[FFmpeg-devel] Contributing to the project?

[Ronald S. Bultje]

Hi George,

On Thu, Jun 11, 2015 at 1:57 PM, George Boyle <george at thebuds.net> wrote:

> Hi Ronald,
>
> On 11/06/15 17:01, Ronald S. Bultje wrote:
> >
> > It would also be helpful if you could identify what you would like to be
> > doing, i.e. do you have any particular aspect that you find particularly
> > exciting or intriguing and would like to learn more about? E.g. you might
> > like one of video codecs or audio codecs or muxing formats more; you
> might
> > prefer decoding algorithms, encoding algorithms or optimizations more;
> you
> > might be more interested in the fundamentals of any one of those (either
> by
> > fixing algorithm bugs/extending algorithms to handle wider aspects of
> > existing formats, or by writing new codecs/(de)muxers from scratch), or
> > initially learning more through testing techniques (code coverage,
> fuzzing).
> >
> > (That sounds like a lot, but it helps point you in a particular direction
> > if we know where you'd like to go.)
> >
>
> Thanks for your reply. Yes, a more specific goal would be useful! In
> thinking about it, there are 2 areas in particular that I find most
> interesting, both originally stemming from being an avid user of the
> ffmpeg command-line tool.
>
> The first is the program itself, and the workings of the project around
> it - its pace, scale and robustness. In that regard, I thought learning
> about the testing system would be a good place to start. This is also
> the closest thing to a match with my existing experience. Plus, I do
> enjoy testing. I think the suggestion of doing some documentation would
> also be a good way to get familiar with the codebase.
>
> Second would be container formats and muxing/demuxing. The matroska and
> ogg formats spring to mind, but I'd like to learn about any of them. The
> idea of coding one up from scratch sounds very interesting, though I'd
> need to study up a lot first.
>
> Those would be the standouts, but I'd enjoy getting involved in any of
> the things you mentioned. I'd like to do some bug-fixing too, and I keep
> an eye on the tracker, but I heven't got enough knowledge to take one on
> yet. Hopefully I will soon.
>
>
> == tl;dr ==
> I was thinking of the following ramp:
> 1. Start with testing and docs
> 2. Learn about and work on container formats and (de)muxing
>
>
> Does that sound like a reasonable way to proceed?


Yeah, absolutely; most of us started with tasks to get us up and running
and familiar with this whole thing. You'll know when you're comfortable
doing that kind of stuff, and that's probably when it's the right time to
write your complete own demuxers (even if you don't think it is yet; it
should be a bit of a challenge, after all). There's always stuff to be done.

Good luck!

Ronald