Closed fippo closed 3 years ago
(note: oddly the firefox issue only shows up when increasing the timeout to 1000ms)
@na-g any idea what could cause twice the NACKs reported in FF stats?
I'm starting to wonder if this is related to us passing RTCP to all receivers.
Not sure I like the current "protocol" of switching the encoding from json to binary. This won't work with datachannel things I think. Sending packets as JSON with type "rtp", "rtcp", "data" might be the way forward..
You don't have to set the encoding
attribute by the way:
not perfectly happy with the test -- currently it relies on payload type asymmetry, i.e. if the answer says "vp8 100 rtx 101" the browser will send on those payload types.
Having a way (endpoint?) to do setRemoteDescription(type=offer) first would be good. Merging anyway.
Not sure I like the current "protocol" of switching the encoding from json to binary. This won't work with datachannel things I think. Sending packets as JSON with type "rtp", "rtcp", "data" might be the way forward..
@nils-ohlmeier: I get a nackCount of 2 in Firefox which increases to six if I send the NACK three times. Looks like you have a stats bug? ;-)