Hi - I wanted to raise the question of using GPU-accelerated decoding from RV.
We recently found that we can get real-time playback of H.265 HEVC at 8k 60fps in Windows Media Player (of all things), which I can only guess is using Nvidia's NVDEC hardware decoding, as it doesn't work on a machine with a pre-Pascal GPU (M6000) but does on one with a GTX 1080.
I'd really like to use RV to have something scriptable, plugin-able, with decent tools for annotation, sync sessions, all that. But we're dealing with clips in the minutes-to-hours size, at 8k (and up) 60fps and H.265 is a good option for a lot of it.
As I understand it you use ffmpeg for decoding, and as far as I can see (https://trac.ffmpeg.org/wiki/HWAccelIntro) there is support there for the NVDEC extensions, if that helps.
I can imagine it may be a pretty significant architectural difference for RV to let the hardware decode the image file directly into a target buffer on the GPU and skip the CPU+RAM altogether, and I don't know if you do other things on the CPU that require the image on that side, but I'd be curious if it's something you've looked into and if others are asking for anything like this.