Does JVB send video to headless clients?

Hi,

I have a docker container that runs a chrome puppeteer instance and joins a jitsi call, then streams audio and video with the --use-fake-device-for-media-stream chrome arg. This works great, I can scale up N clients to join a call and they all stream video and audio as if real users, as I want.

However, I want to be sure the puppeteer instances are acting as if they were real clients.
My dockerised client only receives and sends about 1/4 the data of human clients, which I would not expect to be the case.
Is there an internal measure determining if a client is headless and does not need to receive video?

Thanks!

The fake video does not produce real traffic.
You can pass use-file-for-fake-video-capture= and upload a file there which will be fed as video input.
We are using this file https://media.xiph.org/video/derf/y4m/FourPeople_1280x720_60.y4m but encoded with 30 frames per second.

1 Like

Yep, I’m also using a use-file-for-fake-video-capture= y4m file to get a video feed. :slight_smile:

But the bot clients playing these videos have greater upstream traffic than downstream, where the inverse is true for human clients, and I’m not sure why this is the case.

Cause they look at tile views, this is optimization where clients request just what is needed from the bridge. If you click on tilview to turn it off on clients the download and upload should be comparable as sending HD and receiving HD, and you want to give it a little time 15 seconds or so to rampup.