Hi everyone,
I just received my X-E5 and took some test shots to evaluate the performance, comparing it to my Fujifilm XPro 2 (both RAWs attached). X-E5_test.RAF (36.9 MB) XPro2_test.RAF (21.7 MB)
I noticed that the color temperature and tone in the same shot are significantly shifted between the two cameras. I am using RawTherapee 5.12 on Windows 10. I suppose this issue is similar to this one, reported here a few years ago: Shifted white balance values for Fujifilm X-E3
From this discussion, I thought adjusting the color profile might solve the issue. But even if I correct the colors with a DCP profile, the problem persists, no matter if I use the adobe standard or the one I measured myself.
This issue is however not occuring in Capture One, which offically cooperates with Fuji and uses the factory profiles by Fujifilm.
So I have two questions regarding this issue, which some of you might be able to help me with:
Why is the DCP profile not resolving the white balance issue?
Who may be able to implement the fix into RawTherapee, similar to the last time this issue arose?
Since I have a color reference card and an X-E5, I would happily provide the corresponding DCP profile to distribute with the software.
At the moment no profile of any XTrans 5 sensor is available by default, which may make RawTherapee unattractive for many casual Fuji users.
I used the standard dcp profiles for each from Adobe… and then picked WB off the cube… the wb coefficients are similar but the tint has a very different setting and the two files report quite a differnet wb value…How ever if I disabled the RT tone curves and applied an all adobe color profile base, look and tc using the updated profiles from the latest DNG converter…Things didn’t look too different…
In RT if you used the camera profile vs automatched vs adobe with or without the look you would get quite different versions of the file…
I also notice one camera used metadata for the lens corrections and the vignette and exposure were corrected but not on the other camera…setting them both to metadata also made a visual difference…
So maybe get that proper adobe profile, check the lens corrections and check again???
I tried already using the standard adobe profiles for both cameras.
As you wrote, the color rendering is quite different, but the white balance does not change significantly.
By the way, I used the Capture One image output to visually benchmark the colors. The automatched as well as the adobe profiles produce excessively flat colors. With my own makeshift color calibration, I could significantly improve the color rendering and I am optimistic, that I can solve the issue on the next sunny day with a proper D65 source.
It is only the numerical value of the tint and wb which is puzzeling me.
The reason why I used two different lens correction methods is simply that the lens for the X-E5 just released a few weeks ago and the lens correction profile is not available in RT jet. So I decided to just use the metadata for now and fix the colors and wb first. I tried several permutations of the lens correction options and concluded that the CA-correction doesn’t affect the white balance.
It is indeed interesting, that the wb-coefficients are virtually identical for both cameras. So I assume we are applying the same correction, but somehow RT has to set ridiculous values for wb and tint to get there with the X-E5.
Indeed, if I convert the .RAF files to .DNG first with the adobe DNG converter, I recover the same white balance for both cameras in RT.
Perhaps it is a bug in the interpretation of the .RAF file the X-E5 generates?
I tried to compare the Exif of both images, but at first glance I cannot spot a significant difference.
