I'm taking a phased approach with an agentic host repo that drives the source changes in FFmpeg. generally, i do a "sprint" for adding features, then a sprint reorganizing for coherence with an eye on possible upstreaming.
The fact that there is no newer bug ticket for it in 12 years shows is this a purely academic feature nobody wanted.
And why should they? Most Blu-ray players support the old format (SRT); converting it to PGS gains almost none of the upsides of the newer format (screen placement).
oh yes, every commit is fully open about Claude Code + the model.
honestly, it would've taken me about 2 years to do this otherwise.
there are many FATE tests added and a lot more to add when the hardware verification bits i'm planning are done.
edit: do you know if upstream have ever taken an AI assisted commit? i've not seen anything yet and would hate to be controversial or a burden on hardworking reviewers!
8 comments
is the most recent.
I'm taking a phased approach with an agentic host repo that drives the source changes in FFmpeg. generally, i do a "sprint" for adding features, then a sprint reorganizing for coherence with an eye on possible upstreaming.
edit: plus the blu-ray subtitle encoder itself is new...
And why should they? Most Blu-ray players support the old format (SRT); converting it to PGS gains almost none of the upsides of the newer format (screen placement).
but i do enjoy antialiased blu-ray subtitles of course YMMV :)
honestly, it would've taken me about 2 years to do this otherwise.
there are many FATE tests added and a lot more to add when the hardware verification bits i'm planning are done.
edit: do you know if upstream have ever taken an AI assisted commit? i've not seen anything yet and would hate to be controversial or a burden on hardworking reviewers!