hi John!

Le 5 avril 2012 00:39, John Paul Morrison <[email protected]> a écrit :
> Does mksafe() do its own buffering?
>
> I've got a simple script to read from a named pipe, and it doesn't seem to
> work without mksafe().
>
> Anyway, there seems to be about 20 seconds worth of buffering and I don't
> think it's coming from the writer of the named pipe.
> Is there some way to get around this? The writer on the pipe is controlled
> by an iPod, so it's annoying to wait 20 seconds after pressing pause, or
> skipping to a new track.
>
> Fiddling with buffer/max on input.external, and burst/buffer settings on
> output.harbor don't seem to improve things.
>
> Anything else I can try?
>
>
> def input.shairport(~id="input.shairport",
>         ~restart=true,~restart_on_error=false,
>         ~buffer=0.2,~max=10.,s) =
>  input.external(id=id,restart=restart,
>                 restart_on_error=restart_on_error,
>                 buffer=buffer,max=max,
>                 "dd if=#{quote(s)} 2>/dev/null")
> end
>
> s = mksafe(input.shairport("/tmp/shairtunes.pcm"))
> output.harbor(%mp3(bitrate=256),port=8002,fallible=false,mount="Shairport",s)

Hmm that's a weird situation.. mksafe should not do any buffering..
Its definition is simply: fallback(track_sensitive=false, [s,
blank()]) ...

I'd be more suspicious about something being funky with the blank that
is added by mksafe in relation with output.harbor.. Blank is a very
bad thing with compressed data and output.harbor is still a pretty
experimental operator.. You may want to try with another output to
being with and see how it behaves.. For instance a output.icecast..

Romain

------------------------------------------------------------------------------
Better than sec? Nothing is better than sec when it comes to
monitoring Big Data applications. Try Boundary one-second 
resolution app monitoring today. Free.
http://p.sf.net/sfu/Boundary-dev2dev
_______________________________________________
Savonet-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/savonet-users

Reply via email to