[pulseaudio-discuss] Ideas for GSoC 2013

[Tanu Kaskinen]

On Tue, 2013-04-16 at 16:17 +0200, Michele Bussolotto wrote:
> Dear Sirs,
> 
> I am Michele Bussolotto, an M.Sc. student in Electronic Engineer. I would
> like to partecipate to this edition of Google Summer of Code and I will be
> very glad if you could give me an opinion about a project idea.
> 
> In particular, I would like to pay attention to problems in audio
> transmissions between two devices, like an IP transmission using for
> example module-rtp-send and module-rtp-recv. I am wondering if exist any
> mechanism to avoid underrun/overrun due to clock drift, to maintain
> synchronism between two remote devices, and if could be interesting to
> improve the performance about.

Regarding clock drift handling, module-rtp-recv does that by applying
adaptive resampling, see rtpoll_work_cb() in module-rtp-recv.c.

Fixing the network streaming bugs that we have (especially in the tunnel
module) would be great, of course. In order to fix them you need to be
able to reproduce them, so do you suffer from some particular bugs?

When considering this as a GSoC project, one thing that worries me is
that the amount of work that it takes to fix bugs is hard to estimate,
if you don't know what exactly is wrong in the current code. Having a
very vague schedule makes it hard for the mentor to assess whether the
student should be passed or not, which may cause the mentor to pick some
other project. I don't want to outright reject this project idea,
however, because the network streaming problems have been there for a
long time and cause lots of complaints, but nobody has so far had the
motivation to actually investigate and fix the problems, so getting
someone to work on the problems would be great.

-- 
Tanu