The correction parameters are stored in the file metadata. The raw data are uncorrected. RawTherapee doesn’t know how to read such metadata, so it doesn’t apply those corrections.
Wouldn’t be raw if it was cooked.
I agree, but some cameras are alleged to serve RAW files that have been processed. IIRC, some Sony models applied noise reduction before saving the RAW image, for example.
LR does it automatically with my OM-D. RT does not do it automatically.
Lensfun works well.
Thanks for the feedback! Nice to know the facts.
The interesting thing is though: the distortions are not visible on the camaera’s display in liveview, as far as I remember
when I was taking the shot, I was looking for the wide-angle distortion and was not sure whether my lens is wide enough or had any distaortions at all
And one more thing: In LR there appears to be no way to undo the correction. there is the option tho correct the correction, but the result is that one part of the photo is cropped, if you try to undo the correction manually
Most, if not all, camera displays tend to show you what the JPG will look like. This is something to be aware of if you shoot raw. For example, the display/histogram may show highlight clipping when the raw will not actually clip (because of a high-contrast setting for example).
@betazoid first point:
http://rawpedia.rawtherapee.com/The_Image_Editor_Tab#Eek.21_My_Raw_Photo_Looks_Different_than_the_Camera_JPEG
2 Likes
Micro Four Thirds cameras have a lens correction profile (lcp) saved with the raw file, but not applied to the raw file.
Many software, including some third party software like Adobe PS and LR, apply this lcp automatically when processing the raw file. These images look like the JPEG, which in turn looks like the on-screen live view for all cameras when in live view.
Other raw software ignores the embedded lcp, like DxO and RawTherapee and RPP64. These processed raw files will not look like the live view (or JPEG) seen in the camera.
Can you prove that?
I have a lot of m43 experience and can confirm it from experience.
Proof is not something I have sought.
There is some information here, from 2011.
@T_N_Args, the correction parameters are definitely in the metadata (just use exiftool to see them), but I don’t think they come in the form of an embedded LCP file. Could this be the source of the misunderstanding?
Fair chance you’re confusing opcodes with LCP.
@Morgan_Hardwood I already asked but got no satisfactory answer.
Is there somwhere a libre SW capable to decode the opcodes? (not a hex dump)
@agriggio: Sure, maybe I gave the wrong impression when I wrote ‘lens correction profile’ and people mistakenly thought I wrote it’s a file within a file. m43 raw files do carry unapplied lens correction information, which is a lens correction profile in my book, no matter whether or not it conforms with some profiling file format.
Morgan asked a vague question and I thought the question was ‘can I prove that m43 raw files have lens correction information embedded but not applied’. After all, that’s what I wrote about the raw files.
ok, I think the misunderstanding has been solved. the problem for RT though is that LCP is a documented format, but the embedded corrections parameters are vendor specific (afaik), and we don’t know how to use them
Should those vendors want support and publish or help the developers to understand how to use that data, I’m sure it’d eventually get done.
Users of those companies’ products should let the company know that open is better 
That is not only categorically impossible, but it also does not adhere to facts.