Comments (7)
Is there guidance on what low and high mean here?
from webrtc-stats.
It's well defined for VP8, H.264, VP9 and HEVC, I think. Not redefining it would seem to be a Good Thing.
(low means not very compressed, high means very compressed, if I remember right).
from webrtc-stats.
Yes, the CL states the value is between 1-127. which would important to add. I suppose there is sufficient information to guess what low and high means. A value in the 20s is better than in the 50s.
from webrtc-stats.
To make use of this, we need to know the video codec in use, reset it on codec change, and also ensure this is per-track since each track will have different QP (or even different codecs)..
I think it is a reasonable thing to feed back to the application.
from webrtc-stats.
Hm. I'm trying to work out how this hangs together.
Frames are counted in RTCMediaStreamTrackStats, and you have to divide the sum-of-QP by the number of frames to get the average QP value.
But the codec information (necessary to know what the QP value means) hangs off RTCCodecStats, which is linked to the codecId of the RTCRtpStreamStats. The linkage between the two is taht RTCMediaStreamTrackStats has a sequence of SSRCs that can be used to find the RTCRtpStreamStats involved.
Should we put the sum of QP values on the RTCRtpStreamStats, since that has a single codec? Should we then put the FramesReceived there too, so that if a MediaStreamTrack switches codecs mid-flight, the QP values can be counted sensibly for each codec?
from webrtc-stats.
Frames that are not decoded but received do not have a QP value, ergo, FramesDecoded would be more appropriate than Frames Received.
Frames may be received by the endpoint and not decoded, may occur when a new user joins and starts receiving an existing stream or a user that was not in the filmstrip starts speaking. The decoding depends if the first frame is a full reference or not.
I suppose this would not be an issue if a stream is paused and resumed by the sender?
from webrtc-stats.
Summarizing status:
- The basic approach makes sense. The question is where to place the counters.
- To interpret the number, we need to know the codec. That's only visible in RTCRtpStreamStats (via "codecId" that references an RTCCodecStats object).
- To get the average QP on reception, we need to know the number of decoded frames. At the moment, that's an RTCMediaStreamTrackStats value.
Suggestion:
- Add the proposed counters to RTCRtpStreamStats
- Add a FramesDecoded value to RTCRtpInboundStreamStats
- Add a FramesEncoded value to RTCRtpOutboundStreamStats
I think that should make the counters usable without too much tree-walking.
As long as RTCRtpStream remains 1:1 with RTCMediaStreamTrack, the counters should have the same value as those other counters, but that won't necessarily be true always - the big exception is simulcast, where there would be multiple RtpStreams all pointing to the same RTCMediaStreamTrack.
from webrtc-stats.
Related Issues (20)
- Conformance criteria incompatible with standards-track HOT 7
- RTX and FEC stats are incomplete. HOT 12
- Should we flatten the hierarchy? HOT 5
- Change playoutdelay to jitterbuffertarget HOT 2
- Add clarification that inbound-rtp.bytesReceived includes retransmissions. HOT 6
- RTCAudioSourceStats.audioLevel - what algorithm described in totalAudioEnergy? HOT 3
- Delete the obsolete section, "track" and "stream" stats - they don't exist anymore HOT 3
- Improve test coverage HOT 1
- add framesInput stats to rtcoutboundrtpstreamstats HOT 6
- Stats example uses confusing variable names
- Exposing audio interruption metrics to JavaScript
- Define a mechanism for setting the trigger duration for a video freeze
- Sender-side packetsReceived can be negative from garbage RRs HOT 1
- Change "Does not" to "must not"
- Make "not present" reference "map/exist"
- Is codec.sdpFmtpLine present when there is no fmtp line in the SDP? HOT 2
- Lifetime of outbound-rtp should start BEFORE first packet is sent HOT 3
- Should RTCRtptransceiver.stop() cause inbound-rtp stats to disappear?
- RTCCodecStats.clockRate - media sampling rate or the codec clock rate?
- RTCStats.timestamp - fingerprinting and since epoch
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from webrtc-stats.