RPi3 hardware deinterlace is being actively worked upon and public testing of code should happen soon. Support for 3D is something that can probably be done and one of the Pi Devs has a 3D TV so while it's niche and not a priority for the Foundation, personal interest will probably make it happen eventually. The advantage these "missing" features have is, they also benefit RPi4 support. Software HEVC optimisation for RPi3 is a little different and IMHO unlikely to be reimplemented. One objective of the new video stack is to get everything upstream to make RPi support consistent and better in all distros: previous HEVC support only ever worked in LE and OSMC because we were prepared to add 50,000 LOC patches that normal distros with mature package/patching policies refused. HEVC is probably not technically impossible to do in the new video pipeline, but it would (again) need large unorthodox changes that would never be accepted upstream, and the changes need a ground-up rewrite not adaptation of the previous approach so it's a considerable effort. Meanwhile RPi4 has native hardware decode that works nicely, and while the code is not all upstream yet, ffmpeg is the main missing bit and that effort will start sooner than later. Forcing users to upgrade to profit the Foundation has never been an objective with any of the Pi developers, but upgrading to an RPi4 is the simple and obvious option for anyone who needs HEVC support (and you get 4K, HDR, HBR audio included too).
Matrix is K19/LE10 and there is a release thread. K20 is in a pre-Alpha state so there is no LE11 thread at this time.