Hi guys,

2011/4/28 Brandon Casci <[email protected]>:
> Increasing the buffer size has seemed to make a stable stream last a bit
> longer.

This solution makes sense: if the buffer is big enough, the drift
won't be noticed before the end of the broadcast, assuming it's a live
show. But this should apply equally well to liquidsoap and icecast.
The only difference that I can think of is that we don't buffer
encoded data like icecast but decoded one (hence the buffer parameter
in seconds). This may have interfered in your observations, and may
make a large buffer too memory-consuming. (I think we should change
this eventually, it's a matter of finding a good reason to do this
rather than something else...)

> Geography matters for some people. I don't think that's an issue for
> liquidsoap though. A few icecast servers in different locations for live
> inputs might help some people out. They can choose one that works best for
> them. They could be configured in various ways so liquidsoap only has to
> worry about pulling from a single URL.

That, however, does not look like a solution. If it's a clock drift
(as in, something that could be fixed by using NTP) it's only about
time not space. However, an icecast relay will mean extra buffers
which may delay the problem.

Another possibility is the lack of catchup on libshout-based clients.
If there is a lag in the network (which may happen with low or high
bandwidth links) sending one packet will be delayed, so the next one
should be sent quickly to catchup. The libshout offers a default
synchronization mode that doesn't do that, which creates an
accumulation of small lags. This would look pretty much like a clock
drift, but probably less regular.

Hope this helps...
-- 
David

------------------------------------------------------------------------------
WhatsUp Gold - Download Free Network Management Software
The most intuitive, comprehensive, and cost-effective network 
management toolset available today.  Delivers lowest initial 
acquisition cost and overall TCO of any competing solution.
http://p.sf.net/sfu/whatsupgold-sd
_______________________________________________
Savonet-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/savonet-users

Reply via email to