Fuji RAW and Lightroom

[pluton]

IMHO LR demosaicing of rafs is still inferior to C1 and Iridient.

I agree, and if I have a shot where I need to extract maximum sharp edges from tiny details I will use one of those or Photo Ninja.

[PeterN]

IMHO LR demosaicing of rafs is still inferior to C1 and Iridient.

I am really curious to your findings. Would it be possible to share examples procesed with the recent versions of the software? You use Fuji a lot more than I do. I only use the x100v during biking trips and street photography. In addition, my eyes are probably not as good anymore to see all subtleties.

[PeterN]

Suggested starting Lightroom sharpening settings for X-Trans 16MP:
Amount: 17
Radius: 1.0
Detail: 74
Masking: 0
The basic idea is to be low on the 'Amount' and high on the 'Detail'. If you have the feeling to increase sharpening, increase the 'Amount' slider, but only a bit.  Going from 17 to 30 is a big dose of added sharpening. Don't know if this applies to the newer 26MP X-Trans sensor.

Thanks!

[Eddie Draaisma]

Especially interesting for Fuji X-trans is the "Enhance details..." found under the Photo menu of Lightroom Classic.

It is basically a much more elaborate demosaicing process (trained with machine learning), creating a linear .DNG file next to the raw file.

IMHO it does a very good job on X-Trans files.

[PeterN]

Thought to share 2 full-size jpegs generated from Lightroom 10 and Capture One 20. Photo taken with Fuji x100V. ISO 160 - 1/30sec - f8. Both processed with Fuji Velvia, HL/S correction. I used Topaz Sharpen AI on "auto" settings for sharpening to get an equal baseline in that area.  Unfortunately file size is limited. I can send full-size jpegs, if someone wants to study the details.  I tried to spot differences but failed to do so (details in the sky are comparable when highlights are dampened a bit more in Lightroom.
BTW: I took this photo to explore the possibilities of Luminar and Photoshop for sky replacement.

Lightroom 10.0 example:


Capture One 20:



I haven't checked for other types of photos, especially portraits (which is one of C1's key strengths IMHO)

[Anthony]

Try comparing the two in Photoshop by layering and selecting Difference as the Blend mode. This does show differences, which I can post if you are willing for me to do so. I also noticed that you posted them in Adobe RGB, which some monitors may not represent properly.

There has been much discussion amongst Fuji users about Adobe's processing of X-Trans files. Many have switched to Capture One or Iridient. For example, here is a discussion on an official Photoshop website. https://feedback.photoshop.com/conversations/lightroom-classic/camera-rawlightroom-classic-fuji-xtrans-support/5f5f45684b561a3d424a11b0

Adobe itself recognises that there are issues with its X-Trans processing, and when releasing Enhance Details stated it could "resolve issues that some customers reported with their Fujifilm X-Trans based cameras".

A problem with Enhance Details is that it creates a large additional dng file, in the case of a Fuji file I tested this was 100MB greater than the original raf (which was, of course, retained). So it is not something one would want to use for every image; and I don't want to have to think about whether an image would benefit from it before deciding on the processing workflow.

Photo Ninja and Iridient Developer do a good job on X-Trans files. Iridient also created X-Transformer specifically for the purpose of demosaicing X-Trans files and creating dngs for further processing in LR.

Having said all that, I prefer to leave scientific comparisons to the experts.

[PeterN]

An interesting method, Anthony. I am curious to your findings. I could see differences in colors and aberration handling but no artefacts. Was I wrong? Feel free to post your findings since I may misinterpret what I see. I am curious what yours are.

BTW, I switched to Capture One anyway and do not miss LR at all. I only use Photoshop for specific corrections that are hard to o in C1 and stacking (focus, exposure, panos). It takes some time to get used to C1 but after "getting it" I like it better than LR. I still have to check if Affinity Photo can replace PS in my workflow.
I am also considering a workflow for NEF conversion that starts with PhotoLab4. I am impressed by its capabilities for processing NEF files.

[Anthony]

Thanks, Peter, this is the result of layering one image on the other and selecting Difference as the Blend Mode. I also made a Levels adjustment to brighten the image.

I am not really sure what this indicates as to artefacts as colour differences are picked up. However, on closer inspection I think that in the C1 version I see more detail in the grass below the bush by the river on the lower right; but it is not easy to tell and could relate to specific processing settings.

I agree with you generally about C1, and I like that it has layers with luminosity masking, which I use quite a lot.

[PeterN]

Thanks for sharing and analysing, Anthony. Much appreciated! Your blending looks a bit different than mine. Don't know why.
I turned sharpening off in C1 and LR but just noticed that I used a different clarity setting in C1. Sharpening for both images set to auto in Topaz Sharpen.

[Anthony]

Thanks for sharing and analysing, Anthony. Much appreciated! Your blending looks a bit different than mine. Don't know why.
I turned sharpening off in C1 and LR but just noticed that I used a different clarity setting in C1. Sharpening for both images set to auto in Topaz Sharpen.

Actually, I think I may have discovered a new art form!