Some so called "security" browser extensions override WebRTC APIs so
even if the UA string suggests the browser is supported, it won't work.

Coincidentally, this should also mark Safari in Lockdown Mode as
unsupported, since the WebRTC APIs are not available in that case.

Calling this API on Firefox is causing freezes when the local endpoint is the answerer. https://bugzilla.mozilla.org/show_bug.cgi?id=1917800

* feat(codec-selection): Use the new codec selection API
https://github.com/w3ctag/design-reviews/issues/836. This allows the client to seamlessly switch between the codecs without having to trigger a renegotiation.
This feature is behind the flag testing.enableCodecSelectionAPI in config.js

* fix(stats): Fix local resolution stats.
The video codec for the local video sources needs to identified differently now, from the codecs field in the RTCRtpSendParameters returned by the browser. We no longer munge the remote SDP to switch to a different codec.

* feat(stats): Feed encodeTime stats for all local SSRCs to the codec selection mechanism.

* fix(codec-selection) Continue to mumge SDP for selecting H.264.

* feat(codec-selection) Make screenshare codec configurable.
If no 'screenshareCodec' is set under videoQuality or p2p settings, AV1 will be selected as default.

* squash: Address review comments

* Update modules/RTC/CodecSelection.js

Co-authored-by: Saúl Ibarra Corretgé <s@saghul.net>

* fix(codec-selection) Add codec to existing stats

---------

Co-authored-by: Saúl Ibarra Corretgé <s@saghul.net>

Set the min. required version for Firefox to 91 so that the last 3 ESR versions are supported.

* fix(RTC): Remove support for plan-b SDP.
Since unified plan mode becomes the default mode, so does multi-stream mode.

* squash: Remove plan-b unit tests and fix comments

* squash: address review comments

Add SVC and Simulcast support for AV1 and VP9.

The default mode is the full SVC mode for VP9. It works as expected with the current version of JVB.
AV1 in the client is only supported when the bridge negotiates the Dependency Descriptor extension headers which are needed for reading the layer information for AV1 and H.264 video frames.

* feat: Enable H.264 simulcast support.
Multi-encoding simulcast for H.264 is supported now because of the DD header ext support added on the bridge side.

* fix(video-quality): Add default bitrates for all codecs.
Expect the videoQuality settings in the new format, deprecated configs are still supported.
videoQuality: {
    AV1: {
        maxBitratesVideo: {
            low: 100000,
            standard: 300000,
            high: 1000000,
            ssHigh: 1500000
        },
        useScalabilityModeAPI: {
            enabled: true,
            useSimulcast: true, //(defaults to SVC),
            useKSVC: true //(defaults to L3T3_KEY)
        },
    },
    H264: {
        maxBitratesVideo: {
            low: 200000,
            standard: 500000,
            high: 1500000,
            ssHigh: 2500000
        }
    },
    VP8: {
        maxBitratesVideo: {
            low: 200000,
            standard: 500000,
            high: 1500000,
            ssHigh: 2500000
        }
    },
    VP9: {
        maxBitratesVideo: {
            low: 100000,
            standard: 300000,
            high: 1200000,
            ssHigh: 2500000
        },
        useScalabilityModeAPI: {
            enabled: true,
            useSimulcast: true, //(defaults to SVC),
            useKSVC: true //(defaults to L3T3_KEY)
        },
    }
}

* Disable encodings when requested height for screenshare is 0.

* feat: add unit tests for scalability modes.

* ref: Move all encoding configuration calculation to TPCUtils.

* Add more unit tests for VP9 K-SVC mode.

* Make codec name in settings case insensitive.
Also read the deprecated max bitrates correctly, add unit tests to test it.

* ref(TPC): isSimulcastOn -> isSpatialScalabilityOn.
It makes more sense to call it spatial scalability than simulcast now that full SVC support is available.

* fix: Negotiate AV1 DD header exts only for AV1 and H.264.

* fix: Rename setting useL3T3Mode->useKSVC and adjust the defaults.

* fix(codec-selection) Use a configurable preferred codecs list to select the codecs.
Allow asymmetric codecs to be configured on the endpoints. This means that Firefox and Safari which have bugs with VP9 encode will now encode VP8 but decode VP9 coming in from Chromium endpoints.

* fix(codec-selection) Add unit tests for the codec selection logic.

* feat(codec-selection): Introduce mobileCodecPreferenceOrder setting in config.js.

* fix(codec-selection) Avoid unnecessary renegotiations.
On participant join/leave, check if the new intersection of codecs are already configured to be the top n codecs.

* Address review comments

* fix: Strip the codecs that are not in the codec list from the video m-lines.
Also, ignore remote codecs published in presence for p2p connections.

* fix: Define default codec order for mobile and desktop.
Fix an issue where p2p between mobile and desktop was broken.

* fix: Add default codecs for both p2p and jvb on mobile devices.
Munge the initial offer sent out by RN clients since RN doesn't support RTCRtpTransceiver#setCodecPreferences.

The constructor is too early and we miss it in rtcstats.

* fix(stats) Obtain resolution/fps from 'outbound-rtp' stats.
'Track' based stats were dropped in Chrome 112, therefore send resolution and fps for the simulcast case needs to be calculated based on the 'outbound-rtp' streams that are currently active.

* squash: remove an unwanted log

* Squash: Address review comments.
Use outbound-rtp stats for both Firefox and Safari.
Firefox - Ignore active field if not present in the stats and calc fps using 'framesSent'.

Track based stats have been removed recently and RTCRtpReceiver:getSynchronizationSources has been available in the browsers for a while now.

It does no harm on RN, since it doesn't detect it as a supported browser
and bails out.

RTX in combination with simulcast on Firefox results in bad video quality
therefore we only enable it for version >= 96.

We do not want Firefox sending video when its video muted. https://bugzilla.mozilla.org/show_bug.cgi?id=1735951

* feat(CodecSelection) Disable VP9 on Safari.
Safari does not decode VP9 properly- https://bugs.webkit.org/show_bug.cgi?id=231071 and https://bugs.webkit.org/show_bug.cgi?id=231074.

* squash move VP9 support check to browser capabilities.

There was an error in the pattern match, everything was included in it. In
addition return a boolean in case we succeed in the UA pattern match, since it's
more correct than returning an array with the match.

On mobile Safari, if a user joins audio and video muted, the browser doesn't decode the incoming audio. Workaround is to always add the audio track to pc and mute it if needed.

An alternative to Insertable Streams, currently implemented in Safarii / WebKit.

https://w3c.github.io/webrtc-encoded-transform/

Fixes: https://github.com/jitsi/jitsi-meet/issues/9585

remove legacy APIs which are not even supported anymore

Munge only the m-line that corresponds to the source that the browser will be sending.
Do not select VP9 on Firefox.
Detect support for RTCRtpTransceiver#setCodecPreferences correctly.

This causes the audio to not work.

Call RTCRtpTransceiver#setCodecPreferences with the preferrred codec order before every createOffer/createAnswer. This ensures that the codec preference is enforced even when there is no local description available yet while the preferred codec is being set immediately after media session creation.

This can be controlled through the config.js  option 'enableUnifiedOnChrome'.

Make sure the codec order is munged on all the m-lines for unified plan clients. Implement the logic for setting the preferences through RTCRtpTransceiver#setCodecPreferences.

Capture the audio levels only for the top 5 speakers as RTCRtpReceiver#getSynchronizationSources can be expensive when we have too many audio receivers in the call.
Also, capture the audio levels for track that are unmuted if RTCRtpReceiver#getSynchronizationSources is not supported.
Switch Safari to using getStats since its reporting errorneous values, i.e., 0.000001 as audio level for all remote audio tracks.

Apple added getUserMedia support for WkWebview based browsers like chrome and Firefox on iOS 14.3. These browsers behave as Safari does on iOS. Therefore, extend the Safari checks to these webkit based browsers as well.

All browsers on iOS use WebKit engine so they use the Unified Plan.

It has been `true` for a very long time.

Cleanup some unused code after the bump.

Since we use transferable streams to transfer the streams to the worker
we need to feature detect them as Chrome 86 will launch Insertable Streams
by default without shipping transferable streams yet (status still unclear,
it might still ship too)

See here for the feature detection used:
  https://groups.google.com/a/chromium.org/g/blink-dev/c/1LStSgBt6AM/m/hj0odB8pCAAJ

Make sdp munging logic for changing codec preferences more generic
Do not negotiate High profile H264 codecs on when RN clients are in the call

Add a performance stat around long tasks. Chrome supports PerformanceObserver API that lets us
register for long tasks event. Any task that takes longer than 50ms is considered a long task.

* feat: use getSynchronizationSources on the receiver for remote audio levels
Use getSynchronizationSources if it is supported, fallback to using getStats otherwise.
* feat/ref: Use the local audio levels from LocalStatsCollector
When using getSynchronizationSources, use the audio levels from LocalStatsCollector for NoAudioSignalDetection.js
Remove obsolete code - TalkMutedDetection feature using audio levels is not used anymore