 | 1 | initial version |
SIP endpoints (User-Agents) need a constant stream of RTP to play the audio. Usually RTP packets are send each 20ms (as is the case in your screenshot). Then there is a small buffer (jitter-buffer) to be able to handle little differences in the timings. But after a while, the data is of no use anymore, just because it is too late to play it back after like 100ms.
So, in your output, RTP packets are queued up for 1020 ms (1040ms - 20ms) and then come all at once, so no packetloss, but the packets come too late to be played back. Some devices will just play the audio packets in a hurry, but most devices will just drop the packets that arrive too late.
The way to solve this, is to configure QoS in the network to make sure the RTP packets get sent with priority so they will arrive in time.
