the Moiré suppression in LR? Or color noise suppression? Do the color artifacts fade? Bayer de-mosaicing is a fairly mature technology.

There are cameras for which C1 is distinctly better. For the Fuji x-trans sensor, C1 does a great job that LR just can't match. The maze artifacts in C1 are as well suppressed as the gold standard Iridient. With LR, they are annoying and I, at least, have found no way to avoid them.

Best,

Matt

 

[mristuccia]

Marco,

Of course, I have no idea if this is what you did, but comparing different RAW converters at their default settings is useless.

C1's defaults are usually better tuned, and its images look better out of the box. LR doesn't bother, and requires you to adjust their default settings to get the best (or even decent) results. I have not done this experiment, but what happens if you increase the Moiré suppression in LR? Or color noise suppression? Do the color artifacts fade? Bayer de-mosaicing is a fairly mature technology.

There are cameras for which C1 is distinctly better. For the Fuji x-trans sensor, C1 does a great job that LR just can't match. The maze artifacts in C1 are as well suppressed as the gold standard Iridient. With LR, they are annoying and I, at least, have found no way to avoid them.

Best,

Matt

Hi Matt,

I tried to adjust the color noise reduction and the moiré correction in LR. By means of a localised moiré correction I'm able to remove the color artefacts from the fence and from the water. On the other side, even with the color noise reduction set to maximum the artefacts remain there, especially on the water.
Capture One (v. 12) remains pretty clean even with the chroma noise reduction set to 0 (default = 50) and the moiré set to 0 (which is the default setting).

I have the Fujifilm X-T2 and I agree with you that for X-Trans C1 and Iridient beat LR hands down. That's the (only) reason why I've bought C1 12 Pro for Fujifilm.

Regards,
m/

 

[gerald.d]

Actually, for Japanese manufacturers, they do not by an industry standard use the number of pixels to define resolution. I was actually working as a technical writer for a camera company when the standards were set for describing the technical specifications of a digital camera.

What year was this?

 

[MGrayson]

Hi Matt,

I tried to adjust the color noise reduction and the moiré correction in LR. By means of a localised moiré correction I'm able to remove the color artefacts from the fence and from the water. On the other side, even with the color noise reduction set to maximum the artefacts remain there, especially on the water.
Capture One (v. 12) remains pretty clean even with the chroma noise reduction set to 0 (default = 50) and the moiré set to 0 (which is the default setting).

I have the Fujifilm X-T2 and I agree with you that for X-Trans C1 and Iridient beat LR hands down. That's the (only) reason why I've bought C1 12 Pro for Fujifilm.

Regards,
m/

Fair enough. No one tool is perfect, alas.

M

 

[dougpeterson]

C1's defaults are usually better tuned, and its images look better out of the box. LR doesn't bother, and requires you to adjust their default settings to get the best (or even decent) results. I have not done this experiment, but what happens if you increase the Moiré suppression in LR? Or color noise suppression? Do the color artifacts fade? Bayer de-mosaicing is a fairly mature technology.

One would think. But then you try Adobe Lightroom and Capture One on the same well-shot raw file, tune both to their best results, and it's still not a very subtle difference in my opinion.

I think we all underestimated how much the math for raw demosaicing could be improved. If you asked me when I started in 2003, I would have told you there was little, if any, improvement yet to be made at that point. I was, plainly put, spectacularly wrong.

 

[dougpeterson]

Just to close this analysis, even if it is a little bit out of topic (apologise), I tried Capture One as well (don't ask me how I managed to load the Hasselblad RAW file into C1... )
You were right in saying that C1 manages the demosaicing great, no color artefacts in the fence and in the water (maximise your browser window to better see details).

And moreover I think the overall detail rendering in C1 is still more pictorially pleasant. Look at the background areas where LR is artificially sharpening/"enhancing" subject matter that isn't actually detail, or at the harsh (and not consistently present) grain in the high detail fence areas in LR.

And since you likely had to make the comparison using a hacked raw file (since your camera isn't supported in C1) imagine how much nicer it is when you're working with a camera that is natively supported and for which lens profiles, color profiles, and detail/noise algorithms are all specifically tuned.

The math in C1 is just better; full stop.

That said, if, for whatever reason, you're a LR user, the "enhance details" might be a useful way to get closer to the C1 result, albeit at the cost of speed, file size, and simplicity.

P.S. To my eye I suggest lower the sharpening threshold on the C1 file, and reducing the amount of sharpening, but turning the detail slider up a bit. That should help with the more filmic/naturalistic rendering of the smoother areas of the subject that are currently ever so slightly smoothed out. Of course, it's subjective, so you may or may not agree – and you'd be right either way!

 

[MGrayson]

One would think. But then you try Adobe Lightroom and Capture One on the same well-shot raw file, tune both to their best results, and it's still not a very subtle difference in my opinion.

I agree with this statement, but not, I think, in the way you meant it. I often process the same RAW in both converters, and choose one or the other, depending on what I can achieve. Color leans towards C1, B&W leans towards LR. Fuji x-trans are 100% C1. Leica S is 100% LR :ROTFL:.

We can debate the proper usage of "better" and "worse" another time.

Best,

Matt

 

[Shashin]

What year was this?

Every year since the early 2000s up to the present. The term is "Effective Number of Pixels."

 

[PeterA]

I agree with this statement, but not, I think, in the way you meant it. I often process the same RAW in both converters, and choose one or the other, depending on what I can achieve. Color leans towards C1, B&W leans towards LR. Fuji x-trans are 100% C1. Leica S is 100% LR :ROTFL:.

We can debate the proper usage of "better" and "worse" another time.

Best,

Matt

Interested in your experience /preference for LR over C1 for B&W conversions Matt.

 

[MGrayson]

Interested in your experience /preference for LR over C1 for B&W conversions Matt.

Peter,

If I want a smooth, realistic rendering, C1 does fine. If I want more drama, LR has a wider range of effects - clarity, texture, dehaze. C1 has finer color selection, but if I want to turn a blue sky black, or green leaves white, I can do that more easily in LR. (Edit: never mind, I just figured out how to do that in C1. I was unaware of the interaction between the Color adjustment blocks and the B&W block.)

In short, I can get to the overly dramatic look I sometimes want with a few sliders in LR. The clarity adjustment in C1 covers a much smaller range than in LR. And I haven't found anything in C1 with the same violence as dehaze in LR.

But I work each promising RAW with both converters, and about 1/3 of the time, I prefer the C1 rendering. I have never been successful adjusting one converter's output to match the other's. It is, of course, my interaction with them, and someone else might find them interchangeable, or one a clear superset of the other. I just find them different.

Now that I'm using Fuji x-trans more, I'll spend more time in C1, and probably learn to get all the effects I want.

Best,

Matt

 

[Paul2660]

Sure wish C1 would develop a dehaze. I love the feature in LR.

Paul C

 

[PeterA]

Hi Matt,

Thanks for your response, as a long term LR user I've only recently started with C1 so I am interested to read what other people's experiences / feelings are - thanks for your response.
I guess the main reason I switched to C1 was its support for the GF range of cameras - I find that I too prefer the colour management of C1.
TBH - I'm not particularly happy with B&W conversion from either - I think the drama can be made quite easily - but both lack a degree of managing subtle gradations - I've noticed Alan using software called darktable and am looking into that - since I have been very impressed with some of the B&W renditions in Alan's landscapes - steep learning curve and a tad clunky in use compared to LR/C1 though.

All the Best
Pete

 

[PeterA]

Damn. I just tried Darktable. The BW conversion is simple, but gives quite appealing and - "palpable" is a word that comes to mind - results. Is there a thread somewhere dedicated to BW converter comparisons? I think we need one, if not. Darktable's interface is ... interesting.

M

PS. DT doesn't support the GFX100, so I sent them a RAW picture of a ColorChecker. I hope they add it soon!

Well Allan Stewart mentioned darktable to Bart and I by accident read this mention and googled it. There are many videos on youtube by a bloke called Williams who goes through dark table - but after watching any video I am left not much wiser - I may have to see if there is a manual I can read for myself. I also spent a couple of hours watching a video made by the Chief designer who is French - very clever man and interesting watching but again - one has to know a certain amount about the actual user interface and workflow - before one can even begin to learn something properly. I dislike mucking around so I need to see if there is some systematic way to familiarise myself with the B&W conversion routines in Tone Module.

Well done sending the GFX file to them - yes may well deserve a seperate thread .

Pete

 

[RLB]

In my opinion, "more smoke and mirror" talk from a camera manufacturer. Pixel shift technology has been around a long time (20+ years) and I'm yet to examples that impress me without inherent limitations of the process. That said it looks great to state in marketing materials 1,000,000mp captures. Let us also consider the physical limitation of diffraction, optics, file size, camera functionality, ISO noise limitations, etc., etc.

The cry is always wanting more and more quality for less $. Rarely have I see that actually pan out when the finer points (limitations) are read and fully understood and its reduced to marketing hype. This of course is all just tech talk, and if one can't make an artistically interesting image with the device, the entire argument is mute (unless its just for scientific use).

So Fuji...show me your cards. Yawn...


Robert B

 

[PeterA]

There is a user manual and a book. Both can be downloaded from the darktable website. They suggest the channel mixer for BW conversion. It's a simple design - just R, G, and B response, but it works very well. There are a number of presets in the menu at the very upper right of the channel mixer. I just started learning this today, so I may discover I'm completely wrong about how best to use it.

M

Yep, completely wrong. I watched an hour of BW conversion (applied to portraits, but I saw a lot of tools used). There are a LOT of ways to alter a BW conversion. What makes the videos difficult is that the tools are labeled in small print, so you have to watch at highest resolution. AND their names have changed over time, so it takes quite a lot of searching to find out that the "Equalizer" module is now called the "Contrast Equalizer". There are other Equalizers, so don't be fooled! I guess I have a lot of experience with user interfaces of the throw-in-everything-the-developers-can-think-of variety, so I can get around it ok. The Color Zones module is very powerful for altering BW response.

The latest version of DT(darktable) has two new modules one of which pretty much eliminates the requirement of quite a few other modules and is specifically more powerful than using traditional channel mixer workflow and specifically addresses the workflow that tedious and long winded luminosity masking via channel mixer previously accomplished in PS. CI has now matched LR in terms of recovery tools ( misnamed for my purposes) - but DT beats them both in terms of granular control.

The problem is that both the inventor's AND Williams' explanations are not particularly transparent to new users who as you say have to struggle with the GUI /Interface.

I 'll check the user manual and book.

Thanks
Pete

 

[dougpeterson]

Peter,

If I want a smooth, realistic rendering, C1 does fine. If I want more drama, LR has a wider range of effects - clarity, texture, dehaze. C1 has finer color selection, but if I want to turn a blue sky black, or green leaves white, I can do that more easily in LR. (Edit: never mind, I just figured out how to do that in C1. I was unaware of the interaction between the Color adjustment blocks and the B&W block.)

In short, I can get to the overly dramatic look I sometimes want with a few sliders in LR. The clarity adjustment in C1 covers a much smaller range than in LR. And I haven't found anything in C1 with the same violence as dehaze in LR.

But I work each promising RAW with both converters, and about 1/3 of the time, I prefer the C1 rendering. I have never been successful adjusting one converter's output to match the other's. It is, of course, my interaction with them, and someone else might find them interchangeable, or one a clear superset of the other. I just find them different.

Now that I'm using Fuji x-trans more, I'll spend more time in C1, and probably learn to get all the effects I want.

Best,

Matt

Do I have to drag you kicking and screaming to one of our DT Capture One Masters classes?? We just finished a round in NYC but should have another in a few months.

There is zero chance that you are getting the most out of C1. I say that with confidence because I've been professionally instructing the most advanced classes on Capture One for more than a decade and I still learn new workflows, tips, and tricks every month.

https://www.dtcommercialphoto.com/capture-one-training/

 

[Paul2660]

Would love to see such a class come to the middle of the US, St Louis, or Dallas?

Always seem to west/east coast oriented.

Paul C

 

[Speedgraphic]

In my opinion, "more smoke and mirror" talk from a camera manufacturer. Pixel shift technology has been around a long time (20+ years) and I'm yet to examples that impress me without inherent limitations of the process. That said it looks great to state in marketing materials 1,000,000mp captures. Let us also consider the physical limitation of diffraction, optics, file size, camera functionality, ISO noise limitations, etc., etc.

The cry is always wanting more and more quality for less $. Rarely have I see that actually pan out when the finer points (limitations) are read and fully understood and its reduced to marketing hype. This of course is all just tech talk, and if one can't make an artistically interesting image with the device, the entire argument is mute (unless its just for scientific use).

So Fuji...show me your cards. Yawn...


Robert B

Some photography is more scientific than artistic. When doing reproduction and scanning work, pixel shift aka multi shot is the way to go. If they CAN add the feature to the camera, why not? I camera scan with an S1R and em definitely enjoying the true RGB capture plus the resolution boost. This would be even better with the GFX I'm sure. Why say no? Shooting 8x10 has limitations too. I still do it.

 

[Christopher]

And I have used C1 for ages and would still say there is no easy and fast way to replicate dehaze. I have talked to enough C1 teachers who in the end agree, even if it takes some time.

I’m not saying you can’t get similar results with lot more effort... and I think it’s super important to learn as much about a program as possible. There are just so many amazing ways to edit an image with C1.

I would prefer if they fix the tangent support ASAP, as this shortcoming is embarrassing.

Do I have to drag you kicking and screaming to one of our DT Capture One Masters classes?? We just finished a round in NYC but should have another in a few months.

There is zero chance that you are getting the most out of C1. I say that with confidence because I've been professionally instructing the most advanced classes on Capture One for more than a decade and I still learn new workflows, tips, and tricks every month.

https://www.dtcommercialphoto.com/capture-one-training/