Just a quick heads up - I'm probably going to move the primary rtc.io discussion channel to the organisation channel now that gitter supports such things: https://gitter.im/rtc-io
Hey folks - if people could vote this firefox issue up that would be great. It's pretty bad that they don't support the behaviour outlined in the issue report: https://bugzilla.mozilla.org/show_bug.cgi?id=852665
@kumavis just posted this over on the rtc repository - definitely a worthwhile read: https://www.pkcsecurity.com/blog#webrtc-flow
For anyone trying to use WebRTC where you are trying to maintain a sizeable mesh of connections, it's probably worth starring this Chrome issue: https://code.google.com/p/chromium/issues/detail?id=421253
oops
ANNOUNCE: Thanks to the efforts of @Neograph734 we now have support for signalling over socket.io: https://github.com/rtc-io/rtc-signaller-socket.io
ANNOUNCE: New high level package for doing screensharing in a way that works consistently across webapps, atom-apps, etc: https://github.com/rtc-io/rtc-screen (thanks to @maxogden) for the push
Firefox screensharing support is still to be done but I should be able to tackle that early next week
Here a suggestion is made : rtc-io/rtc-mesh#8 but I'm not sure how to use the library
Is there an implication to browserify
I must have missed
I'm trying to code a prototype of a mobile device streaming live video to a server and the server echoing back a processed version of the stream to the mobile client. We see how to handle WebRTC on the mobile device, but have not been able to find any examples of server-side code that handles WebRTC and let's me access the frames for processing (without running a browser). Could someone please point me in the right direction or give any suggestions?
@hamsal apologies for not getting back to you on this - I think what you are talking about would be probably within the domain of janus (https://github.com/meetecho/janus-gateway)
I think they've already implemented an echo functionality, and may have processing capabilities also
you won't be in the realm of JS though
this might be worth a look to: https://github.com/js-platform/node-webrtc (a great project, but AFAIK still doesn't have support for streams)
how are things going?
im looking to setup a hub-spoke connection graph
i want to use webrtc b/c it exposes reliable + unreliable data channels
how can i most elegantly setup a connection with only the hub and not the other spokes?
let me have a think about it - primarily I would think the way to do it would be to include a qualifier in the announce message of the hub which the spokes would look for before attempting to set up a new peer connection
you have the ability to filter out announce messages based on criteria by overriding the
allow attribute of the peer:filter event (documented here: https://github.com/rtc-io/rtc-signaller#events-regarding-peer-state)
I think that should work...
and then save the stream that gets sent to it to a file?
Basically the flow would be:
- Get a MediaStream object from the browser.
- Create a client and add that object.
- Create a client on the server that the client from the browser connects to.
Trying to get around the fact that Chrome hasn't implemented the MediaRecorder API
@vhmth it's definitely possible using something like Janus (https://github.com/meetecho/janus-gateway) but I'm not aware of any pure JS clients running on the server side yet. Some folks have run headless chrome for doing this kind of thing on a server, but given that you are wanting to save media I think Janus is probably your best bet...
@DamonOehlman thanks man - do you remember how they got the media stream by running a headless browser? Just curious. I'm actually setting up my backend stack with Docker right now and am going to run Kurento Media Server to record the media stream to a file and then send that file to S3 from my node app server.
I had issues that I thought stemmed from NAT and port exposure issues with Docker, so I abandoned that.
But I ended up having to set up secure websockets on the Kurento server to get it to work within the VM, so perhaps I'll revisit to see if I can get Docker working with my newfound knowledge
Really don't want to be developing and running my stack off a VM vs. my host OS (Mac OSX).
Is it possible to use rtc.io with https://github.com/webcast/webcaster, and stream webrtc over websockets?
