All right, after the first 3 parts, let’s get out into the park and have a look at how packet jitter looks like during a voice call when there’s pretty much nothing between my mobile device and the radio tower(s). As before, I used tcpdump / Wireshark and a little program of my own to capture and visualize the inter-packet time distances. For this test I modified my test setup a bit, as I had both devices for the call with me. Hence I couldn’t trace on my Wi-Fi router at home. Instead I traced right in the middle, i.e. on the TURN server in the network. That’s one of the nice benefits of running my own services. That being said, the graph above shows the inter-packet distance during a 3 minute call.
The graph looks very smooth, most packets arrive well below 30 milliseconds from each other and there are only a few that arrive with a timing of 40 milliseconds in between. In other words: During that three minute call, no handover was made, as that would have delayed at least one packet by around 80 ms. See my previous post on the topic for the details on this. I have to say I was very positively surprised to see such a clean result!
Next I had a look at a somewhat more difficult terrain: I left the park and walked through the nearby streets for 7 minutes and went around a number of corners to make sure I had quickly changing RF conditions and some handovers during the call. The second graph in this post shows how this looked like. Don’t be deceived by the initial clean look. It only looks clean because the y-axis has changed. During the call, I had 3 packets which only arrived after the previous packet after more than 400 milliseconds! That’s quite a bit away from the smooth 20 ms inter-distance arrival when the channel is clear. I’m not sure what caused those 3 packets to be delayed for such a long time. I don’t think the jitter buffer could compensate for such sporadic delays, so those packets certainly had an impact on speech quality.
There are also a number of packets that arrived around 80 to 120 ms after the previous packets and quite a number of them were probably triggered by network handovers. See my previous post for the details why this delay is typical for that.
The next question: Does one hear such occasional delays and if so, how? I can’t answer that in this post, because I had both devices I used for the call with me, but I didn’t listen to the audio channel closely. So I have to defer the answer to this question to one the next posts.