[Members] XSF roadmap

jehan at zemarmot.net jehan at zemarmot.net
Sat Nov 20 05:13:30 CST 2010


On Sat, 20 Nov 2010 11:47:22 +0100, Alexander Gnauck
<gnauck at ag-software.de> wrote:
> Justin Karneges wrote:
>> There are many issues here:
>>    - audio/video is hard
>>    - going around NATs is hard
>>    - wrong protocol (incomplete implementations, mismatched codecs, GTalk)
> I totally agree with Justin.
> I still think that some libraries for NAT and RTC under a liberal
> license with bindings to the major programming languages would help.
> The same applies to codecs. Most commercial software is using GIPS
> which has excellent codecs but is often incompatible to the free
> codecs used in Open Source software.
> Did anything change on GIPS licensing since they were acquired by Google?
> Most people use Skype because it A) just works and B) has excellent
> voice quality. We can fix A with interop testing, but I don't know how
> we can fix B. Is somebody on the list which has experience with free
> codecs?

Yes I was only worried about A) but B) is very true as well. I finally
manage to get one sound call with both side the other day (but video was
failing though, both side, so sound only) for the first time ever. I
don't know what was different from other times. I was on Pidgin under
Linux and the other side was on the Google talk soft under Windows.
Yet I could hear ok enough, but the other side had a lot of static
noise so we had to end soon because it was very annoying to her. And
with the same material, a call under Skype had good sound (but long
video calls on Skype crash my computer, it seems. The linux version does
not seem reliable, I have a lot of problem, I don't know for the Windows

If we want the technology to work, we must make call pass AND be of
good quality. Both are indeed needed!
Yet I don't know if the problem is only about codecs, but it is maybe
also about material support sometimes.

As for the codec, I think to remember discussions (in IETF probably?)
were going on about having a new codec derived from the current existing
ones because speex or others would not be that efficient for voip (or so
I think to remember being told, I have no knowledge on the matter). What
about this point?


More information about the Members mailing list