I generally never use aperture under diffraction limit.
Is photolab ai demosaiking and lens correction takes diffraction into account and make some corrections (never heard about this) ?
Hi
I don’t remember if DxO has said anything about diffraction.
One thing is certain, Lens correction counteracts the general weakness, which I consider to include diffraction.
Pascal
Including diffraction would, as I imagine it given the type of correction DxO applies, push back the diffraction limit (gaining 1 or 2 stops ?).
So ?
I dont know if PL do anything with diffraction, but some years ago I tested when I really could see the effect. As I use cameras with a 4/3 sensor it’s of actual interest. The test was with my old Olympus E-30 and the 50-200 f/2.8 - 3.5. As far as I remember you would be hard pressed to see it at f/8. At f/11 it was visible and at f/16 it was very visible.
my camera limit is about f/8.
And any response I’ve found about this till now is, … far beyond diffraction limit 
I don’t know what camera you have.
Photopills has a diffraction calculator that distinguishes between diffraction in a 100% view and diffraction when printing.
George
Check these posts by DxO staff member @Benoit (Aug 2018): https://forum.dxo.com/t/dpl-and-lens-diffraction/4463/5
and three posts later:
"
Consequently, as diffraction contributes more and more to lens sharpness when aperture decreases, its effect is taken into account in our dxo modules, and thus lens sharpness tool has an impact on it.
"
and
"
As lens sharpness correction already includes the action on diffraction, and as it would be quite costly to split in each DxO modules the part of the blur caused by diffraction, we don’t think about such a feature for the short term.
"
…and there’s more in the referenced post:
This said, when diffraction is strong (or more generally when input lens sharpness is very poor), our dxo modules limit their correction by default, as fully correcting the picture might lead to artifacts (and when a small detail is very blurry, there’s not that much to do to recover it…). Nonetheless, adjusting manually lens sharpness sliders enables to increase again sharpness.
…would be interesting to know the threshold when correction gets limited.
Years ago, when Olympus introduced diffraction compensation (deconvolution) into its Truepic engine, I compared Truepic VII to DxO software at the time using my sharpest lens. The micro four thirds cameras and lenses I use close the aperture down to f/22. Certain lenses will produce their sharpest results at f/5.6-f/8 in the center and around f/11 in the corners. But the best lenses are sharpest around f/2.8-f/4. So testing the 60mm macro telephoto lens, I found that up to f/16 DxO was very competitive with the Truepic VII out-of-camera JPEGs. Only at f/22 did the DxO software fall short. I doubt anything’s changed since then, given Benoit’s comments, but haven’t retested.
I doubt whether a distinction can be made between lens sharpness (or lack thereof) and diffraction.
From DPL and lens diffraction - #5 by Benoit
This said, when diffraction is strong (or more generally when input lens sharpness is very poor), our dxo modules limit their correction by default, as fully correcting the picture might lead to artifacts (and when a small detail is very blurry, there’s not that much to do to recover it…). Nonetheless, adjusting manually lens sharpness sliders enables to increase again sharpness.
George
I don’t think we can draw that conclusion, George. PhotoLab limits correction in both the specific case of diffraction and the general case of poor sharpness, but that doesn’t mean that it can’t distinguish one from another to some extent. Truepic can, and so can other image processors that employ deconvolution when the diffraction limit of the optic is known.