Will client-heavy avatar-based teleconferencing be a thing soon?

twitter logo github logo ・1 min read

Teleconferencing is notoriously iffy. We're pretty happy with Zoom, but even the slightest connectivity issues tend to go make things go south quickly.

Transmitting high quality live video strikes me as the bottleneck we could do without, in favor of sending minimal data in order to recreate my presence on the other end.

Our avatars could be sent once upon connection (and cached for future convos), and from there we only send audio. And even that could hypothetically be sent in a really lossy way and likely pieced together with advancements in AI. (Big risks in not getting this right)

I've seen concepts of stuff like this, but I'm wondering what people think the timeline is, and what the quality and reliability will be like.

This conversation tends to relate to VR, but it doesn't seem entirely necessary that they are intertwined.

twitter logo DISCUSS (12)
markdown guide

I had a meeting on hubs.mozilla.com which is a WebVR meeting room and it was pretty cool/fast but we kept getting distracted by throwing VR objects around. It's open source though so you could make your own version and disable those features.


That is very neat! Definitely an experiment worth keeping an eye on in the next few years. As noted in another comment, the lack of expressive facial mapping will be a problem for making this a truly comparable experience for basic communication the way video call provides, but very cool!

I really like to see this coming from Mozilla. I definitely support their efforts over, say, Facebook. I still fear Facebook as the undeniable winner eventually with all their resources.


I just tested the hub and its really interesting experience even without VR equipment it is a good move toward quality meetings with far less bandwidth usage. :)


in order for this to be more effective than just an audio call, you need to convey a lot of expressivity. Animoji-style mapping faces to avatars could be powerful, and so could VR-style head and hand posing.

What's tricky is those two methods are (using today's technology!) mutually exclusive — there's no way good way to transmit facial expression when you're wearing a VR headset, and you'll also feel rather silly waving around hand controllers if you're not wearing a HMD.


Animoji-style mapping faces to avatars could be powerful, and so could VR-style head and hand posing.

Yeah, I think that is definitely the key, and if it's not basically perfect you're in the uncanny valley of never good enough.


What would really change the game is spatial audio to be able to distinguish voices.

I'm thinking avatar-based conferencing in a VR environment.

Also the bandwidth I have to all the locations I use is ridiculously high, at this point the fact that it never works properly is a true mystery to me.


Never been a fan of video conferencing (or anything that wants a static- or video-camera version of my visage). When I'm in a camera-enabled conference room, I always try to find the cameras' blind-spots. When I'm working remote, I always disable my camera in favor of a static image (the same one I use for Dev.To, actually).


Sounds pretty interesting, I haven't seen much of this yet. My question would be how it would affect our perceptions knowing that who we're watching is simply a recreation of their likeness, assuming the recreation is intended to be photorealistic. I suppose most are talking about video game-like avatars, but that would be another interesting point.

As an aside, I've been wary of Zoom ever since this huge bug last year.


High quality video encoding and transmission is definitely already quite doable, depending on CPU power and network bandwidth and latency. Working on the team that creates the Web Client for GoToMeeting, I should probably know. Also, as new and improved video codecs arrive and CPU power is constantly increasing, it'd really need an anstonishing breakthrough to make VR meetings feasible before media quality improves so much that you would still wish for them to be possible.


I would totally be on board with avatar styled persona's for video conferencing. The average group is probably not using video all the time anyways, and a personalized avatar would add a much needed relaxation of the seriousness of conference calls. Think old xbox live avatar style vs a static image.

I love the idea.

Generic avatar guy approves


My guess is probably never. I see this as a novelty feature whereas you can just turn off your video and communication works fine for remote teams audio-only.


Not sure I need the avatar. The screen sharing is the bigger issue IMHO, which is worse than the video sometimes.

This is a really hard space.

Classic DEV Post from Aug 27 '19

How do *you* pronounce sudo?

Ben Halpern profile image
A Canadian software developer who thinks he’s funny. He/Him.

dev.to is where software developers stay in the loop and avoid career stagnation.

Sign up (for free)