RF 24-105 f4-7.1 IS STM distortion correction at 24 mm

[orewa]

It is well known that this lens has a very heavy barrel distortion at the wide end. Camera corrects it pretty good, however I'm trying to find an answer how it is corrected - is it originally (distorted) 24 mm and after correction it turns into say 28mm image or Canon (knowing this issue) made it slightly wider (maybe about 22mm) and after correction the image turns into real 24mm.

Worth mentioning that with this lens you can't turn the distortion correction off in camera, so Canon definitely was aware of that "behavior" of this lens.

[Tronhard]

I would counter that this could be expressed not an awareness of a flaw that required correction, but rather as a design characteristic that is a device to create lenses that are more easily engineered and cheaper but incorporate firmware and software counterpoints to complete the process of lens to image alignment as an integral package.
Now that lenses are governed more and more by firmware, and interface with the camera via other software components, this becomes possible when, before this digitization, it all had to be done by pure optics.  One could describe is as a form of applied computational photography.

This now appears in several lenses from Canon.  I have it in the RF 24-240 superzoom camera, where the optic has an uncorrected FoV view of 21mm, but with correction comes out at 24mm.  It is not limited to non-L lenses: my RF14-35 f/4 L uses the same methodology.

One could argue that finally dedicated cameras are starting to include and apply devices that have become the norm in cell phones for a long time.