Wonderful. Plus being able to directly stream to peercast is the ice on the
cake :-)

Thanks a lot for the information, and thanks in advance for the planned
improvements!

Xavier.

On 12/18/05, David Baelde <[EMAIL PROTECTED]> wrote:
>
>
> > But I'd also like to intervene anytime through the sound
> > card and/or the network for a live session, whithout having to manually
> > fade and mix. I just want the live source to fully take
> > the lead whenever it comes up, and give it back whenever it shuts down.
>
> As far as I understand, what you want is possible with a pure savonet
> system. We have an http source which does exactly what you said: "I just
> want the live source to fully take the lead whenever it comes up, and
> give it back whenever it shuts down.". The http source listens for an
> http ogg/vorbis stream.
>
> Here is a piece of the configuration used for the next version of
> RadioPi (http://delroy.via.ecp.fr:8080/reggae.ogg):
>
> ========
> # One output for icecast, one directly sent to peercast
> let out = output.icecast ~host:"localhost" ~port:8080 ~password:"XXX"
> let p2p = output.icecast ~host:"localhost" ~port:7144 ~password:"XXX"
>
> # A file for playing during failures
> let interlude = one_file "/bla/some.mp3"
>
> # A normal playlist and a relay for live session
> let reggae = playlist "/data/reggae/playlist"
> let live = http "http://localhost:8080/live.ogg";
>
> let reggae = out ~mount:"reggae.ogg" (p2p ~mount:"reggae.ogg"
>    (hard_fallback [ live ; fallback [ reggae ; interlude ]]))
> ========
>
> However, the switch from the playlists to the live currently doesn't
> support any kind of transition. This is something we should do soon,
> since radiopi wants it too. Another thing you may not like is that the
> ogg/vorbis stream is decoded to WAV and then encoded back to ogg/vorbis,
> even the settings are exactly the same. This is not a fundamental
> limitation of the system, but it may not be improved so soon.
>
> I hope you'll have fun with liquidsoap!
> Feel free to ask anything here.
> --
> David
>

Reply via email to