[Debconf-video] irc chat with KarlH (probably not an issue for us)
Most likely we won't run into this, but just wanted to post it somewhere
for reference sake in case we run into it. Karl Heyes is pretty much the
only active icecast developer now and is almost always very responsive.
Here he was proactively so, pm'ing me, just to let me know. I don't
think this was the mplayer issue some people saw w/my test stream as
what Karl describes is not client specific and plenty of other clients
had no problems (or at least not this problem) with my test streams.
-Eric
21:46 -!- Irssi: Starting query in freenode with karlH
21:46 <karlH> are you working to setup theora streams?
21:46 <edrz> will be soonish.
21:46 <edrz> probably going to sleep soon.
21:47 <edrz> but hopefully tomorrow we'll have test streams running in
preperation for DebConf8
21:47 <karlH> well it's just that I've had a bug reported for low
bandwidth theora, fixed in 2.3.2-kh1
21:47 <edrz> low bandwidth as in how low?
21:48 <karlH> but if you have a .2.3.2 setup then I have a tarball for
that
21:48 <edrz> we've got a 9mbps uplink going out to our master relay.
21:48 <edrz> and only enough encoder cpu power to do one quality.
21:48 <karlH> not sure of the trigger point, but the recent improvements
in theora has shown this up more
21:49 <edrz> what are the symptoms?
21:49 <karlH> listener connects but receives nothing
21:50 <karlH> it has to do with the startup markers
21:50 <edrz> hmmm. i saw something that might have been that when I was
testing at home but assumed it was a configuration vagary on my part.
21:50 <edrz> though ... it wasn't that they were getting nothing.
21:51 <karlH> you get the http headers part
21:52 <edrz> my issue was more that the queue/burst seemed to have stale
data when I tried having very large queues on multiple levels of relay.
21:53 <karlH> haven't heard anything about that
21:54 <edrz> lowering all but the first mostly took care of it.
21:54 <edrz> sounds different from what you're describing.
21:55 <edrz> does what you're talking about happen for all clients? it's
definitely on the icecast side?
21:55 <karlH> yes
21:55 <karlH> yes
21:55 <edrz> i'll know more tomorrow.
21:55 <edrz> thanks for the heads up.
21:55 <karlH> ok
--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Reply to: