I love to work remotely. My dream workplace would probably be some hundred kilometers north of Oslo in a house outside the village overlooking the fjord.

But a very big part of my job is to communicate with people. And here’s the thing: I would rather do that in person. Of course I have tried all the different video chat solutions, I have invested in a good microphone and good headphones and a great internet connection. But it just isn’t the same.

When I am talking to somebody via a video call, I can never quite shake off the feeling that something is off. The in-person conversion flows, and the video call just… doesn’t.

The reason for this is latency.

When voice is transmitted over the internet, it needs some time to arrive at its destination. This delay is called latency. In an in-person conversation, the latency is less than 10 milliseconds, governed by the speed of sound. In a video call, that latency is often 200 to 500 milliseconds. That is a huge difference. It means that from the moment you say someting, it takes between 400 and 1000 milliseconds for a reaction to arrive.

As humans, we are used to conversations in a low-latency environment. We expect an immediate reaction when we say something. If the reaction comes with a delay, we subconsciously conclude that the person we are talking to is a little bit stupid. The latency also makes it hard to throw in a well-timed “hmm” or an interjection - crucial for a constructive exchange.

Therefore, the video call conversation does not flow.

But what causes this latency?

A modern fiber connection transmits signals with roughly two third of the speed of light, about 200 kilometers per millisecond. But in most video calls, the conversation partners are not more than a few thousand kilometers apart from each other. Even accounting for the detour to a relay server 2000 kilometers away, this does not explain more than a few milliseconds of latency.

This means that almost all latency must be caused by delays in signal processing, both on the high level (video encoding), in the operating system and on the TCP/IP package switching level.

To me, this is good news. Because it means that the problem can be fixed. And fixing it will do to video calling what the iPhone did to the smartphone: It will still be video calling, but it will suddenly feel RIGHT. The implications of this could hardly be overstated.

Subscribe via RSS