Comment #3 on issue 2070 by baranowb: Support for stream reassembly after silence period
http://code.google.com/p/mobicents/issues/detail?id=2070
seeeeeeee comment above. When there is silence, buffer will contain single packet. Now it has stamp=X, seq=N. Now after some time - like T, we get P+1, which has stamp=X+T,seq=N+1. Now what will happen? We do evolve on stream, impl gets data from buffer. In buffer we will set duration of packet with seq=N to T. What happens next? We stream that old packet(which essentially has no meaning to UA, since its loooong late) and schedule next read to happen after T.
So what happens from UA POV, we do silence, than on activity, we get long late packet, than we have to wait period equal to silence to get voice. Thats bad.
