...
But surprise surprise surprise... the tensioner is moving up and down as the crank is spun. Between the chainring and the cog, there's one or more eccentricities. I will investigate, remediate, and report back.
When I change chainrings, there is a lot of slop in the holes in the crank spider arms and the chainrings. Thus, if I just put a ring on and tighten the bolts, the chainring will not be concentric with bottom bracket spindle. I have to semi-tighten up the bolts so that the chainring and crank arms can be shifted in position with a little pressure on them, and then it is an iterative process of trying to get the chainring in exactly the right spot where the chain slack has minimal change when I spin the crank backwards before I tighten up the chainring bolts.
I use different chainring sizes for touring versus riding around near home (36T for touring, 44T for riding around near home). Thus, chainrings get changed on a regular basis, so I am accustomed to doing this. And when I change chain rings, add or subtract four chain links.
I do not have a tensioner like you have, thus I need a small amount of slack in my chain, see photo. It is an old photo, I was using the 44T chainring as a bash guard, but now have a real bashguard sized for the 36T chainring. The chain looks tight in the photo, but if you hold a straight edge on your computer screen, you can see some slack in the chain.
That said, I would not expect an off-concentric chainring to have any noticeable feel when pedaling with a sprung tensioner. But perhaps your tensioner movement for part of the time causes something else to rub?