@tensamin/audio 0.1.11 → 0.1.12

package/README.md

@@ -101,8 +101,8 @@ vad: {
   energyVad?: {
     smoothing: number;                 // Default: 0.95
     initialNoiseFloor: number;         // Default: 0.001
-    noiseFloorAdaptRateQuiet: number;  // Default: 0.005
-    noiseFloorAdaptRateLoud: number;   // Default: 0.01
+    noiseFloorAdaptRateQuiet: number;  // Default: 0.01
+    noiseFloorAdaptRateLoud: number;   // Default: 0.05
     minSNR: number;                    // Default: 10.0 (dB)
     snrRange: number;                  // Default: 10.0 (dB)
     minEnergy: number;                 // Default: 0.001

package/dist/{chunk-GFLVGUTU.mjs → chunk-DYY2MXMU.mjs}

@@ -61,25 +61,33 @@ var VADStateMachine = class {
         newState = "silent";
         this.lastSilenceTime = timestamp;
       }
-    } else if (this.currentState === "speech_starting" || this.currentState === "speaking") {
+    } else if (this.currentState === "speech_starting") {
+      if (probability >= stopThreshold) {
+        const speechDuration = timestamp - this.speechStartTime;
+        if (speechDuration >= minSpeechDurationMs) {
+          newState = "speaking";
+        } else {
+          newState = "speech_starting";
+        }
+        this.lastSpeechTime = timestamp;
+      } else {
+        newState = "silent";
+        this.lastSilenceTime = timestamp;
+      }
+    } else if (this.currentState === "speaking") {
       if (probability >= stopThreshold) {
         newState = "speaking";
         this.lastSpeechTime = timestamp;
       } else {
         const timeSinceSpeech = timestamp - this.lastSpeechTime;
-        const speechDuration = timestamp - this.speechStartTime;
         if (timeSinceSpeech < hangoverMs) {
           newState = "speaking";
-        } else if (speechDuration < minSpeechDurationMs) {
-          newState = "silent";
-          this.lastSilenceTime = timestamp;
         } else {
           newState = "speech_ending";
           this.lastSilenceTime = timestamp;
         }
       }
     }
-    if (newState === "speech_starting") newState = "speaking";
     if (newState === "speech_ending") newState = "silent";
     this.currentState = newState;
     return {

package/dist/{chunk-B36JBXOK.mjs → chunk-KEWK2OKV.mjs}

@@ -3,8 +3,8 @@ var createEnergyVadWorkletCode = (vadConfig) => {
   const energyParams = vadConfig?.energyVad || {};
   const smoothing = energyParams.smoothing ?? 0.95;
   const initialNoiseFloor = energyParams.initialNoiseFloor ?? 1e-3;
-  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 5e-3;
-  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.01;
+  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 0.01;
+  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.05;
   const minSNR = energyParams.minSNR ?? 10;
   const snrRange = energyParams.snrRange ?? 10;
   const minEnergy = energyParams.minEnergy ?? 1e-3;
@@ -61,9 +61,9 @@ class EnergyVadProcessor extends AudioWorkletProcessor {
       
       let multiplier = 1.0;
       if (this.isSpeaking) {
-        multiplier = 0.01;
+        multiplier = 0.05;
       } else if (snrDb > 20) {
-        multiplier = 0.1;
+        multiplier = 0.2;
       }
       
       const adaptRate = this.noiseFloorAdaptRateLoud * multiplier;

package/dist/{chunk-RLZVZ6D6.mjs → chunk-Q2I22TJG.mjs}

@@ -1,6 +1,6 @@
 import {
   createAudioPipeline
-} from "./chunk-I5AR7XQD.mjs";
+} from "./chunk-SMZJFNRU.mjs";
 
 // src/livekit/integration.ts
 async function attachProcessingToTrack(track, config = {}) {

package/dist/{chunk-I5AR7XQD.mjs → chunk-SMZJFNRU.mjs}

@@ -1,6 +1,6 @@
 import {
   VADStateMachine
-} from "./chunk-GFLVGUTU.mjs";
+} from "./chunk-DYY2MXMU.mjs";
 import {
   getAudioContext,
   registerPipeline,
@@ -9,7 +9,7 @@ import {
 import {
   getNoiseSuppressionPlugin,
   getVADPlugin
-} from "./chunk-3I4OQD2L.mjs";
+} from "./chunk-XZSFQJW4.mjs";
 
 // src/pipeline/audio-pipeline.ts
 import mitt from "mitt";

package/dist/{chunk-3I4OQD2L.mjs → chunk-XZSFQJW4.mjs}

@@ -3,7 +3,7 @@ import {
 } from "./chunk-XO6B3D4A.mjs";
 import {
   EnergyVADPlugin
-} from "./chunk-B36JBXOK.mjs";
+} from "./chunk-KEWK2OKV.mjs";
 
 // src/extensibility/plugins.ts
 var nsPlugins = /* @__PURE__ */ new Map();

package/dist/extensibility/plugins.js

@@ -106,8 +106,8 @@ var createEnergyVadWorkletCode = (vadConfig) => {
   const energyParams = vadConfig?.energyVad || {};
   const smoothing = energyParams.smoothing ?? 0.95;
   const initialNoiseFloor = energyParams.initialNoiseFloor ?? 1e-3;
-  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 5e-3;
-  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.01;
+  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 0.01;
+  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.05;
   const minSNR = energyParams.minSNR ?? 10;
   const snrRange = energyParams.snrRange ?? 10;
   const minEnergy = energyParams.minEnergy ?? 1e-3;
@@ -164,9 +164,9 @@ class EnergyVadProcessor extends AudioWorkletProcessor {
       
       let multiplier = 1.0;
       if (this.isSpeaking) {
-        multiplier = 0.01;
+        multiplier = 0.05;
       } else if (snrDb > 20) {
-        multiplier = 0.1;
+        multiplier = 0.2;
       }
       
       const adaptRate = this.noiseFloorAdaptRateLoud * multiplier;

package/dist/extensibility/plugins.mjs

@@ -3,9 +3,9 @@ import {
   getVADPlugin,
   registerNoiseSuppressionPlugin,
   registerVADPlugin
-} from "../chunk-3I4OQD2L.mjs";
+} from "../chunk-XZSFQJW4.mjs";
 import "../chunk-XO6B3D4A.mjs";
-import "../chunk-B36JBXOK.mjs";
+import "../chunk-KEWK2OKV.mjs";
 export {
   getNoiseSuppressionPlugin,
   getVADPlugin,

package/dist/index.js

@@ -158,8 +158,8 @@ var createEnergyVadWorkletCode = (vadConfig) => {
   const energyParams = vadConfig?.energyVad || {};
   const smoothing = energyParams.smoothing ?? 0.95;
   const initialNoiseFloor = energyParams.initialNoiseFloor ?? 1e-3;
-  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 5e-3;
-  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.01;
+  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 0.01;
+  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.05;
   const minSNR = energyParams.minSNR ?? 10;
   const snrRange = energyParams.snrRange ?? 10;
   const minEnergy = energyParams.minEnergy ?? 1e-3;
@@ -216,9 +216,9 @@ class EnergyVadProcessor extends AudioWorkletProcessor {
       
       let multiplier = 1.0;
       if (this.isSpeaking) {
-        multiplier = 0.01;
+        multiplier = 0.05;
       } else if (snrDb > 20) {
-        multiplier = 0.1;
+        multiplier = 0.2;
       }
       
       const adaptRate = this.noiseFloorAdaptRateLoud * multiplier;
@@ -412,25 +412,33 @@ var VADStateMachine = class {
         newState = "silent";
         this.lastSilenceTime = timestamp;
       }
-    } else if (this.currentState === "speech_starting" || this.currentState === "speaking") {
+    } else if (this.currentState === "speech_starting") {
+      if (probability >= stopThreshold) {
+        const speechDuration = timestamp - this.speechStartTime;
+        if (speechDuration >= minSpeechDurationMs) {
+          newState = "speaking";
+        } else {
+          newState = "speech_starting";
+        }
+        this.lastSpeechTime = timestamp;
+      } else {
+        newState = "silent";
+        this.lastSilenceTime = timestamp;
+      }
+    } else if (this.currentState === "speaking") {
       if (probability >= stopThreshold) {
         newState = "speaking";
         this.lastSpeechTime = timestamp;
       } else {
         const timeSinceSpeech = timestamp - this.lastSpeechTime;
-        const speechDuration = timestamp - this.speechStartTime;
         if (timeSinceSpeech < hangoverMs) {
           newState = "speaking";
-        } else if (speechDuration < minSpeechDurationMs) {
-          newState = "silent";
-          this.lastSilenceTime = timestamp;
         } else {
           newState = "speech_ending";
           this.lastSilenceTime = timestamp;
         }
       }
     }
-    if (newState === "speech_starting") newState = "speaking";
     if (newState === "speech_ending") newState = "silent";
     this.currentState = newState;
     return {

package/dist/index.mjs

@@ -1,13 +1,13 @@
 import "./chunk-WBQAMGXK.mjs";
 import {
   attachProcessingToTrack
-} from "./chunk-RLZVZ6D6.mjs";
+} from "./chunk-Q2I22TJG.mjs";
 import {
   createAudioPipeline
-} from "./chunk-I5AR7XQD.mjs";
+} from "./chunk-SMZJFNRU.mjs";
 import {
   VADStateMachine
-} from "./chunk-GFLVGUTU.mjs";
+} from "./chunk-DYY2MXMU.mjs";
 import {
   closeAudioContext,
   getAudioContext,
@@ -21,13 +21,13 @@ import {
   getVADPlugin,
   registerNoiseSuppressionPlugin,
   registerVADPlugin
-} from "./chunk-3I4OQD2L.mjs";
+} from "./chunk-XZSFQJW4.mjs";
 import {
   RNNoisePlugin
 } from "./chunk-XO6B3D4A.mjs";
 import {
   EnergyVADPlugin
-} from "./chunk-B36JBXOK.mjs";
+} from "./chunk-KEWK2OKV.mjs";
 export {
   EnergyVADPlugin,
   RNNoisePlugin,

package/dist/livekit/integration.js

@@ -127,8 +127,8 @@ var createEnergyVadWorkletCode = (vadConfig) => {
   const energyParams = vadConfig?.energyVad || {};
   const smoothing = energyParams.smoothing ?? 0.95;
   const initialNoiseFloor = energyParams.initialNoiseFloor ?? 1e-3;
-  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 5e-3;
-  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.01;
+  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 0.01;
+  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.05;
   const minSNR = energyParams.minSNR ?? 10;
   const snrRange = energyParams.snrRange ?? 10;
   const minEnergy = energyParams.minEnergy ?? 1e-3;
@@ -185,9 +185,9 @@ class EnergyVadProcessor extends AudioWorkletProcessor {
       
       let multiplier = 1.0;
       if (this.isSpeaking) {
-        multiplier = 0.01;
+        multiplier = 0.05;
       } else if (snrDb > 20) {
-        multiplier = 0.1;
+        multiplier = 0.2;
       }
       
       const adaptRate = this.noiseFloorAdaptRateLoud * multiplier;
@@ -375,25 +375,33 @@ var VADStateMachine = class {
         newState = "silent";
         this.lastSilenceTime = timestamp;
       }
-    } else if (this.currentState === "speech_starting" || this.currentState === "speaking") {
+    } else if (this.currentState === "speech_starting") {
+      if (probability >= stopThreshold) {
+        const speechDuration = timestamp - this.speechStartTime;
+        if (speechDuration >= minSpeechDurationMs) {
+          newState = "speaking";
+        } else {
+          newState = "speech_starting";
+        }
+        this.lastSpeechTime = timestamp;
+      } else {
+        newState = "silent";
+        this.lastSilenceTime = timestamp;
+      }
+    } else if (this.currentState === "speaking") {
       if (probability >= stopThreshold) {
         newState = "speaking";
         this.lastSpeechTime = timestamp;
       } else {
         const timeSinceSpeech = timestamp - this.lastSpeechTime;
-        const speechDuration = timestamp - this.speechStartTime;
         if (timeSinceSpeech < hangoverMs) {
           newState = "speaking";
-        } else if (speechDuration < minSpeechDurationMs) {
-          newState = "silent";
-          this.lastSilenceTime = timestamp;
         } else {
           newState = "speech_ending";
           this.lastSilenceTime = timestamp;
         }
       }
     }
-    if (newState === "speech_starting") newState = "speaking";
     if (newState === "speech_ending") newState = "silent";
     this.currentState = newState;
     return {

package/dist/livekit/integration.mjs

@@ -1,12 +1,12 @@
 import {
   attachProcessingToTrack
-} from "../chunk-RLZVZ6D6.mjs";
-import "../chunk-I5AR7XQD.mjs";
-import "../chunk-GFLVGUTU.mjs";
+} from "../chunk-Q2I22TJG.mjs";
+import "../chunk-SMZJFNRU.mjs";
+import "../chunk-DYY2MXMU.mjs";
 import "../chunk-OZ7KMC4S.mjs";
-import "../chunk-3I4OQD2L.mjs";
+import "../chunk-XZSFQJW4.mjs";
 import "../chunk-XO6B3D4A.mjs";
-import "../chunk-B36JBXOK.mjs";
+import "../chunk-KEWK2OKV.mjs";
 export {
   attachProcessingToTrack
 };

package/dist/pipeline/audio-pipeline.js

@@ -125,8 +125,8 @@ var createEnergyVadWorkletCode = (vadConfig) => {
   const energyParams = vadConfig?.energyVad || {};
   const smoothing = energyParams.smoothing ?? 0.95;
   const initialNoiseFloor = energyParams.initialNoiseFloor ?? 1e-3;
-  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 5e-3;
-  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.01;
+  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 0.01;
+  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.05;
   const minSNR = energyParams.minSNR ?? 10;
   const snrRange = energyParams.snrRange ?? 10;
   const minEnergy = energyParams.minEnergy ?? 1e-3;
@@ -183,9 +183,9 @@ class EnergyVadProcessor extends AudioWorkletProcessor {
       
       let multiplier = 1.0;
       if (this.isSpeaking) {
-        multiplier = 0.01;
+        multiplier = 0.05;
       } else if (snrDb > 20) {
-        multiplier = 0.1;
+        multiplier = 0.2;
       }
       
       const adaptRate = this.noiseFloorAdaptRateLoud * multiplier;
@@ -373,25 +373,33 @@ var VADStateMachine = class {
         newState = "silent";
         this.lastSilenceTime = timestamp;
       }
-    } else if (this.currentState === "speech_starting" || this.currentState === "speaking") {
+    } else if (this.currentState === "speech_starting") {
+      if (probability >= stopThreshold) {
+        const speechDuration = timestamp - this.speechStartTime;
+        if (speechDuration >= minSpeechDurationMs) {
+          newState = "speaking";
+        } else {
+          newState = "speech_starting";
+        }
+        this.lastSpeechTime = timestamp;
+      } else {
+        newState = "silent";
+        this.lastSilenceTime = timestamp;
+      }
+    } else if (this.currentState === "speaking") {
       if (probability >= stopThreshold) {
         newState = "speaking";
         this.lastSpeechTime = timestamp;
       } else {
         const timeSinceSpeech = timestamp - this.lastSpeechTime;
-        const speechDuration = timestamp - this.speechStartTime;
         if (timeSinceSpeech < hangoverMs) {
           newState = "speaking";
-        } else if (speechDuration < minSpeechDurationMs) {
-          newState = "silent";
-          this.lastSilenceTime = timestamp;
         } else {
           newState = "speech_ending";
           this.lastSilenceTime = timestamp;
         }
       }
     }
-    if (newState === "speech_starting") newState = "speaking";
     if (newState === "speech_ending") newState = "silent";
     this.currentState = newState;
     return {

package/dist/pipeline/audio-pipeline.mjs

@@ -1,11 +1,11 @@
 import {
   createAudioPipeline
-} from "../chunk-I5AR7XQD.mjs";
-import "../chunk-GFLVGUTU.mjs";
+} from "../chunk-SMZJFNRU.mjs";
+import "../chunk-DYY2MXMU.mjs";
 import "../chunk-OZ7KMC4S.mjs";
-import "../chunk-3I4OQD2L.mjs";
+import "../chunk-XZSFQJW4.mjs";
 import "../chunk-XO6B3D4A.mjs";
-import "../chunk-B36JBXOK.mjs";
+import "../chunk-KEWK2OKV.mjs";
 export {
   createAudioPipeline
 };

package/dist/types.d.mts

@@ -97,12 +97,12 @@ interface AudioProcessingConfig {
             initialNoiseFloor?: number;
             /**
              * Rate at which noise floor adapts to quiet signals (0-1).
-             * Default: 0.005 (slower downward drift)
+             * Default: 0.01
              */
             noiseFloorAdaptRateQuiet?: number;
             /**
              * Rate at which noise floor adapts to loud signals (0-1).
-             * Default: 0.01
+             * Default: 0.05 (faster tracking of rising noise)
              */
             noiseFloorAdaptRateLoud?: number;
             /**

package/dist/types.d.ts

@@ -97,12 +97,12 @@ interface AudioProcessingConfig {
             initialNoiseFloor?: number;
             /**
              * Rate at which noise floor adapts to quiet signals (0-1).
-             * Default: 0.005 (slower downward drift)
+             * Default: 0.01
              */
             noiseFloorAdaptRateQuiet?: number;
             /**
              * Rate at which noise floor adapts to loud signals (0-1).
-             * Default: 0.01
+             * Default: 0.05 (faster tracking of rising noise)
              */
             noiseFloorAdaptRateLoud?: number;
             /**

package/dist/vad/vad-node.js

@@ -27,8 +27,8 @@ var createEnergyVadWorkletCode = (vadConfig) => {
   const energyParams = vadConfig?.energyVad || {};
   const smoothing = energyParams.smoothing ?? 0.95;
   const initialNoiseFloor = energyParams.initialNoiseFloor ?? 1e-3;
-  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 5e-3;
-  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.01;
+  const noiseFloorAdaptRateQuiet = energyParams.noiseFloorAdaptRateQuiet ?? 0.01;
+  const noiseFloorAdaptRateLoud = energyParams.noiseFloorAdaptRateLoud ?? 0.05;
   const minSNR = energyParams.minSNR ?? 10;
   const snrRange = energyParams.snrRange ?? 10;
   const minEnergy = energyParams.minEnergy ?? 1e-3;
@@ -85,9 +85,9 @@ class EnergyVadProcessor extends AudioWorkletProcessor {
       
       let multiplier = 1.0;
       if (this.isSpeaking) {
-        multiplier = 0.01;
+        multiplier = 0.05;
       } else if (snrDb > 20) {
-        multiplier = 0.1;
+        multiplier = 0.2;
       }
       
       const adaptRate = this.noiseFloorAdaptRateLoud * multiplier;

package/dist/vad/vad-node.mjs

@@ -1,6 +1,6 @@
 import {
   EnergyVADPlugin
-} from "../chunk-B36JBXOK.mjs";
+} from "../chunk-KEWK2OKV.mjs";
 export {
   EnergyVADPlugin
 };

package/dist/vad/vad-state.js

@@ -85,25 +85,33 @@ var VADStateMachine = class {
         newState = "silent";
         this.lastSilenceTime = timestamp;
       }
-    } else if (this.currentState === "speech_starting" || this.currentState === "speaking") {
+    } else if (this.currentState === "speech_starting") {
+      if (probability >= stopThreshold) {
+        const speechDuration = timestamp - this.speechStartTime;
+        if (speechDuration >= minSpeechDurationMs) {
+          newState = "speaking";
+        } else {
+          newState = "speech_starting";
+        }
+        this.lastSpeechTime = timestamp;
+      } else {
+        newState = "silent";
+        this.lastSilenceTime = timestamp;
+      }
+    } else if (this.currentState === "speaking") {
       if (probability >= stopThreshold) {
         newState = "speaking";
         this.lastSpeechTime = timestamp;
       } else {
         const timeSinceSpeech = timestamp - this.lastSpeechTime;
-        const speechDuration = timestamp - this.speechStartTime;
         if (timeSinceSpeech < hangoverMs) {
           newState = "speaking";
-        } else if (speechDuration < minSpeechDurationMs) {
-          newState = "silent";
-          this.lastSilenceTime = timestamp;
         } else {
           newState = "speech_ending";
           this.lastSilenceTime = timestamp;
         }
       }
     }
-    if (newState === "speech_starting") newState = "speaking";
     if (newState === "speech_ending") newState = "silent";
     this.currentState = newState;
     return {

package/dist/vad/vad-state.mjs

@@ -1,6 +1,6 @@
 import {
   VADStateMachine
-} from "../chunk-GFLVGUTU.mjs";
+} from "../chunk-DYY2MXMU.mjs";
 export {
   VADStateMachine
 };

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@tensamin/audio",
-  "version": "0.1.11",
+  "version": "0.1.12",
   "main": "dist/index.js",
   "module": "dist/index.mjs",
   "types": "dist/index.d.ts",