On Sat, 2012-10-27 at 01:46 +0000, Sun, Xiaodong wrote: > Tanu, > > Thanks for your detailed explanation of sink-input. I am clear of it > now. > In term of the solution, what else would you like to suggest me to try > from system point of view? It seems only changing module-combine-sink > module is not a feasible solution. Why not? The blocker problem seems to be that module-tunnel reports the latency wrong, and I suggested a fix for that: avoid the pa_sink_render() call in module-tunnel.c until the latency information is available. pop_cb() in module-combine-sink.c is called by pa_sink_render(), so if pa_sink_render() is called only when the latency information is correct, then the latency information will always be correct also in pop_cb(). -- Tanu
