After processing my RAF file with DxO PureRAW 3.2.0, I have observed notable disparities in the edge rendering when compared to the out-of-camera JPEG. Regrettably, it appears that the installed profile named “Fujifilm X-T4 + FUJINON XC 35mm F2 (RAW)” has not effectively corrected the distortion in the image.
Are you sure? The out-of-camera JPEG is not the gold standard for lens corrections. DxO tends to produce a different rendering based on its own measurements of how a given lens behaves. If you see objects that are clearly distorted in the output from RAW, that’s a problem. Can you provide samples?
Since I am unable to share the original photos where I initially made this observation, I created new images to further investigate the issue. I discovered that a door frame serves as an excellent subject for testing distortion correction.
Initially, I expected the out-of-camera (OOC) JPEG to be the most accurate representation since Fuji is knowledgeable about their lenses. However, it turns out that you were right, and relying solely on the OOC JPEG is not reliable. I have become accustomed to a view that is evidently distorted.
In the sample images provided below, you can observe significant distortion with the XC 35mm F2 lens, making it challenging to work with in this regard. The OOC JPEG clearly exhibits flaws, but I still believe that PureRAW overcorrected the distortion, whereas the OOC JPEG did not correct it enough.
Upon analyzing the left line using Affinity Photo, I conclude that the correction applied by Capture One appears to be the most accurate for this particular shot.
capture_one|690x460
ooc_jpeg|690x460
pureraw|690x460
without|690x460
You can always share your original photos through services like Google drive, wetransfer and others. Original files can help to locate the issue(s) that might be caused by any link in your processing chain.
The images attached have enough resolution to see the problem.
Maybe this is just another case where PureRAW is not enough, because I saw that I can lower the intensity of the correction in PhotoLab an around 90% it looks right. Somehow this particular lens profile tends to over-correct.
If this should be the case, DxO might want to have a look at specifics.
@sgospodarenko , can you check this with @CommanderSteps and your colleagues?
@CommanderSteps Hello, Please, provide me with original files with the issues so we can fix it if needed?Upload them here - http://upload.dxo.com and let me know when ready. Thank you !
I uploaded the original files.
Compare the Capture One correction with PureRAW. Capture One looks more correct as it has a straight line.
btw - does DxO PL6 decode focusing distance from X-T4 raw files ? it does not, for example, for X-H1 raw files and as a result it can’t properly apply its own optics correction UNLESS you manually enter focusing distance in “Focusing Distance” tool
for example ( X-H1 + XF35/2.0 )
$0.02
1 Like
I downloaded a X-T4 raw from Imaging Resource ( it seems btw it is alive again ) and no DxO PL6 can’t decode focusing distance from X-T4 raw files → you need to make sure that it is not the reason for some ill effects… I am not saying it is but do check !
1 Like
Thank you for pointing that out. Yes, look like its indeed part of the problem.
I’m glad that Fuji X-Trans finally came to PhotoLab, because I wasn’t able to use it since I switched to Fuji and they hadn’t support. But it does not look like that the support is good enough right now.
My hopes are with PhotoLab 7 
then you are out of luck… it seems that nobody is capable to decode focusing distance in Fujifilm raw files and DxO does not want users to be able / to opt to use optics corrections data that is already present in Fujifilm / or any other raw files/ raw files ( if users are not happy w/ what DxO makes or DxO simply does not have a module in place and ~never will )
With some cameras the problem is fixable, but DxO dragging their bottom part, for example for Sony A7R2 it is known for a long time that focusing distance is stored in exif tag = “0x002d FocusPosition2” and formula for a full frame mode for example is ( 2 ^ ( {0x002d FocusPosition2} / 16 - 5 ) + 1 ) * {lens focal distance} / 1000 … but no, DxO still refuses to implement and as a result you MUST manually fix this for every shot in DxO UI ( or generate DOP files first and write a script to mod .DOP files based on that calculation ) or you might end up wasting their optics correction
color that STUPID !
1 Like
Wow, that’s an very interesting insight. Thank you for sharing that with me.
Return of the attaboy?
1 Like