There is a distortion bug in WMR OXR using SS 50, 55 and 60 with Reproj. This is most likely what you’ve seen.
However using TAA90 is not maximizing it at all because the hardware you’re running and the shader code can give you better at less cost. That’s the whole point balancing these settings.
I’ve posted some finding about this here too:
EFIS resolution depends on render resolution, not post-processing resolution.
And some render resolution facts here:
Comparing Reverb G2 and Valve Index resolution and what does it mean with SteamVR and OpenXR
In other words, if you’re shooting for best EFIS legibility, you rather use TAA100 and adjust OXR SS until your system can’t deliver. This will be much better than TAA90 always.
Otherwise if you’re shooting for best outside details, you rather use TAA70 and OXR SS 100. You might even be surprised with TAA60 + OXR 100 with analogue gauges aircraft (and get much more room for raising other things).
The 70/70 balance is running fine with the 2070S without reproj. and is good overall for inside/outside view details on the Reverb G2, I have no doubt it will run really good on the 2080TI with lots of leeway.
The key in the end is that you’re loosing details for no fps gain in using TAA90, and using TAA70 (with higher OXR SS) or using TAA100 (with lower OXR SS if your card can’t deliver) will give you better visuals and better fps altogether.
NB: Motion Smoothing/Reproj. is computed on the OXR SS render res and when enabling this, you’d have to balance the SS value maybe lower in order to give it room