Thanks. Positional audio teleconf in the browser was something I hacked on back 5 years ago when WebVR first came out. Unfortunately, Google broke positional audio, then WebVR, then changed a lot of how WebAudio worked.
Lately, I'm involved with a few meetups that miss the dynamic of a real meetup. So I threw this together to hopefully get some semblance of the meetup space again.
It's really simple. Jitsi Meet has a web app API that basically loads a standard Jitsi Meet install in an iframe, with some inter-window communication setup using Window.postMessage. I add a small hack on Jitsi Meet's default install to take my own postMessage calls to sniff out the HTML audio elements Jitsi Meet creates and modifies their volume.
For the position of participants, I use Jitsi Meets inter-user text messaging system to send data packages. Then I just scale each user's audio volume based on a linear scale. Anyone closer than 3 squares is rendered in full volume, and anyone further than 14 is muted. And of course, settings in the scripts to scale these values. They could even be per-user, if they wanted to scale how well they can hear.
There's no stereo pan or ambisonics, but hopefully the volume level scaling is enough to make for useful meetups.
Lately, I'm involved with a few meetups that miss the dynamic of a real meetup. So I threw this together to hopefully get some semblance of the meetup space again.
It's really simple. Jitsi Meet has a web app API that basically loads a standard Jitsi Meet install in an iframe, with some inter-window communication setup using Window.postMessage. I add a small hack on Jitsi Meet's default install to take my own postMessage calls to sniff out the HTML audio elements Jitsi Meet creates and modifies their volume.
For the position of participants, I use Jitsi Meets inter-user text messaging system to send data packages. Then I just scale each user's audio volume based on a linear scale. Anyone closer than 3 squares is rendered in full volume, and anyone further than 14 is muted. And of course, settings in the scripts to scale these values. They could even be per-user, if they wanted to scale how well they can hear.
There's no stereo pan or ambisonics, but hopefully the volume level scaling is enough to make for useful meetups.