Calm discussions are great
What I meant regarding the 70-200 and 50 on the R5 vs R was that I didn't see any improvement – but I do see the imperfections that were there on the R become amplified on the R5. That's not to say their performance on the R5 is any worse, but I am better able see the imperfections on the higher res sensor because I'm looking more closely at the image at 1:1 magnification on the R5 than I am on the R at 1:1.
The 15-35, I almost returned it when I was using it on my R. I'm glad I kept it and tried it on the R5. I do think they may have changed the microlens design on the R5 so that it improves corner performance for extreme wide angles, but I don't have the testing acumen to prove it. I'll have to leave that to the brick-wall tripod testers out there (for which I am thankful, I don't say that in a derogatory manner).
If there are any optical physics gurus in the audience: does sensor stabilization have a more dramatic effect in the corners for extreme wide angles? I know the movement of the camera is the same between lenses, but on an extreme wide angle, the same amount of movement at the camera should result in the image blur covering a larger distance on a more distant target. So even though the IBIS correction will be the same at 15mm as it will at 50mm for the same camera movement, more of the target image passes across the same corner of the image at 15mm versus 50mm, and that image corner will be less-detailed to begin with and blur will have a more noticeable effect. For example, if I put the same tree in the corner of my image at 50mm and 15mm, I will see more detail in the branches at 50mm than I do at 15mm. So a small amount of blur will have the same effect on both images, but since the subjects contain different levels of detail, the 15mm shot should appear less detailed when blurred, thus giving the impression that IBIS has a bigger impact on the corners the wider the angle of view.