Hi guys,
I ran some tests. Here is my protocol:
# Server, start it first
liquidsoap 'from_live = fun(a,b)->b ; to_live =
fun(a,b)->add(normalize=false,[sequence([blank(duration=10.),b]),
fade.final(duration=10.,a)]) ;
output.pulseaudio(fallback(track_sensitive=false,[input.harbor("foo.ogg",port=9876,max=30.),amplify(0.2,noise())],transitions=[to_live,from_live]))'
# The clients streams "one, two, three, four, ..."
liquidsoap 'c = ref 0 ; next = { c := !c + 1 ;
request.create("say:#{!c}") } ; s = request.dynamic(next) ;
output.icecast(%vorbis,port=9876,mount="foo.ogg",s,fallible=true)'
Without the max=30. parameter in the input.harbor(), I can only hear
the stream starting at 4. The logs are pretty clear and verbose: the
buffer becomes overfull and data is dropped. Without the max parameter
set to 30., no data is lost.
This test may not reflect your situation. In particular, it could be
that you have plugged an output.dummy() on your input.harbor()
precisely to avoid those noisy "Buffer overrun" logs. (Every now and
then, we advise people to do this, it's not a very good situation, but
your example shows that it's not obvious to make a default choice.) If
you have an output.dummy() it will consume data from the buffer as
soon as there is enough, and so the max parameter will become useless.
In short: Romain's advice is still good, but make sure you don't have
an output.dummy in there.
Please tell us how this works for you. I don't think we'll change
anything before 1.0.1 but after the release we should do some serious
thinking to try to get better default behaviors.
Cheers,
--
David
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Savonet-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/savonet-users
