Transcoding is no joke, for nostr:npub120zh308gs69j3vwgvvrzgglv58vsk3q0vef43uajyu0lc3mtldsq2p33r2 Stories, I decided to offload the transcoding process to your device so servers aren’t overwhelmed and the UX trade off in terms of users having to wait is ok

This doesn’t solve the quality issue, either we prioritize quality over disk space which benefits UX or prioritize disk space over quality which benefits instance admins.

I’m leaning towards prioritizing disk space for now, and am looking into HLS + delayed backend processing to solve this long term