News for hackers. Stuff that matters.
Re:WebRTC-HTTP ingestion protocol (WHIP) by marksomnian on Tuesday February 15, 2022 @06:29PM (#30354392)
Very interested to see if there is development in the WebRTC streaming ingest world. As far as I remember, the last major implementation was Mixer’s FTL (which has an open-source SDK implemented in OBS, but the server side was proprietary - though it has been reverse-engineered).

Also interested in comparisons between WebRTC and SRT, in particular for low-latency high-quality streaming (the niche between traditional streaming like RTMP and more conventional WebRTC apps where latency trumps quality).

Re:WebRTC-HTTP ingestion protocol (WHIP) by xyk2 on Wednesday February 16, 2022 @05:01AM (#30358103)
SRT serve well as a first-mile contribution protocol, so things like delivering live video from remote field cameras back to HQ or enabling reliable cellular broadcasting over a congested 4G network. The error-correcting ability combined with a configurable fixed latency makes SRT very attractive for this type of production, and has been slowly replacing RTMP/MPEG-TS workflows in the broadcast world. I've seen networks where a RTMP stream did not work at all and switching to SRT worked straight away.

WebRTC seems like the future for last-mile delivery to viewers, though the popularity of LL-HLS and LL-DASH (and the ease of deploying them on existing infrastructure) is confining WebRTC to video-calling and some smaller applications where every bit of latency matters.

Re:WebRTC-HTTP ingestion protocol (WHIP) by pabs3 on Tuesday February 15, 2022 @07:17PM (#30354795)