Most probably I allready do know the answer, but I want to make sure that I follow the best (possible) workflow.
My camera has the option to use a built-in/internal correction of the lenses. To be honest I’m not sure whether this correction affects only JPG or also CR3. As I do use PL in post process, I record my images in CR3. And I have turned of all camera internal image corrections. Is DXO PL capable of handling the difference if the auto corrections in the camera are swithced on or off? Or are the Modules desigend for one one state of camera-setup? I guess the best outcome would be to switch all camera correction off, right?
Good questions, @Flrbb and I admit that I never tried to find out. What I see from my Canon files in Lightroom (and PhotoLab) is the following:
JPG files show straight lines while CR* files show barrel and pincushion distortions
→ the assumption is true that Canon applies corrections to JPGs only
Whether PhotoLab is smart enough to add/not add corrections to JPGs depending use of camera internal corrections, I cannot say. Although I record (small) JPGs in parallel for blogs and shares, I only work on RAW files in PL and Lr.
You can easily tests this by comparing JPGs corrected by Canon or PhotoLab. The easiest way is to set distortions and crop so that they don’t limit the image to the original aspect ratio or cut off protruding corners or edges.
Only Canon’s own DPP software will read the lens correction data that a Canon camera can write to a .CR2 or .CR3 RAW file. Other RAW convertors cannot read such Canon specific data.
With JPEGs though, if the lens correction data is set in-camera then those corrections are baked into the image. That means if your default PL profile for RGB files is one that includes lens corrections PL will correct already corrected data.
I don’t bother with any of the in-camera lens corrections settings in my Canon 90D. I let PL ‘fix’ things when I process the .CR3 files. I never edit any of the SOOC JPEGS in PL. I only use those files for rapid browsing.
In general, raw processors don’t/can’t read camera maker lens corrections (there are some exceptions, obviously including the maker’s own software), partly because the manufacturers hide this kind of data in proprietary maker settings.
I doubt that any camera file contains actual correction data. Corrections are applied based on lens model and - setting metadata and actual correction parameters stored or loaded (like in a DxO module) in(to) camera firmware. Canon’s EOS utility allows to add and remove such lens info and if the respective data is missing, corrections cannot be enabled and then, RAW and JPEG files show uncorrected images.
If camera manufacturers made correction parameters available to third party apps, we’d need no manufacturer software or apps like PL, C1 or Lr.
Anyways, the best use of DxO modules happens when they are applied to RAW files, the second best is their application to JPEG files. JPEGs? Yes, There are modules for specific camera/lens combos that have two modules, one for RAW and the other for JPEG. They are used withJPEG for DxO specific corrections that manufacturers don’t do or that DxO deems better than the manufacturer’s.
I just did a test for a photo for which I had both RAW and JPEG saved, and I have let Photolab apply its intended lens correction to both. For that photo the cheap RF-S 10-18mm les was used, so significant vignetting and distortion applied. Also in-camera correction was enabled so the JPEG file already that applied to the image. What I can see is this. Photolab fixed the RAW as expected. Photolab apples a definately different correction to the JPEG file and that correction was taking into account the fact that the in-camera correction was already applied to the image. The correction applied to the JPEG only very slightly further improved distortion and vignetting to better match the corrected RAW,. It also applied sharpening to the edges of the frame.
So my conclusion is that the DXO module made for JPEG is intended for an already corrected JPEG, and allow to bring that JPEG to an higher lever of correction, beyond what the camera already did. But for sure it does not cause a double correction to be applied which kind would cause the inverse aberrations to appear in the picture (white vignetting and pincushion instead of barrel distortion).
Canon’s corrections often seem to work to less than 100% by design. possibly depending on overall lighting conditions and/or exposure. This helps to prevent e.g. overly brightened and noisy looking corners in an otherwise dark scene.
Some of the respective metadata can be found with ExifTool:
[MakerNotes] Peripheral Lighting Setting : On
[MakerNotes] Chromatic Aberration Setting : On
[MakerNotes] Distortion Correction Setting : On
[MakerNotes] Digital Lens Optimizer Setting : On
[MakerNotes] Peripheral Illumination Corr : Off
[MakerNotes] Auto Lighting Optimizer : Off
[MakerNotes] Highlight Tone Priority : Off
[MakerNotes] Long Exposure Noise Reduction : Off
[MakerNotes] High ISO Noise Reduction : Off
[MakerNotes] Digital Lens Optimizer : Standard
...
[Composite] Light Value : 5.6
If some of the above relates to what actually happens, “improvement” should/can be seen with DxO Modules being applied to JPEGs.
Usually, we can proceed with corrections in any sequence and the confidence the PL “does the right thing”. Still, best practices exist, e.g applying profiles, LUTs, optical corrections (and possibly others) before anything else.