- From: Varun Singh <varun@callstats.io>
- Date: Fri, 16 Aug 2019 14:04:59 +0000
- To: "henbos" <sysbot+gh@w3.org>
- Cc: public-webrtc@w3.org
- Message-ID: <jze6nihy.28eb028d-4808-4cab-8158-a7e2e7472700@we.are.superhuman.com>
The suggestion to move the implementation name to the inbound and outbound local stream instead of attaching RTCCodec to the stream. Founder, CEO, callstats.io ( http://callstats.io ) http://www.callstats.io Interested in networking, media quality, and diagnostics. We are hiring!: www.callstats.io/jobs/ ( http://www.callstats.io/jobs/ ) On Fri, Aug 16 2019 at 2:43 PM, Harald Alvestrand via GitHub < sysbot+gh@w3.org > wrote: > > > > henbos has just submitted a new pull request for > https://github.com/w3c/webrtc-stats: > > > > == Move RTCCodecStats.implementation to > RTC[In/Out]boundRtpStreamStats.[decoder/encoder]Implementation == Fixes > #445 > > > > Which encoder or decoder is used is a property of the RTP stream, not a > property of the codec. For example, it is possible that one stream uses an > H264 hardware decoder but another stream, having experienced decoding > errors and fallen back to software, uses an H264 software decoder. Both > RTCInboundRtpStream objects would reference the same RTCCodecStats. > > > > This PR allows Chrome to ship this in M78: > https://webrtc-review.googlesource.com/c/src/+/149168 > > > > See https://github.com/w3c/webrtc-stats/pull/473 > > >
Received on Friday, 16 August 2019 14:05:25 UTC