Web GUI

[mvcisback]

No description provided.

[mvcisback]

So, I suspect it may be easiest to do a lot of the classification server side (although it'd be nice to do it client side, it'd mean rewriting code in JS....ew).

For that it looks like this is a thing:
https://github.com/meetecho/janus-gateway

Or in other words:

[ffaghri1]

Good point.
I certainly vote for the server side too, easier to scale in future and
have no interest touching JS!

On Mon, Nov 10, 2014 at 2:01 PM, Marcell Vazquez-Chanlatte <
[email protected]> wrote:

So, I suspect it may be easiest to do a lot of the classification server
side (although it'd be nice to do it client side, it'd mean rewriting code
in JS....ew).

For that it looks like this is a thing:
https://github.com/meetecho/janus-gateway

Or in other words:

https://camo.githubusercontent.com/e57adcf62eeda2df8a16d35daee52471d231151c/687474703a2f2f6d65646961312e67697068792e636f6d2f6d656469612f7375796d7a58794462754550432f67697068792e676966

Reply to this email directly or view it on GitHub
#22 (comment).

[mvcisback]

@ffaghri1 yeah, I was also thinking a good start would be an echo client that takes what you say for 3 sec, repeats the video unaltered, then returns a video showing where the activated face is with altered audio. That way its easy to tell the difference and you only need 1 person to demo

[mvcisback]

I've looked into how to get the video from webrtc. Looks like modifying these plugins might be the best way:

• https://github.com/meetecho/janus-gateway/blob/master/plugins/janus_streaming.c
• https://github.com/meetecho/janus-gateway/blob/master/plugins/janus_echotest.c

One shows how to hook into gstreamer, which might serve as a good way to pass around the video data.
The other shows how to just echo back the data unchanged.