[Live-devel] question about ServerMediaSubsession

[Ross Finlayson]

>     1. In the test programmes, I wonder why they use 
>PassiveServerMediaSubsession to create a new subsession and to begin 
>streaming through 
>"sessionState.sink->startPlaying(*sessionState.source, afterPlaying, 
>NULL)"? And I found that inside of PassiveServerMediaSubssion we did 
>nothing
>about streaming.(Maybe I missed)
>         In my opinion, RTSP is to control RTP(my 
>understanding), so the situation should be like this: 
>method startPlaying should be inside of ServerMediaSubssion so the 
>streaming can be automatical. It seems that 
>OnDemandServerMediaSubssion is doing so but I still don't quite 
>understand.

"PassiveServerMediaSubsession" is used to implement a RTSP server for 
a stream (usually multicast) that already exists before the client 
accesses it (using RTSP).  Therefore, for such a stream, we call 
"startPlaying" once, before (and independent of) each RTSP request.

"OnDemandServerMediaSubsession", however, is used to implement a RTSP 
server that creates and serves new streams, on demand.  I.e., a new 
stream is created for each RTSP client request.  Therefore, for such 
a stream we call "startPlaying" for each RTSP request.  This is done 
in the "OnDemandServerMediaSubsession" code.

>     2. When I test testMP3Steamer, the unicast address is pretty 
>good, however, when the address is multicast address such as 
>232.255.42.42, although I add USE_SSM inside of Makefile, client 
>using VLC within windows does not work while within linux works.
>         Why? I am puzzled.

So am I.  I suggest using the "testOnDemandRTSPServer" demo 
application (a unicast RTSP server) instead.
-- 

Ross Finlayson
Live Networks, Inc.
http://www.live555.com/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.live555.com/pipermail/live-devel/attachments/20070211/513c1412/attachment.html