@stream-io/video-client 1.47.0 → 1.48.0

package/CHANGELOG.md

@@ -2,6 +2,12 @@
 
 This file was generated using [@jscutlery/semver](https://github.com/jscutlery/semver).
 
+## [1.48.0](https://github.com/GetStream/stream-video-js/compare/@stream-io/video-client-1.47.0...@stream-io/video-client-1.48.0) (2026-04-28)
+
+### Features
+
+- **rn:** remove peer connection usage in speech detection ([#2200](https://github.com/GetStream/stream-video-js/issues/2200)) ([1c73d10](https://github.com/GetStream/stream-video-js/commit/1c73d10cc25761c08a8f9350e44137afaee33acf))
+
 ## [1.47.0](https://github.com/GetStream/stream-video-js/compare/@stream-io/video-client-1.46.1...@stream-io/video-client-1.47.0) (2026-04-15)
 
 ### Features

package/dist/index.browser.es.js

@@ -6284,7 +6284,7 @@ const getSdkVersion = (sdk) => {
     return sdk ? `${sdk.major}.${sdk.minor}.${sdk.patch}` : '0.0.0-development';
 };
 
-const version = "1.47.0";
+const version = "1.48.0";
 const [major, minor, patch] = version.split('.');
 let sdkInfo = {
     type: SdkType.PLAIN_JAVASCRIPT,
@@ -11932,192 +11932,6 @@ const createNoAudioDetector = (audioStream, options) => {
     return stop;
 };
 
-class RNSpeechDetector {
-    constructor(externalAudioStream) {
-        this.pc1 = new RTCPeerConnection({});
-        this.pc2 = new RTCPeerConnection({});
-        this.isStopped = false;
-        this.externalAudioStream = externalAudioStream;
-    }
-    /**
-     * Starts the speech detection.
-     */
-    async start(onSoundDetectedStateChanged) {
-        let detachListeners;
-        let unsubscribe;
-        try {
-            this.isStopped = false;
-            const audioStream = this.externalAudioStream != null
-                ? this.externalAudioStream
-                : await navigator.mediaDevices.getUserMedia({ audio: true });
-            this.audioStream = audioStream;
-            const onPc1IceCandidate = (e) => {
-                this.forwardIceCandidate(this.pc2, e.candidate);
-            };
-            const onPc2IceCandidate = (e) => {
-                this.forwardIceCandidate(this.pc1, e.candidate);
-            };
-            const onTrackPc2 = (e) => {
-                e.streams[0].getTracks().forEach((track) => {
-                    // In RN, the remote track is automatically added to the audio output device
-                    // so we need to mute it to avoid hearing the audio back
-                    // @ts-expect-error _setVolume is a private method in react-native-webrtc
-                    track._setVolume(0);
-                });
-            };
-            this.pc1.addEventListener('icecandidate', onPc1IceCandidate);
-            this.pc2.addEventListener('icecandidate', onPc2IceCandidate);
-            this.pc2.addEventListener('track', onTrackPc2);
-            detachListeners = () => {
-                this.pc1.removeEventListener('icecandidate', onPc1IceCandidate);
-                this.pc2.removeEventListener('icecandidate', onPc2IceCandidate);
-                this.pc2.removeEventListener('track', onTrackPc2);
-            };
-            audioStream
-                .getTracks()
-                .forEach((track) => this.pc1.addTrack(track, audioStream));
-            const offer = await this.pc1.createOffer({});
-            await this.pc2.setRemoteDescription(offer);
-            await this.pc1.setLocalDescription(offer);
-            const answer = await this.pc2.createAnswer();
-            await this.pc1.setRemoteDescription(answer);
-            await this.pc2.setLocalDescription(answer);
-            unsubscribe = this.onSpeakingDetectedStateChange(onSoundDetectedStateChanged);
-            return () => {
-                detachListeners?.();
-                unsubscribe?.();
-                this.stop();
-            };
-        }
-        catch (error) {
-            detachListeners?.();
-            unsubscribe?.();
-            this.stop();
-            const logger = videoLoggerSystem.getLogger('RNSpeechDetector');
-            logger.error('error handling permissions: ', error);
-            return () => { };
-        }
-    }
-    /**
-     * Stops the speech detection and releases all allocated resources.
-     */
-    stop() {
-        if (this.isStopped)
-            return;
-        this.isStopped = true;
-        this.pc1.close();
-        this.pc2.close();
-        if (this.externalAudioStream != null) {
-            this.externalAudioStream = undefined;
-        }
-        else {
-            this.cleanupAudioStream();
-        }
-    }
-    /**
-     * Public method that detects the audio levels and returns the status.
-     */
-    onSpeakingDetectedStateChange(onSoundDetectedStateChanged) {
-        const initialBaselineNoiseLevel = 0.13;
-        let baselineNoiseLevel = initialBaselineNoiseLevel;
-        let speechDetected = false;
-        let speechTimer;
-        let silenceTimer;
-        const audioLevelHistory = []; // Store recent audio levels for smoother detection
-        const historyLength = 10;
-        const silenceThreshold = 1.1;
-        const resetThreshold = 0.9;
-        const speechTimeout = 500; // Speech is set to true after 500ms of audio detection
-        const silenceTimeout = 5000; // Reset baseline after 5 seconds of silence
-        const checkAudioLevel = async () => {
-            try {
-                const stats = await this.pc1.getStats();
-                const report = flatten(stats);
-                // Audio levels are present inside stats of type `media-source` and of kind `audio`
-                const audioMediaSourceStats = report.find((stat) => stat.type === 'media-source' &&
-                    stat.kind === 'audio');
-                if (audioMediaSourceStats) {
-                    const { audioLevel } = audioMediaSourceStats;
-                    if (audioLevel) {
-                        // Update audio level history (with max historyLength sized array)
-                        audioLevelHistory.push(audioLevel);
-                        if (audioLevelHistory.length > historyLength) {
-                            audioLevelHistory.shift();
-                        }
-                        // Calculate average audio level
-                        const avgAudioLevel = audioLevelHistory.reduce((a, b) => a + b, 0) /
-                            audioLevelHistory.length;
-                        // Update baseline (if necessary) based on silence detection
-                        if (avgAudioLevel < baselineNoiseLevel * silenceThreshold) {
-                            if (!silenceTimer) {
-                                silenceTimer = setTimeout(() => {
-                                    baselineNoiseLevel = Math.min(avgAudioLevel * resetThreshold, initialBaselineNoiseLevel);
-                                }, silenceTimeout);
-                            }
-                        }
-                        else {
-                            clearTimeout(silenceTimer);
-                            silenceTimer = undefined;
-                        }
-                        // Speech detection with hysteresis
-                        if (avgAudioLevel > baselineNoiseLevel * 1.5) {
-                            if (!speechDetected) {
-                                speechDetected = true;
-                                onSoundDetectedStateChanged({
-                                    isSoundDetected: true,
-                                    audioLevel,
-                                });
-                            }
-                            clearTimeout(speechTimer);
-                            speechTimer = setTimeout(() => {
-                                speechDetected = false;
-                                onSoundDetectedStateChanged({
-                                    isSoundDetected: false,
-                                    audioLevel: 0,
-                                });
-                            }, speechTimeout);
-                        }
-                    }
-                }
-            }
-            catch (error) {
-                const logger = videoLoggerSystem.getLogger('RNSpeechDetector');
-                logger.error('error checking audio level from stats', error);
-            }
-        };
-        const intervalId = setInterval(checkAudioLevel, 250);
-        return () => {
-            clearInterval(intervalId);
-            clearTimeout(speechTimer);
-            clearTimeout(silenceTimer);
-        };
-    }
-    cleanupAudioStream() {
-        if (!this.audioStream) {
-            return;
-        }
-        this.audioStream.getTracks().forEach((track) => track.stop());
-        if (
-        // @ts-expect-error release() is present in react-native-webrtc
-        typeof this.audioStream.release === 'function') {
-            // @ts-expect-error called to dispose the stream in RN
-            this.audioStream.release();
-        }
-    }
-    forwardIceCandidate(destination, candidate) {
-        if (this.isStopped ||
-            !candidate ||
-            destination.signalingState === 'closed') {
-            return;
-        }
-        destination.addIceCandidate(candidate).catch(() => {
-            // silently ignore the error
-            const logger = videoLoggerSystem.getLogger('RNSpeechDetector');
-            logger.info('cannot add ice candidate - ignoring');
-        });
-    }
-}
-
 class MicrophoneManager extends AudioDeviceManager {
     constructor(call, devicePersistence, disableMode = 'stop-tracks') {
         super(call, new MicrophoneManagerState(disableMode, call.tracer), TrackType.AUDIO, devicePersistence);
@@ -12430,13 +12244,16 @@ class MicrophoneManager extends AudioDeviceManager {
                 return;
             await this.teardownSpeakingWhileMutedDetection();
             if (isReactNative()) {
-                this.rnSpeechDetector = new RNSpeechDetector();
-                const unsubscribe = await this.rnSpeechDetector.start((event) => {
+                const speechActivity = globalThis.streamRNVideoSDK?.nativeEvents?.speechActivity;
+                if (!speechActivity) {
+                    this.logger.warn('Native speech activity not available, make sure the "@stream-io/react-native-webrtc" peer dependency version is satisfied');
+                    return;
+                }
+                const unsubscribe = speechActivity.subscribe((event) => {
                     this.state.setSpeakingWhileMuted(event.isSoundDetected);
                 });
                 this.soundDetectorCleanup = async () => {
                     unsubscribe();
-                    this.rnSpeechDetector = undefined;
                 };
             }
             else {
@@ -16151,7 +15968,7 @@ class StreamClient {
         this.getUserAgent = () => {
             if (!this.cachedUserAgent) {
                 const { clientAppIdentifier = {} } = this.options;
-                const { sdkName = 'js', sdkVersion = "1.47.0", ...extras } = clientAppIdentifier;
+                const { sdkName = 'js', sdkVersion = "1.48.0", ...extras } = clientAppIdentifier;
                 this.cachedUserAgent = [
                     `stream-video-${sdkName}-v${sdkVersion}`,
                     ...Object.entries(extras).map(([key, value]) => `${key}=${value}`),
@@ -16787,5 +16604,5 @@ const humanize = (n) => {
     return String(n);
 };
 
-export { AudioSettingsRequestDefaultDeviceEnum, AudioSettingsResponseDefaultDeviceEnum, browsers as Browsers, Call, CallRecordingFailedEventRecordingTypeEnum, CallRecordingReadyEventRecordingTypeEnum, CallRecordingStartedEventRecordingTypeEnum, CallRecordingStoppedEventRecordingTypeEnum, CallState, CallType, CallTypes, CallingState, CameraManager, CameraManagerState, CreateDeviceRequestPushProviderEnum, DebounceType, DeviceManager, DeviceManagerState, DynascaleManager, ErrorFromResponse, FrameRecordingSettingsRequestModeEnum, FrameRecordingSettingsRequestQualityEnum, FrameRecordingSettingsResponseModeEnum, IndividualRecordingSettingsRequestModeEnum, IndividualRecordingSettingsResponseModeEnum, IngressAudioEncodingOptionsRequestChannelsEnum, IngressSourceRequestFpsEnum, IngressVideoLayerRequestCodecEnum, LayoutSettingsRequestNameEnum, MicrophoneManager, MicrophoneManagerState, NoiseCancellationSettingsModeEnum, OwnCapability, RNSpeechDetector, RTMPBroadcastRequestQualityEnum, RTMPSettingsRequestQualityEnum, RawRecordingSettingsRequestModeEnum, RawRecordingSettingsResponseModeEnum, RecordSettingsRequestModeEnum, RecordSettingsRequestQualityEnum, rxUtils as RxUtils, ScreenShareManager, ScreenShareState, events as SfuEvents, SfuJoinError, models as SfuModels, SpeakerManager, SpeakerState, StartClosedCaptionsRequestLanguageEnum, StartTranscriptionRequestLanguageEnum, StreamSfuClient, StreamVideoClient, StreamVideoReadOnlyStateStore, StreamVideoWriteableStateStore, TranscriptionSettingsRequestClosedCaptionModeEnum, TranscriptionSettingsRequestLanguageEnum, TranscriptionSettingsRequestModeEnum, TranscriptionSettingsResponseClosedCaptionModeEnum, TranscriptionSettingsResponseLanguageEnum, TranscriptionSettingsResponseModeEnum, VideoSettingsRequestCameraFacingEnum, VideoSettingsResponseCameraFacingEnum, ViewportTracker, VisibilityState, checkIfAudioOutputChangeSupported, combineComparators, conditional, createSoundDetector, defaultSortPreset, descending, deviceIds$, disposeOfMediaStream, dominantSpeaker, getAudioBrowserPermission, getAudioDevices, getAudioOutputDevices, getAudioStream, getClientDetails, getDeviceState, getScreenShareStream, getSdkInfo, getVideoBrowserPermission, getVideoDevices, getVideoStream, getWebRTCInfo, hasAudio$1 as hasAudio, hasPausedTrack, hasScreenShare, hasScreenShareAudio, hasVideo, humanize, isPinned, livestreamOrAudioRoomSortPreset, logToConsole, name, noopComparator, paginatedLayoutSortPreset, pinned, publishingAudio, publishingVideo, reactionType, resolveDeviceId, role, screenSharing, setDeviceInfo, setOSInfo, setPowerState, setSdkInfo, setThermalState, setWebRTCInfo, speakerLayoutSortPreset, speaking, videoLoggerSystem, withParticipantSource };
+export { AudioSettingsRequestDefaultDeviceEnum, AudioSettingsResponseDefaultDeviceEnum, browsers as Browsers, Call, CallRecordingFailedEventRecordingTypeEnum, CallRecordingReadyEventRecordingTypeEnum, CallRecordingStartedEventRecordingTypeEnum, CallRecordingStoppedEventRecordingTypeEnum, CallState, CallType, CallTypes, CallingState, CameraManager, CameraManagerState, CreateDeviceRequestPushProviderEnum, DebounceType, DeviceManager, DeviceManagerState, DynascaleManager, ErrorFromResponse, FrameRecordingSettingsRequestModeEnum, FrameRecordingSettingsRequestQualityEnum, FrameRecordingSettingsResponseModeEnum, IndividualRecordingSettingsRequestModeEnum, IndividualRecordingSettingsResponseModeEnum, IngressAudioEncodingOptionsRequestChannelsEnum, IngressSourceRequestFpsEnum, IngressVideoLayerRequestCodecEnum, LayoutSettingsRequestNameEnum, MicrophoneManager, MicrophoneManagerState, NoiseCancellationSettingsModeEnum, OwnCapability, RTMPBroadcastRequestQualityEnum, RTMPSettingsRequestQualityEnum, RawRecordingSettingsRequestModeEnum, RawRecordingSettingsResponseModeEnum, RecordSettingsRequestModeEnum, RecordSettingsRequestQualityEnum, rxUtils as RxUtils, ScreenShareManager, ScreenShareState, events as SfuEvents, SfuJoinError, models as SfuModels, SpeakerManager, SpeakerState, StartClosedCaptionsRequestLanguageEnum, StartTranscriptionRequestLanguageEnum, StreamSfuClient, StreamVideoClient, StreamVideoReadOnlyStateStore, StreamVideoWriteableStateStore, TranscriptionSettingsRequestClosedCaptionModeEnum, TranscriptionSettingsRequestLanguageEnum, TranscriptionSettingsRequestModeEnum, TranscriptionSettingsResponseClosedCaptionModeEnum, TranscriptionSettingsResponseLanguageEnum, TranscriptionSettingsResponseModeEnum, VideoSettingsRequestCameraFacingEnum, VideoSettingsResponseCameraFacingEnum, ViewportTracker, VisibilityState, checkIfAudioOutputChangeSupported, combineComparators, conditional, createSoundDetector, defaultSortPreset, descending, deviceIds$, disposeOfMediaStream, dominantSpeaker, getAudioBrowserPermission, getAudioDevices, getAudioOutputDevices, getAudioStream, getClientDetails, getDeviceState, getScreenShareStream, getSdkInfo, getVideoBrowserPermission, getVideoDevices, getVideoStream, getWebRTCInfo, hasAudio$1 as hasAudio, hasPausedTrack, hasScreenShare, hasScreenShareAudio, hasVideo, humanize, isPinned, livestreamOrAudioRoomSortPreset, logToConsole, name, noopComparator, paginatedLayoutSortPreset, pinned, publishingAudio, publishingVideo, reactionType, resolveDeviceId, role, screenSharing, setDeviceInfo, setOSInfo, setPowerState, setSdkInfo, setThermalState, setWebRTCInfo, speakerLayoutSortPreset, speaking, videoLoggerSystem, withParticipantSource };
 //# sourceMappingURL=index.browser.es.js.map