First off, apologies for the flurry of posts, I know I tend to do this sometimes. XD
So I found a way to improve calculation of quality for autoquality. I have updated the code which can be found in main.txt in the link below. What I am doing is iterating through each parent object to find its scaling and factoring this into the calculation of the overall quality scaling. I purposefully exclude the top level parent canvas, as this is oftentimes an overlay, and can have weird, arbitrary scaling. I'm not sure how you might want to handle that, though. This can be very helpful if a SuperTextMesh object is embedded in layers of prefabs which are scaled to arbitrary values.
