[FFmpeg-devel] [RFC] 5 year plan & Inovation

Tomas Härdin git at haerdin.se
Thu Apr 25 01:50:02 EEST 2024


ons 2024-04-17 klockan 15:58 +0200 skrev Michael Niedermayer:

> * ffchat
>     (expand into realtime chat / zoom) this would
>     bring in more users and developers, and we basically have almost
>     all parts for it already but some people where against it

You mean inventing a new chat protocol? If so then please don't. We
don't need even more fragmentation in that space - heretical projects
like Matrix are bad enough. It's also widely out of scope.

> * AI / neural network filters and codecs
>     The future seems to be AI based. Future Filters and Codecs will
> use
>     neural networks. FFmpeg can be at the forefront, developing these

New codecs are better developed as separate projects. The IETF has a
group dedicated to codec development. In fact they seem to be working
on ML codecs right now: https://datatracker.ietf.org/wg/mlcodec/about/

> * [your idea here]

A large long term project that would help immensely with security is
moving to a proper parsing framework, rather than the present shotgun
parsing approach. But this might be such a large undertaking that it's
better to start from scratch.

A more modest proposal is to improve subtitle support. Streaming
support could also be improved, and would be very much with the times.
The fact that we can't pass MPEG-TS through unmolested isn't great.

/Tomas


More information about the ffmpeg-devel mailing list