@defcomm/meetingsdk 0.6.8 → 0.6.9

package/dist/index.d.mts

@@ -339,6 +339,12 @@ declare class VideoSDKCore {
     getMeetingSummary(roomCode: string, sessionId: string): Promise<{
         summary: MeetingSummary | null;
     }>;
+    private acquireOptionalLocalMedia;
+    /**
+     * Wakes the active-speaker Web Audio pipeline. Call this from a user gesture
+     * whenever possible (Join, first click inside the meeting, etc.).
+     */
+    resumeActiveSpeaker(): Promise<boolean>;
     joinMeeting(config: Omit<MeetingConfig, "participantId">): Promise<void>;
     private connect;
     private createSFUClient;
@@ -427,6 +433,7 @@ type MeetingContextValue = {
     }>;
     toggleMic: () => void;
     toggleCam: () => void;
+    resumeActiveSpeaker: () => Promise<boolean>;
     startScreenShare: () => Promise<MediaStream | null>;
     stopScreenShare: () => void;
     sendMessage: (input: ChatInput) => void;
@@ -524,6 +531,7 @@ declare const useMeeting: (handlers?: {
     }>;
     toggleMic: () => void;
     toggleCam: () => void;
+    resumeActiveSpeaker: () => Promise<boolean>;
     startScreenShare: () => Promise<MediaStream | null>;
     stopScreenShare: () => void;
     sendMessage: (input: ChatInput) => void;
@@ -630,6 +638,15 @@ interface ActiveSpeakerDetectorOptions {
     releaseMs?: number;
     onActiveSpeakersChanged?: (participantIds: string[]) => void;
 }
+/**
+ * Detects currently speaking participants from the actual audio tracks that
+ * are being received/published by WebRTC.
+ *
+ * The AudioContext is explicitly resumable so an application can wake it from
+ * a real user gesture (for example, the Join button). This matters in browsers
+ * that create Web Audio contexts in the "suspended" state until the page has
+ * received user interaction.
+ */
 declare class ActiveSpeakerDetector {
     private audioContext;
     private speakers;
@@ -642,7 +659,12 @@ declare class ActiveSpeakerDetector {
     private lastEmitted;
     constructor(options?: ActiveSpeakerDetectorOptions);
     private ensureAudioContext;
-    resume(): Promise<void>;
+    /**
+     * Create/wake the Web Audio context. Call this from a user gesture when
+     * possible. Returns true when the context is running after the call.
+     */
+    resume(): Promise<boolean>;
+    getAudioContextState(): AudioContextState | "unavailable";
     addTrack(participantId: string, track: MediaStreamTrack): void;
     removeParticipant(participantId: string): void;
     removeTrack(track: MediaStreamTrack): void;
@@ -650,6 +672,7 @@ declare class ActiveSpeakerDetector {
     start(): void;
     stop(): void;
     private sample;
+    private getActiveSpeakers;
     private getActiveSpeakerIds;
     private emitIfChanged;
     destroy(): Promise<void>;

package/dist/index.d.ts

@@ -339,6 +339,12 @@ declare class VideoSDKCore {
     getMeetingSummary(roomCode: string, sessionId: string): Promise<{
         summary: MeetingSummary | null;
     }>;
+    private acquireOptionalLocalMedia;
+    /**
+     * Wakes the active-speaker Web Audio pipeline. Call this from a user gesture
+     * whenever possible (Join, first click inside the meeting, etc.).
+     */
+    resumeActiveSpeaker(): Promise<boolean>;
     joinMeeting(config: Omit<MeetingConfig, "participantId">): Promise<void>;
     private connect;
     private createSFUClient;
@@ -427,6 +433,7 @@ type MeetingContextValue = {
     }>;
     toggleMic: () => void;
     toggleCam: () => void;
+    resumeActiveSpeaker: () => Promise<boolean>;
     startScreenShare: () => Promise<MediaStream | null>;
     stopScreenShare: () => void;
     sendMessage: (input: ChatInput) => void;
@@ -524,6 +531,7 @@ declare const useMeeting: (handlers?: {
     }>;
     toggleMic: () => void;
     toggleCam: () => void;
+    resumeActiveSpeaker: () => Promise<boolean>;
     startScreenShare: () => Promise<MediaStream | null>;
     stopScreenShare: () => void;
     sendMessage: (input: ChatInput) => void;
@@ -630,6 +638,15 @@ interface ActiveSpeakerDetectorOptions {
     releaseMs?: number;
     onActiveSpeakersChanged?: (participantIds: string[]) => void;
 }
+/**
+ * Detects currently speaking participants from the actual audio tracks that
+ * are being received/published by WebRTC.
+ *
+ * The AudioContext is explicitly resumable so an application can wake it from
+ * a real user gesture (for example, the Join button). This matters in browsers
+ * that create Web Audio contexts in the "suspended" state until the page has
+ * received user interaction.
+ */
 declare class ActiveSpeakerDetector {
     private audioContext;
     private speakers;
@@ -642,7 +659,12 @@ declare class ActiveSpeakerDetector {
     private lastEmitted;
     constructor(options?: ActiveSpeakerDetectorOptions);
     private ensureAudioContext;
-    resume(): Promise<void>;
+    /**
+     * Create/wake the Web Audio context. Call this from a user gesture when
+     * possible. Returns true when the context is running after the call.
+     */
+    resume(): Promise<boolean>;
+    getAudioContextState(): AudioContextState | "unavailable";
     addTrack(participantId: string, track: MediaStreamTrack): void;
     removeParticipant(participantId: string): void;
     removeTrack(track: MediaStreamTrack): void;
@@ -650,6 +672,7 @@ declare class ActiveSpeakerDetector {
     start(): void;
     stop(): void;
     private sample;
+    private getActiveSpeakers;
     private getActiveSpeakerIds;
     private emitIfChanged;
     destroy(): Promise<void>;

package/dist/index.js

@@ -63,13 +63,16 @@ var ActiveSpeakerDetector = class {
     this.timer = null;
     this.lastEmitted = "";
     this.intervalMs = options.intervalMs ?? 100;
-    this.threshold = options.threshold ?? 0.025;
-    this.releaseThreshold = options.releaseThreshold ?? 0.015;
-    this.releaseMs = options.releaseMs ?? 600;
+    this.threshold = options.threshold ?? 0.02;
+    this.releaseThreshold = options.releaseThreshold ?? 0.01;
+    this.releaseMs = options.releaseMs ?? 650;
     this.onActiveSpeakersChanged = options.onActiveSpeakersChanged ?? (() => {
     });
   }
   ensureAudioContext() {
+    if (typeof window === "undefined") {
+      throw new Error("AudioContext is only available in the browser");
+    }
     if (!this.audioContext) {
       const AudioContextClass = window.AudioContext || window.webkitAudioContext;
       if (!AudioContextClass) {
@@ -79,28 +82,52 @@ var ActiveSpeakerDetector = class {
     }
     return this.audioContext;
   }
+  /**
+   * Create/wake the Web Audio context. Call this from a user gesture when
+   * possible. Returns true when the context is running after the call.
+   */
   async resume() {
-    if (!this.audioContext) return;
-    if (this.audioContext.state === "suspended") {
-      try {
-        await this.audioContext.resume();
-      } catch {
+    try {
+      const context = this.ensureAudioContext();
+      if (context.state === "suspended") {
+        try {
+          await context.resume();
+        } catch (error) {
+          console.warn("[ActiveSpeaker] AudioContext resume failed", error);
+        }
       }
+      return context.state === "running";
+    } catch (error) {
+      console.warn("[ActiveSpeaker] AudioContext unavailable", error);
+      return false;
     }
   }
+  getAudioContextState() {
+    if (!this.audioContext) return "unavailable";
+    return this.audioContext.state;
+  }
   addTrack(participantId, track) {
-    if (!participantId) return;
-    if (track.kind !== "audio") return;
+    if (!participantId || track.kind !== "audio") return;
     this.removeParticipant(participantId);
-    const context = this.ensureAudioContext();
+    let context;
+    try {
+      context = this.ensureAudioContext();
+    } catch (error) {
+      console.warn("[ActiveSpeaker] Could not create AudioContext", error);
+      return;
+    }
     const stream = new MediaStream([track]);
     const source = context.createMediaStreamSource(stream);
     const analyser = context.createAnalyser();
     analyser.fftSize = 512;
-    analyser.smoothingTimeConstant = 0.65;
+    analyser.smoothingTimeConstant = 0.35;
     analyser.minDecibels = -90;
     analyser.maxDecibels = -10;
     source.connect(analyser);
+    const sink = context.createGain();
+    sink.gain.value = 0;
+    analyser.connect(sink);
+    sink.connect(context.destination);
     const data = new Float32Array(analyser.fftSize);
     const state = {
       participantId,
@@ -108,15 +135,20 @@ var ActiveSpeakerDetector = class {
       stream,
       source,
       analyser,
+      sink,
       data,
       active: false,
+      level: 0,
       lastActiveAt: 0
     };
     this.speakers.set(participantId, state);
     track.addEventListener(
       "ended",
       () => {
-        this.removeParticipant(participantId);
+        const current = this.speakers.get(participantId);
+        if (current?.track === track) {
+          this.removeParticipant(participantId);
+        }
       },
       { once: true }
     );
@@ -127,15 +159,19 @@ var ActiveSpeakerDetector = class {
     const state = this.speakers.get(participantId);
     if (!state) return;
     state.active = false;
+    state.level = 0;
     try {
       state.source.disconnect();
     } catch {
     }
-    state.stream.getTracks().forEach((track) => {
-      if (track !== state.track) {
-        track.stop();
-      }
-    });
+    try {
+      state.analyser.disconnect();
+    } catch {
+    }
+    try {
+      state.sink.disconnect();
+    } catch {
+    }
     this.speakers.delete(participantId);
     this.emitIfChanged();
   }
@@ -147,7 +183,7 @@ var ActiveSpeakerDetector = class {
     }
   }
   clear() {
-    for (const participantId of this.speakers.keys()) {
+    for (const participantId of [...this.speakers.keys()]) {
       this.removeParticipant(participantId);
     }
     this.speakers.clear();
@@ -167,6 +203,7 @@ var ActiveSpeakerDetector = class {
     const now = performance.now();
     for (const state of this.speakers.values()) {
       if (state.track.readyState !== "live" || state.track.muted || !state.track.enabled) {
+        state.level = 0;
         if (state.active) {
           state.active = false;
         }
@@ -179,30 +216,33 @@ var ActiveSpeakerDetector = class {
         sumSquares += sample * sample;
       }
       const rms = Math.sqrt(sumSquares / state.data.length);
+      state.level = rms;
       if (!state.active) {
         if (rms >= this.threshold) {
           state.active = true;
           state.lastActiveAt = now;
         }
-      } else {
-        if (rms >= this.releaseThreshold) {
-          state.lastActiveAt = now;
-        } else if (now - state.lastActiveAt >= this.releaseMs) {
-          state.active = false;
-        }
+      } else if (rms >= this.releaseThreshold) {
+        state.lastActiveAt = now;
+      } else if (now - state.lastActiveAt >= this.releaseMs) {
+        state.active = false;
       }
     }
     this.emitIfChanged();
   }
+  getActiveSpeakers() {
+    return Array.from(this.speakers.values()).filter((speaker) => speaker.active).sort((a, b) => {
+      if (b.level !== a.level) return b.level - a.level;
+      return a.participantId.localeCompare(b.participantId);
+    });
+  }
   getActiveSpeakerIds() {
-    return Array.from(this.speakers.values()).filter((speaker) => speaker.active).map((speaker) => speaker.participantId);
+    return this.getActiveSpeakers().map((speaker) => speaker.participantId);
   }
   emitIfChanged(force = false) {
-    const ids = this.getActiveSpeakerIds().sort();
+    const ids = this.getActiveSpeakerIds();
     const key = ids.join("|");
-    if (!force && key === this.lastEmitted) {
-      return;
-    }
+    if (!force && key === this.lastEmitted) return;
     this.lastEmitted = key;
     this.onActiveSpeakersChanged(ids);
   }
@@ -2347,26 +2387,11 @@ var VideoSDKCore = class {
   async initLocal(video, name) {
     this.participantName = name;
     try {
-      this.localStream = await navigator.mediaDevices.getUserMedia({
-        video: {
-          width: { ideal: 1280 },
-          height: { ideal: 720 }
-        },
-        audio: {
-          echoCancellation: true,
-          noiseSuppression: true,
-          autoGainControl: true
-        }
-      });
+      this.localStream = await this.acquireOptionalLocalMedia();
       const localAudioTrack = this.localStream.getAudioTracks()[0];
       if (localAudioTrack) {
         this.activeSpeakerDetector.addTrack(this.myId, localAudioTrack);
       }
-      const hasVideo = this.localStream.getVideoTracks().some((t) => t.readyState === "live");
-      const hasAudio = this.localStream.getAudioTracks().some((t) => t.readyState === "live");
-      if (!hasVideo || !hasAudio) {
-        throw new Error(`Missing tracks: video=${hasVideo}, audio=${hasAudio}`);
-      }
       video.srcObject = this.localStream;
       this.updateLocalMedia();
     } catch (err) {
@@ -2383,8 +2408,8 @@ var VideoSDKCore = class {
         stream: this.localStream,
         cameraTrack: this.localStream.getVideoTracks()[0],
         audioTrack: this.localStream.getAudioTracks()[0],
-        micEnabled: this.localStream.getAudioTracks()[0]?.enabled ?? !audioMuted,
-        camEnabled: this.localStream.getVideoTracks()[0]?.enabled ?? !videoMuted,
+        micEnabled: this.localStream.getAudioTracks()[0]?.enabled ?? false,
+        camEnabled: this.localStream.getVideoTracks()[0]?.enabled ?? false,
         isScreenSharing: this.isScreenSharing,
         screenStream: this.screenStream,
         screenTrack: this.screenStream?.getVideoTracks()[0]
@@ -2411,6 +2436,70 @@ var VideoSDKCore = class {
   async getMeetingSummary(roomCode, sessionId) {
     return this.sfuClient.getMeetingSummary(roomCode, sessionId);
   }
+  async acquireOptionalLocalMedia(cameraDeviceId, microphoneDeviceId) {
+    const combined = new MediaStream();
+    if (!navigator.mediaDevices?.getUserMedia) {
+      console.warn(
+        "[SDK] Camera/microphone access is not supported; joining without local media."
+      );
+      return combined;
+    }
+    const cameraPromise = navigator.mediaDevices.getUserMedia({
+      video: cameraDeviceId ? { deviceId: { exact: cameraDeviceId } } : {
+        width: { ideal: 1280 },
+        height: { ideal: 720 }
+      },
+      audio: false
+    });
+    const microphonePromise = navigator.mediaDevices.getUserMedia({
+      video: false,
+      audio: microphoneDeviceId ? {
+        deviceId: { exact: microphoneDeviceId },
+        echoCancellation: true,
+        noiseSuppression: true,
+        autoGainControl: true
+      } : {
+        echoCancellation: true,
+        noiseSuppression: true,
+        autoGainControl: true
+      }
+    });
+    const [cameraResult, microphoneResult] = await Promise.allSettled([
+      cameraPromise,
+      microphonePromise
+    ]);
+    if (cameraResult.status === "fulfilled") {
+      for (const track of cameraResult.value.getVideoTracks()) {
+        combined.addTrack(track);
+      }
+    } else {
+      console.warn("[SDK] Camera unavailable; continuing without video.", {
+        name: cameraResult.reason?.name,
+        message: cameraResult.reason?.message
+      });
+    }
+    if (microphoneResult.status === "fulfilled") {
+      for (const track of microphoneResult.value.getAudioTracks()) {
+        combined.addTrack(track);
+      }
+    } else {
+      console.warn(
+        "[SDK] Microphone unavailable; continuing without audio.",
+        {
+          name: microphoneResult.reason?.name,
+          message: microphoneResult.reason?.message
+        }
+      );
+    }
+    return combined;
+  }
+  /**
+   * Wakes the active-speaker Web Audio pipeline. Call this from a user gesture
+   * whenever possible (Join, first click inside the meeting, etc.).
+   */
+  async resumeActiveSpeaker() {
+    return this.activeSpeakerDetector.resume();
+  }
   // JOIN
   async joinMeeting(config) {
     const {
@@ -2454,15 +2543,10 @@ var VideoSDKCore = class {
     this.room.id = roomId;
     this.room.name = name;
     if (!this.localStream) {
-      this.localStream = await navigator.mediaDevices.getUserMedia({
-        video: cameraDeviceId ? { deviceId: { exact: cameraDeviceId } } : true,
-        audio: microphoneDeviceId ? {
-          deviceId: { exact: microphoneDeviceId },
-          echoCancellation: true,
-          noiseSuppression: true,
-          autoGainControl: true
-        } : true
-      });
+      this.localStream = await this.acquireOptionalLocalMedia(
+        cameraDeviceId,
+        microphoneDeviceId
+      );
       const localAudioTrack = this.localStream.getAudioTracks()[0];
       if (localAudioTrack) {
         this.activeSpeakerDetector.addTrack(this.myId, localAudioTrack);
@@ -2507,8 +2591,8 @@ var VideoSDKCore = class {
     await this.sfuClient.getToken(roomId, {
       displayName: name,
       participantId: this.myId,
-      audioMuted: !(this.localStream?.getAudioTracks()[0]?.enabled ?? true),
-      videoMuted: !(this.localStream?.getVideoTracks()[0]?.enabled ?? true)
+      audioMuted: !(this.localStream?.getAudioTracks()[0]?.enabled ?? false),
+      videoMuted: !(this.localStream?.getVideoTracks()[0]?.enabled ?? false)
     });
   }
   createSFUClient(token) {
@@ -3568,6 +3652,7 @@ var MeetingProvider = ({
       getMeetingSummary: sdk.getMeetingSummary.bind(sdk),
       toggleMic: sdk.toggleMic.bind(sdk),
       toggleCam: sdk.toggleCam.bind(sdk),
+      resumeActiveSpeaker: sdk.resumeActiveSpeaker.bind(sdk),
       startScreenShare: sdk.startScreenShare.bind(sdk),
       stopScreenShare: sdk.stopScreenShare.bind(sdk),
       sendMessage: sdk.sendChatMessage.bind(sdk),