amazon-chime-sdk-js - npm Package Versions

[3.9.0] - 2022-09-21

Added

• Add the audio output gain and frequency to the meeting readiness checker's configuration. The readiness checker uses this value to set the "Play Tone" gain and frequency.
• Add support for background filter starting from iOS 16 for major browsers Safari, Chrome, and Firefox (except on iPad).

Removed

Changed

• Add metric derived from string metric in metric report.
• Send audioInputFailed and videoInputFailed if there is error getting device labels.
• Updated DefaultVideoFrameProcessorPipeline to clone audio tracks to the output MediaStream instead of dropping them. This allows more straightforward usage of the pipeline on content sharing like file playback.

Fixed

• Fixed missing videos, or unnecessarily long freezes when switching simulcast streams.

[3.8.0] - 2022-08-18

Added

• Add encoder/decoder is in hardware metric and stream dimension to signaling protocol
• Report encode/decode time and if encoder/decoder is in hardware, and add encoder/decoder name as a metric dimension

Removed

Changed

• Fix a confusing function name from millisecondsPerSecond to averageTimeSpentPerSecondInMilliseconds.
• Move requiresPlaybackLatencyHintForAudioContext to ExtendedBrowserBehavior.

Fixed

• MessagingSession reconnect loop did not break on error past reconnect deadline. Infinite reconnect loop was caused due to firstConnectionAttemptTimestamp not being set as startedConnectionAttempt was not invoked. Check https://github.com/aws/amazon-chime-sdk-js/issues/2372 for details.
• MessagingSession getMessagingSessionEndpoint call is now backwards compatible with AWS JS SDK v2.
• Use a default "playback" latencyHint when creating the AudioContext on Windows. Also adds a setDefaultLatencyHint API to DefaultDeviceController to allow for overriding.
• Fix behavior of websocket disconnects before a session is connected. Session.start() promise shall fail in the scenario.
• Queue messages before a messaging session is established to avoid dropping them.

[3.7.0] - 2022-07-05

Added

• Add support in MessagingSession to allow websocket connection for messaging service to select sort by for Prefetch.
• Added two new meeting events: sendingAudioFailed to indicate a failure in sending audio packets out and sendingAudioRecovered to indicate recovery after a sendingAudioFailed event.
• Added a new ConnectionHealthPolicy named SendingAudioFailureConnectionHealthPolicy to evaluate and trigger the above events. Also, added sendingAudioFailureSamplesToConsider and sendingAudioFailureInitialWaitTimeMs to ConnectionHealthPolicyConfiguration fields for configuring the new health policy.
• Added a new field consecutiveStatsWithNoAudioPacketsSent in ConnectionHealthData for evaluating the SendingAudioFailureConnectionHealthPolicy.

Removed

Changed

• Update package.json to include Node 18.
• Refactored connectionHealthDidChange() in MonitorTask by creating a new applyHealthPolicy() method to reduce redundancy in the health policy evaluation.
• Modified SignalingAndMetricsConnectionMonitor to update consecutiveStatsWithNoAudioPacketsSent.
• Reset connectionHealthData on every connect (as opposed to doing it only when signalingClient is not initialized). This allows us to honor the cool-down time in both SendingAudioFailureConnectionHealthPolicy and UnusableAudioWarningConnectionHealthPolicy in case when the connection is failing due to an error like "no ice candidates were gathered".

Fixed

• Fix AbortError when turning video ON in Safari.
• MessagingSession reconnects with refreshed endpoint and credentials if needed. EndpointUrl on MessagingSessionConfiguration is deprecated as it is resolved by calling getMessagingSessionEndpoint internally.

[3.6.0] - 2022-06-23

Added

• Add a new API enableSimulcastForContentShare to enable simulcast for content share so that content share could be shown in network constrained clients. The lower quality layer has 300 kbps max bitrate, resolution scale factor of 2, and 5 max framerate.
• Add APIs setVideoCodecSendPreferences and setContentShareVideoCodecPreferences to allow configuration of codec being used to send. See the JS SDK guide for more details.
• Added tracer logs to missing subscribe and unsubscribe methods.
• Added opt-in server side network adaption enablement flag ServerSideNetworkAdaption.EnableBandwidthProbing. See this section in the guide for more details.

Removed

Changed

Fixed

• Fix issue closing a web worker in Amazon Voice Focus's inline worklet.

[3.5.0] - 2022-06-02

Added

• Add a workaround for https://bugs.webkit.org/show_bug.cgi?id=241152 to play a paused video element in Safari.

Removed

• Removed deprecated WebRTC constraints googCpuOveruseDetection and googCombinedAudioVideoBwe which were being removed or were already no-ops

Changed

• Bump protobufjs from 6.8.8 to 6.11.3.
• Update DOMBlobMock to accommodate @types/node changes.

Fixed

• Fix issue where Amazon Voice Focus stops working after changing device while muted.
• Remove the isChanged flag as it breaks some old muting funtionality causing Amazon Voice Focus to stop working.
• Fix issue where Amazon Voice Focus could stop working after changing the mute state of a null device or changing the device shortly after.
• Fix an issue for mute local when there is no audio input.
• Fix trucation of video subscriptions not occuring if the resubscribe was driven by MonitorTask.
• Fix protobuf generation script for upgrade.
• Optional chain signaling client observer removal to fix issue if audioVideo.stop() is called before audioVideo.start().

[3.4.0] - 2022-05-24

Added

• Add the reserved status code AudioDisconnectAudio.
• Add support in MessagingSession to allow websocket connection for messaging service to enable Prefetch feature.

Removed

Changed

• Add reset function to uplink policy interface, and ignore indexes in nscale policy if the number of published videos did not change.

Fixed

• Rate limited CPU warnings to at most once a minute in Voice Focus library, so that builder logs are not flooded.

[3.3.0] - 2022-05-12

Added

• Add support for hosting meetings in US GovCloud regions.
• Add support for starting live transcription in US GovCloud regions.

Removed

Changed

• Assume SDP section is sendrecv if no direction is present. This should have no impact on media negotiation.

Fixed

• Replace startVideoInput(null) and startAudioInput(null) withstopVideoInput and stopAudioInput for video, audio test in meeting readiness checker to stop video, audio input.
• Replace the deprecated API getRTCPeerConnectionStats with metricsDidReceive in meeting readiness checker.
• Prevent realtimeUnsubscribeFromVolumeIndicator from causing a fatal error when there are no subscriptions for the attendeeId.
• Subscribe to audioOutputDidChange in audio mix controller to fix the issue where the audio output is not updated before meeting start.

[3.2.0] - 2022-04-27

Added

• Add browser support information to content share guide.
• Readd layers allocation negotiation in Chromium based browsers to avoid resubscribing to preemptively turn off simulcast streams or to switch layers. Avoid duplicate RTP header extension and changing extension id.

Removed

Changed

• Clean up the HTML video element bounded to VideoTileState using unbindVideoElement API to fix Safari memory leak. If you do not intend to clean the video element, call unbindVideoElement API with cleanUpVideoElement set to false. Check PR#2217 for detailed information.

Fixed

• Fix issue where video resolution and framerate changes when toggle video transform.