My reasons for thinking about Babylon are, first of all, the ones you mention. It seems more stable, the forum (and you!) responded quickly and helpfully, it seems like it does a lot, and it seems straightforward, with good support (though not terrific documentation).
I built a demo showing VR video running in a webVR site using AFrame. Aframe has its own style, and one must go beyond that to three.js (& more, sometimes). In particular, it was hard to get VR video to work. It required an external component, and some fixes to that component and its documentation to make VR video work. The upgrade to v. 9 then broke the code. (External components do not upgrade automatically).
I tried react360, but it did not display VR video well, required node.js all the time, and it seemed too much trouble to go that route.
I’ll be writing games, displaying text and graphics, using text-to-speech (hopefully), and showing framed and 180 and 360, 2D and 3D, photos and videos. The simpler, more powerful, better documented, better supported, less buggy, and more stable the code is, the less time I’ll spend programming, and the more time I’ll spend creating. So I’m hoping some investment in Babylon will prove a good choice.
I’m the organizer of NYXR - the New York eXtended Reality Meetup, so I’m very interested in XR, in which I include not just xR (xR = AR + MR + VR), but also all the other technologies enabled by the exponential explosion of digital technology, such as AI, 5G, wearables, IoT, and derivative technologies like robotics, autonomous vehicles, and cryptocurrencies PLUS the content it enables: good, like VR games, and bad, like deep fakes, and surveillance. However, in terms of xR programming and content, I’m mostly working on VR right now, and waiting for AR and MR to evolve. I was an early adopter of the HoloLens, but my tests convinced me that MR still needed time and deeper resources than mine to create the infrastructure to create anything interesting.
Thanks for your help and comments!