Generating marker previews and other tasks is quite slow on the tiny quad-core ARM processor of my Raspberry Pi 4. Has anyone managed to get hardware acceleration working with a Stash instance running on an RPi4?
I’ve tried adding -hwaccel v4l2m2m as additional output or input arguments, but nothing seems to work. I have to admit, my knowledge of FFmpeg command line options is pretty limited.
Details:
OS: Raspberry Pi OS Lite 64-Bits (Bullseye)
Not using Docker: I just run the Stash binary directly (v0.28.1)
FFmpeg: version 4.3.9-0+deb11u1+rpt1. Installed from apt
My main priority is marker previews, but if could also improve transcoding, that would be a great bonus. Any advice or examples of a working configuration would be greatly appreciated!
This is the issue to follow for VAAPI and general hwaccel for generation, I don’t have a new enough Pi to test V4L2 but will def look into it once they’re back to reasonable prices
