@voltkit-rtc/react-native 0.1.5 → 0.1.7

package/CHANGELOG.md

@@ -1,5 +1,40 @@
 # Changelog
 
+## 0.1.7
+
+2026-11-09. The echo canceller tuned for loudspeakers. Same peer dependencies as 0.1.6. Build the app again after upgrading:
+the setting is read from the manifest when the app starts, and the version check reads this package's version.
+
+- **WebRTC's canceller tuned for phones whose loudspeaker leaks into their own microphone** (`android.echoSuppression`
+  in the Expo plugin, meta-data `in.voltkit.rtc.ECHO_SUPPRESSION`: `"auto"`, the default, applies the tuning on phones whose
+  own canceller is Android's software one or absent, read from Android's effect descriptors, the gate Agora's SDK applies to
+  the phone's canceller; `"strong"` applies it everywhere; `"webrtc"` is 0.1.6's behaviour everywhere).
+  Measured on an OPPO CPH2477 at full loudspeaker volume, whose speaker compresses the sound and whose microphone clips: with
+  no processing the far voice came back 15 dB louder than it was sent, with the phone's own canceller 10 dB louder, with 0.1.6's
+  canceller 6 dB louder, and with this tuning 3 dB quieter (measured through the Flutter SDK; this SDK runs the same
+  AEC3 and the same tuning). Why it had let the echo through: AEC3 estimates the echo from what it plays with a path gain of 1
+  until its filter converges; on this phone the real gain was about +15 dB, so the residual looked like near-end speech and the
+  canceller went transparent. The tuning (WebRTC's `WebRTC-Aec3SuppressorTuningOverride` field trial, `Aec3Tuning.kt`) lowers
+  the suppressor's echo-to-near-end thresholds and makes its dominant-near-end detector insensitive. A phone whose canceller
+  coped before (a POCO X3 Pro) stays as good. The price: someone speaking into such a phone while the room is loud is cut more
+  often than before. The init provider adds the trial to `WebRTCModuleOptions.fieldTrials` (an app's own trials stay); it does
+  nothing while the phone's canceller is in use (`hardwareEchoCancellation: true`). Measurement:
+  `docs/implementation-progress/ECHO_LAB_2026-11-09.md`. Tests: `RtcConfigTest`, `Aec3TuningTest`, the plugin's.
+
+## 0.1.6
+
+2026-11-06. WebRTC's own echo canceller by default on Android. Same peer dependencies as 0.1.5. Build the app again after
+upgrading: the setting is read from the manifest when the app starts, and the version check reads this package's version.
+
+- **WebRTC's echo canceller (AEC3) and noise suppressor by default** (`android.hardwareEchoCancellation` now defaults to
+  `false`; meta-data `in.voltkit.rtc.HW_AEC`). On Android, while WebRTC uses the phone's own canceller it switches its own
+  off, so the result depended on the phone: on an Oppo on the loudspeaker the person on the other phone heard themselves
+  with distortion, in this SDK and the Flutter one (a developer's finding, 2026-11-06). Agora and ZEGO tell developers to
+  use the SDK's canceller for such phones; theirs keep one, WebRTC does not. AEC3 is the canceller browsers use, and the
+  one this SDK's `react-native-webrtc` build carries. To keep the phone's, as before 0.1.6, set
+  `android: { hardwareEchoCancellation: true }` in the plugin options (or `in.voltkit.rtc.HW_AEC` to `true`). iOS is
+  unchanged: it uses the system's voice processing.
+
 ## 0.1.5
 
 2026-11-05. A developer's findings on 0.1.4, and the round of 2026-11-05 (decisions D2 and D10). Peer dependencies

package/README.md

@@ -134,7 +134,7 @@ you edit no native file:
 | `android.callServiceCamera` | `false` | Adds the `camera` type and `FOREGROUND_SERVICE_CAMERA`, for an app that keeps the camera on in the background (`background={{ pauseCamera: false }}`) |
 | `android.screenShare` | `false` | Turns on react-native-webrtc's screen-capture service for [screen sharing](#screen-sharing), with `FOREGROUND_SERVICE_MEDIA_PROJECTION` and the `mediaProjection` type on the call service |
 | `android.audioMode` | `"communication"` | `"media"` for an app whose users only listen: see [Audio routing](#audio-routing) |
-| `android.hardwareEchoCancellation` | `true` | The phone's own echo canceller and noise suppressor, where it has them |
+| `android.hardwareEchoCancellation` | `false` | `false` (the default from 0.1.6): WebRTC's own echo canceller and noise suppressor (AEC3). `true`: the phone's own, where it has them, as before 0.1.6. See [Echo on Android](#echo-on-android) |
 | `ios.backgroundAudio` | `true` | Adds `audio` to `UIBackgroundModes`, so a call keeps its sound in the background |
 | `ios.voip` | `false` | Adds `voip` to `UIBackgroundModes`, for an app that integrates CallKit itself |
 
@@ -741,11 +741,12 @@ export function OutputPicker() {
   tried again until the user picks it, a device is added, or the call ends.
 - **No Bluetooth permission** is needed: the routing goes through Android's
   `AudioManager`.
-- **Echo on one phone model.** The phone's own echo canceller and noise
-  suppressor are used where it has them. If one model echoes, turn them off with
-  the Expo plugin's `android.hardwareEchoCancellation: false` (or the
-  `in.voltkit.rtc.HW_AEC` meta-data, below), and WebRTC's own echo canceller
-  does the work.
+- **Echo cancellation.** From 0.1.6, WebRTC's own echo canceller and noise
+  suppressor (AEC3, the one browsers use), not the phone's built-in ones, which
+  some phones cannot run well on the loudspeaker.
+  `android: { hardwareEchoCancellation: true }` in the Expo plugin's options (or
+  the `in.voltkit.rtc.HW_AEC` meta-data, below) brings the phone's back. See
+  [Echo on Android](#echo-on-android).
 - **Apps whose users only listen** can set the audio mode to `media` (the
   plugin's `android.audioMode`, or the `in.voltkit.rtc.AUDIO_MODE` meta-data):
   sound plays at the media volume and on Bluetooth's high-quality profile, with
@@ -765,8 +766,8 @@ In a React Native app without Expo, the audio settings are meta-data in
 <application>
   <!-- communication (calls and audio rooms) or media (listen-only apps) -->
   <meta-data android:name="in.voltkit.rtc.AUDIO_MODE" android:value="communication" />
-  <!-- the phone's own echo canceller and noise suppressor, where it has them -->
-  <meta-data android:name="in.voltkit.rtc.HW_AEC" android:value="true" />
+  <!-- false: WebRTC's own echo canceller and noise suppressor (from 0.1.6); true: the phone's own, where it has them -->
+  <meta-data android:name="in.voltkit.rtc.HW_AEC" android:value="false" />
   <!-- react-native-webrtc's screen-capture service: see Screen sharing -->
   <meta-data android:name="in.voltkit.rtc.SCREEN_SHARE" android:value="false" />
 </application>
@@ -1433,12 +1434,62 @@ The emulator has only the loudspeaker, so `useAudioOutput().outputs` lists one
 output and `setSpeakerphoneOn(false)` has nowhere else to go. Test the earpiece
 and headsets on a phone.
 
-### Echo on one phone model
+### Echo on Android
 
-Turn off the phone's own echo canceller with the Expo plugin's
-`android.hardwareEchoCancellation: false` (or `in.voltkit.rtc.HW_AEC` set to
-`false`) and build again; WebRTC's own canceller takes over. A headset or the
-earpiece removes echo entirely.
+From 0.1.6 the SDK uses WebRTC's own echo canceller and noise suppressor (AEC3,
+the one browsers use) on Android, not the phone's built-in ones. While WebRTC
+uses a phone's canceller it switches its own off, and some phones' cancellers
+do not cope with the loudspeaker: the person on the other phone hears their own
+voice back, even in another room. An app on 0.1.5 or earlier that sees this
+should upgrade and build again.
+
+To use the phone's own canceller instead, as before 0.1.6, set it in the Expo
+plugin's options, in `app.json` or `app.config.js`:
+
+```json
+{
+  "expo": {
+    "plugins": [
+      ["@voltkit-rtc/react-native", { "android": { "hardwareEchoCancellation": true } }]
+    ]
+  }
+}
+```
+
+Then make a new build of the app (`npx expo run:android`, or a new EAS build):
+the setting is read when the app starts, so reloading the JavaScript does not
+apply it. In a React Native app without Expo, set the meta-data in
+`AndroidManifest.xml` instead, and build again:
+
+```xml
+<meta-data android:name="in.voltkit.rtc.HW_AEC" android:value="true" />
+```
+
+**Tuned for loudspeakers (0.1.7).** WebRTC's canceller, as tuned for laptops,
+still let the echo through on a phone whose loudspeaker leaks into its own
+microphone: measured on an OPPO CPH2477 at full volume, the far voice came back
+6 dB louder than it was sent, because the canceller underestimated the echo some
+thirty times and took the rest for someone speaking. From 0.1.7 the SDK hands
+WebRTC a tuning for such phones: on that phone the echo comes back 3 dB
+quieter than it was sent. The default, `android.echoSuppression: "auto"`, applies
+it only on phones whose own echo canceller is Android's software one or absent,
+which are the phones that echo (Android's effect descriptors say which; the gate
+Agora's SDK applies to the phone's canceller); phones with a vendor's canceller
+keep WebRTC's own tuning and full duplex. The price on the tuned phones is that
+someone speaking into them while the room is loud is cut more often. `"strong"`
+applies the tuning everywhere; `"webrtc"` keeps WebRTC's own tuning everywhere,
+as 0.1.6 had. Set it in the plugin's options
+(`{ "android": { "echoSuppression": "webrtc" } }`), or the meta-data
+`in.voltkit.rtc.ECHO_SUPPRESSION` without Expo, and build again. The tuning is WebRTC's `WebRTC-Aec3SuppressorTuningOverride` field
+trial, handed to react-native-webrtc through `WebRTCModuleOptions.fieldTrials`
+(field trials your app set itself stay); it does nothing while the phone's own
+canceller is in use. The measurement: docs/implementation-progress/ECHO_LAB_2026-11-09.md.
+
+An Expo app uses the plugin option, not a hand edit: `expo prebuild` writes the
+manifest again, and the edit is lost. Test either choice on the phone that
+echoed and on another before you ship it. A headset or the earpiece on the
+phone that causes the echo also stops it. Where echo comes from, including two
+phones in one room: https://voltkit.in/docs/troubleshooting#echo.
 
 ### Google Play rejects the app for 16 KB page sizes
 

package/android/src/main/java/com/voltkit/rtc/reactnative/Aec3Tuning.kt

@@ -0,0 +1,60 @@
+package com.voltkit.rtc.reactnative
+
+/**
+ * The WebRTC field trial behind the SDK's echo suppression (0.1.7): AEC3's suppressor thresholds and its dominant-
+ * near-end detector, in WebRTC's override format (`WebRTC-Aec3SuppressorTuningOverride`). AEC3 estimates the echo from
+ * what it plays with an assumed path gain of 1 until its filter converges; on the measured phone (an OPPO CPH2477,
+ * whose loudspeaker compresses the sound and whose microphone clips) the real gain was about +15 dB, so the echo was
+ * underestimated some thirty times, the residual looked like near-end speech and the canceller went transparent: the
+ * people on the other phones heard themselves, louder than they had spoken. These thresholds are WebRTC's defaults
+ * divided by that factor, and the near-end detector made as much less sensitive. From the lab of 2026-11-09
+ * (docs/implementation-progress/ECHO_LAB_2026-11-09.md). The same string is used by the Flutter and Kotlin SDKs.
+ * react-native-webrtc hands `WebRTCModuleOptions.fieldTrials` to WebRTC when it starts.
+ */
+internal object Aec3Tuning {
+    const val LOUDSPEAKER: String = "WebRTC-Aec3SuppressorTuningOverride/" +
+        "normal_tuning_mask_lf_enr_transparent:0.01,normal_tuning_mask_lf_enr_suppress:0.013," +
+        "normal_tuning_mask_hf_enr_transparent:0.0023,normal_tuning_mask_hf_enr_suppress:0.0033," +
+        "nearend_tuning_mask_lf_enr_transparent:0.03,nearend_tuning_mask_lf_enr_suppress:0.04," +
+        "nearend_tuning_mask_hf_enr_transparent:0.003,nearend_tuning_mask_hf_enr_suppress:0.01," +
+        "dominant_nearend_detection_enr_threshold:0.008,dominant_nearend_detection_hold_duration:25/"
+
+    /**
+     * The phone's own echo canceller as Android describes it (`AudioEffect.queryEffects()`): a vendor's DSP canceller
+     * (Qualcomm Fluence, Samsung, ...) or the AOSP software one (`libaudiopreprocessing`, the old WebRTC mobile canceller
+     * run by the audio server), which the phones that echo on the loudspeaker ship (the OPPO CPH2477 of the 2026-11-09
+     * lab: MediaTek's build, uuid 9f35ed76-...; the stock one is bb392ec0-...). Agora's SDK gates the phone's canceller
+     * on the same descriptor.
+     */
+    data class PhoneCanceller(val available: Boolean, val implementor: String?, val uuid: String?) {
+        val aosp: Boolean
+            get() = (implementor?.contains("Android Open Source") == true) ||
+                uuid.equals("bb392ec0-8d4d-11e0-a896-0002a5d5c51b", ignoreCase = true) ||
+                uuid.equals("9f35ed76-0b66-4330-8f79-e39ca266dc7c", ignoreCase = true)
+
+        fun describe(): String = if (!available) "none" else "${implementor ?: "unknown"} ($uuid)"
+    }
+
+    fun phoneCanceller(): PhoneCanceller {
+        val available = runCatching { android.media.audiofx.AcousticEchoCanceler.isAvailable() }.getOrDefault(false)
+        val d = runCatching {
+            android.media.audiofx.AudioEffect.queryEffects().firstOrNull { it.type == android.media.audiofx.AudioEffect.EFFECT_TYPE_AEC }
+        }.getOrNull()
+        return PhoneCanceller(available, d?.implementor, d?.uuid?.toString())
+    }
+
+    /** Whether the loudspeaker tuning goes in: `strong` always, `webrtc` never, `auto` on the phones that echo (the AOSP software canceller, or none). */
+    fun tunes(suppression: EchoSuppression, phone: PhoneCanceller): Boolean = when (suppression) {
+        EchoSuppression.strong -> true
+        EchoSuppression.webrtc -> false
+        EchoSuppression.auto -> !phone.available || phone.aosp
+    }
+
+    /** [LOUDSPEAKER] added to field trials the app already set, once; an app's own AEC3 tuning stays. */
+    fun merge(existing: String?): String {
+        val have = existing?.trim().orEmpty()
+        if (have.contains("WebRTC-Aec3SuppressorTuningOverride/")) return have
+        if (have.isEmpty()) return LOUDSPEAKER
+        return (if (have.endsWith("/")) have else "$have/") + LOUDSPEAKER
+    }
+}

package/android/src/main/java/com/voltkit/rtc/reactnative/RtcConfig.kt

@@ -20,22 +20,53 @@ internal enum class AudioMode {
     media,
 }
 
+/**
+ * How hard WebRTC's own echo canceller (AEC3) suppresses what is left of the echo (meta-data in.voltkit.rtc.ECHO_SUPPRESSION;
+ * 0.1.7). In force only while WebRTC's canceller is in use (HW_AEC false, the default): with the phone's own, WebRTC's is off.
+ */
+@Suppress("EnumEntryName")
+internal enum class EchoSuppression {
+    /**
+     * The default: [strong] on phones whose own echo canceller is Android's software one or absent (the phones that echo
+     * on the loudspeaker), [webrtc] on phones with a vendor's DSP canceller (which do not echo and keep full duplex).
+     */
+    auto,
+
+    /**
+     * AEC3 tuned for phones whose loudspeaker leaks into their own microphone (an OPPO CPH2477 returned the far voice
+     * louder than it arrived, and WebRTC's canceller as tuned for laptops mistook that echo for someone speaking and let
+     * it through). Someone speaking into the phone while the room is loud is cut more often than with [webrtc].
+     */
+    strong,
+
+    /** WebRTC's own tuning, as 0.1.6 used. */
+    webrtc,
+}
+
 /**
  * The process-wide settings the init provider reads from the app's manifest before react-native-webrtc starts
  * (plan §5.3), as the Expo config plugin writes them or a hand-written manifest gives them:
  *
  *     <meta-data android:name="in.voltkit.rtc.AUDIO_MODE" android:value="communication" />  communication | media
- *     <meta-data android:name="in.voltkit.rtc.HW_AEC" android:value="true" />               the handset's own echo
- *                                                                                            canceller and noise suppressor
+ *     <meta-data android:name="in.voltkit.rtc.HW_AEC" android:value="false" />              the handset's own echo
+ *                                                                                            canceller and noise suppressor;
+ *                                                                                            false (the default from 0.1.6):
+ *                                                                                            WebRTC's own (AEC3)
  *     <meta-data android:name="in.voltkit.rtc.SCREEN_SHARE" android:value="false" />        react-native-webrtc's
  *                                                                                            screen-share service
+ *     <meta-data android:name="in.voltkit.rtc.ECHO_SUPPRESSION" android:value="auto" />     auto (the default from 0.1.7:
+ *                                                                                            AEC3 tuned for loudspeakers
+ *                                                                                            on phones whose own canceller
+ *                                                                                            is Android's software one) |
+ *                                                                                            strong | webrtc
  *
  * Parsing only, so it can be tested off a device. A value that cannot be read keeps its default and is reported.
  */
 internal data class RtcConfig(
     val audioMode: AudioMode = AudioMode.communication,
-    val hardwareEchoCancellation: Boolean = true,
+    val hardwareEchoCancellation: Boolean = false,
     val screenShare: Boolean = false,
+    val echoSuppression: EchoSuppression = EchoSuppression.auto,
     /** What could not be read, with the default used instead; the init provider logs it. */
     val problems: List<String> = emptyList(),
 ) {
@@ -43,6 +74,7 @@ internal data class RtcConfig(
         const val AUDIO_MODE = "in.voltkit.rtc.AUDIO_MODE"
         const val HW_AEC = "in.voltkit.rtc.HW_AEC"
         const val SCREEN_SHARE = "in.voltkit.rtc.SCREEN_SHARE"
+        const val ECHO_SUPPRESSION = "in.voltkit.rtc.ECHO_SUPPRESSION"
 
         /** From the application's meta-data: [meta] gives one key's value (a Bundle's `get`), or null. */
         fun from(meta: (String) -> Any?): RtcConfig {
@@ -53,10 +85,17 @@ internal data class RtcConfig(
                     ?: AudioMode.communication.also { problems += "$AUDIO_MODE: '$raw' is not communication or media; using communication" }
                 else -> AudioMode.communication.also { problems += "$AUDIO_MODE: '$raw' is not communication or media; using communication" }
             }
+            val suppression = when (val raw = meta(ECHO_SUPPRESSION)) {
+                null -> EchoSuppression.auto
+                is String -> EchoSuppression.values().firstOrNull { it.name == raw.trim().lowercase() }
+                    ?: EchoSuppression.auto.also { problems += "$ECHO_SUPPRESSION: '$raw' is not auto, strong or webrtc; using auto" }
+                else -> EchoSuppression.auto.also { problems += "$ECHO_SUPPRESSION: '$raw' is not auto, strong or webrtc; using auto" }
+            }
             return RtcConfig(
                 audioMode = mode,
-                hardwareEchoCancellation = flag(meta(HW_AEC), HW_AEC, default = true, problems),
+                hardwareEchoCancellation = flag(meta(HW_AEC), HW_AEC, default = false, problems),
                 screenShare = flag(meta(SCREEN_SHARE), SCREEN_SHARE, default = false, problems),
+                echoSuppression = suppression,
                 problems = problems,
             )
         }

package/android/src/main/java/com/voltkit/rtc/reactnative/VoltkitRtcInitProvider.kt

@@ -49,6 +49,15 @@ class VoltkitRtcInitProvider : ContentProvider() {
         config.problems.forEach { Log.w(TAG, "manifest meta-data: $it") }
         val options = WebRTCModuleOptions.getInstance()
         if (config.screenShare) options.enableMediaProjectionService = true
+        // WebRTC's canceller tuned for loudspeakers (0.1.7): a field trial react-native-webrtc hands WebRTC when it
+        // starts, read when the canceller is built. Nothing to tune while the phone's own canceller is in use (WebRTC's
+        // is then off). Field trials the app set itself stay; this one is added.
+        if (!config.hardwareEchoCancellation) {
+            val phone = Aec3Tuning.phoneCanceller()
+            val tune = Aec3Tuning.tunes(config.echoSuppression, phone)
+            if (tune) options.fieldTrials = Aec3Tuning.merge(options.fieldTrials)
+            Log.i(TAG, "echo suppression ${config.echoSuppression.name}: the phone's canceller is ${phone.describe()}; WebRTC's canceller ${if (tune) "tuned for loudspeakers" else "with its own tuning"}")
+        }
         if (options.audioDeviceModule != null) {
             Log.i(TAG, "react-native-webrtc already has an audio device module; keeping it")
             return

package/lib/commonjs/native.js

@@ -21,7 +21,7 @@ function _interopRequireDefault(e) { return e && e.__esModule ? e : { default: e
  */
 
 /** The version of this package: what getDiagnostics() reports, and what the native side was built from. */
-const SDK_VERSION = exports.SDK_VERSION = '0.1.5';
+const SDK_VERSION = exports.SDK_VERSION = '0.1.7';
 
 /** The native module, or null where there is none (iOS, Expo Go, an app not rebuilt). */
 function nativeModule() {

package/lib/module/native.js

@@ -9,7 +9,7 @@ import { Platform } from 'react-native';
 import NativeVoltkitRtc from "./specs/NativeVoltkitRtc.js";
 import { iosUnsupported, logError, nativeMissing } from "./errors.js";
 /** The version of this package: what getDiagnostics() reports, and what the native side was built from. */
-export const SDK_VERSION = '0.1.5';
+export const SDK_VERSION = '0.1.7';
 
 /** The native module, or null where there is none (iOS, Expo Go, an app not rebuilt). */
 export function nativeModule() {

package/lib/typescript/commonjs/native.d.ts

@@ -2,7 +2,7 @@ import type { EventSubscription } from 'react-native';
 import { type AudioDevice, type Constants, type Spec } from './specs/NativeVoltkitRtc';
 import type { MediaDevice } from '@voltkit-rtc/client';
 /** The version of this package: what getDiagnostics() reports, and what the native side was built from. */
-export declare const SDK_VERSION = "0.1.5";
+export declare const SDK_VERSION = "0.1.7";
 /** The native module, or null where there is none (iOS, Expo Go, an app not rebuilt). */
 export declare function nativeModule(): Spec | null;
 /** The module's constants, read once: the audio mode, screen sharing, the call service as the manifest declares them. */

package/lib/typescript/module/native.d.ts

@@ -2,7 +2,7 @@ import type { EventSubscription } from 'react-native';
 import { type AudioDevice, type Constants, type Spec } from './specs/NativeVoltkitRtc.js';
 import type { MediaDevice } from '@voltkit-rtc/client';
 /** The version of this package: what getDiagnostics() reports, and what the native side was built from. */
-export declare const SDK_VERSION = "0.1.5";
+export declare const SDK_VERSION = "0.1.7";
 /** The native module, or null where there is none (iOS, Expo Go, an app not rebuilt). */
 export declare function nativeModule(): Spec | null;
 /** The module's constants, read once: the audio mode, screen sharing, the call service as the manifest declares them. */

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@voltkit-rtc/react-native",
-  "version": "0.1.5",
+  "version": "0.1.7",
   "description": "React Native SDK for VoltKit RTC: calls, audio rooms and live video on Android (iOS in preview), with the hooks of @voltkit-rtc/react, background calls, audio routing and an Expo config plugin.",
   "keywords": [
     "voltkit",

package/plugin/build/android.d.ts

@@ -6,6 +6,7 @@ export declare const CALL_SERVICE_CLASS = "com.voltkit.rtc.reactnative.VoltkitCa
 export declare const META_AUDIO_MODE = "in.voltkit.rtc.AUDIO_MODE";
 export declare const META_HW_AEC = "in.voltkit.rtc.HW_AEC";
 export declare const META_SCREEN_SHARE = "in.voltkit.rtc.SCREEN_SHARE";
+export declare const META_ECHO_SUPPRESSION = "in.voltkit.rtc.ECHO_SUPPRESSION";
 /** A service type by its manifest name (CallServiceTypes.kt), with what it needs on Android 14 and later. */
 type ServiceType = 'microphone' | 'mediaPlayback' | 'camera' | 'mediaProjection';
 /** What the plugin writes for these options, given the permissions the app blocks. */

package/plugin/build/android.js

@@ -1,6 +1,6 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.withVoltkitAndroid = exports.META_SCREEN_SHARE = exports.META_HW_AEC = exports.META_AUDIO_MODE = exports.CALL_SERVICE_CLASS = void 0;
+exports.withVoltkitAndroid = exports.META_ECHO_SUPPRESSION = exports.META_SCREEN_SHARE = exports.META_HW_AEC = exports.META_AUDIO_MODE = exports.CALL_SERVICE_CLASS = void 0;
 exports.blockedPermissions = blockedPermissions;
 exports.planAndroid = planAndroid;
 exports.setVoltkitManifest = setVoltkitManifest;
@@ -25,8 +25,9 @@ exports.setVoltkitManifest = setVoltkitManifest;
  *     <uses-feature android:name="android.hardware.microphone" android:required="false" />        with RECORD_AUDIO
  *     <application>
  *       <meta-data android:name="in.voltkit.rtc.AUDIO_MODE" android:value="communication" />   RtcConfig.kt
- *       <meta-data android:name="in.voltkit.rtc.HW_AEC" android:value="true" />
+ *       <meta-data android:name="in.voltkit.rtc.HW_AEC" android:value="false" />
  *       <meta-data android:name="in.voltkit.rtc.SCREEN_SHARE" android:value="false" />
+ *       <meta-data android:name="in.voltkit.rtc.ECHO_SUPPRESSION" android:value="auto" />   auto | strong | webrtc (0.1.7)
  *       <service android:name="com.voltkit.rtc.reactnative.VoltkitCallService" android:exported="false"
  *                android:foregroundServiceType="microphone|mediaPlayback" />                    VoltkitCallService.kt
  *     </application>
@@ -85,6 +86,7 @@ exports.CALL_SERVICE_CLASS = 'com.voltkit.rtc.reactnative.VoltkitCallService';
 exports.META_AUDIO_MODE = 'in.voltkit.rtc.AUDIO_MODE';
 exports.META_HW_AEC = 'in.voltkit.rtc.HW_AEC';
 exports.META_SCREEN_SHARE = 'in.voltkit.rtc.SCREEN_SHARE';
+exports.META_ECHO_SUPPRESSION = 'in.voltkit.rtc.ECHO_SUPPRESSION';
 const permission = (name) => `android.permission.${name}`;
 const CAMERA = permission('CAMERA');
 const RECORD_AUDIO = permission('RECORD_AUDIO');
@@ -246,6 +248,7 @@ function setVoltkitManifest(manifest, options, configBlocked = []) {
     config_plugins_1.AndroidConfig.Manifest.addMetaDataItemToMainApplication(application, exports.META_AUDIO_MODE, options.audioMode);
     config_plugins_1.AndroidConfig.Manifest.addMetaDataItemToMainApplication(application, exports.META_HW_AEC, String(options.hardwareEchoCancellation));
     config_plugins_1.AndroidConfig.Manifest.addMetaDataItemToMainApplication(application, exports.META_SCREEN_SHARE, String(plan.screenShare));
+    config_plugins_1.AndroidConfig.Manifest.addMetaDataItemToMainApplication(application, exports.META_ECHO_SUPPRESSION, options.echoSuppression);
     setCallService(application, plan.serviceTypes);
     if (plan.serviceDropped) {
         config_plugins_1.WarningAggregator.addWarningAndroid(`${options_1.PLUGIN_NAME} android.callService`, `VoltkitCallService is not declared: ${plan.serviceDropped}. Calls will drop a few seconds after the app leaves ` +

package/plugin/build/index.d.ts

@@ -2,8 +2,8 @@ import type { ConfigPlugin } from 'expo/config-plugins';
 import type { VoltkitRtcPluginProps } from './options';
 declare const _default: ConfigPlugin<void | VoltkitRtcPluginProps | null>;
 export default _default;
-export { CALL_SERVICE_CLASS, META_AUDIO_MODE, META_HW_AEC, META_SCREEN_SHARE, blockedPermissions, planAndroid, setVoltkitManifest, withVoltkitAndroid, } from './android';
+export { CALL_SERVICE_CLASS, META_AUDIO_MODE, META_ECHO_SUPPRESSION, META_HW_AEC, META_SCREEN_SHARE, blockedPermissions, planAndroid, setVoltkitManifest, withVoltkitAndroid, } from './android';
 export type { AndroidPlan } from './android';
 export { setVoltkitInfoPlist, withVoltkitIos } from './ios';
 export { DEFAULT_CAMERA_PERMISSION, DEFAULT_MICROPHONE_PERMISSION, PLUGIN_NAME, resolveOptions } from './options';
-export type { AndroidOptions, AudioMode, IosOptions, PluginOptions, VoltkitRtcAndroidProps, VoltkitRtcIosProps, VoltkitRtcPluginProps, } from './options';
+export type { AndroidOptions, AudioMode, IosOptions, PluginOptions, VoltkitRtcAndroidProps, VoltkitRtcIosProps, VoltkitRtcPluginProps, EchoSuppression, } from './options';

package/plugin/build/index.js

@@ -1,6 +1,6 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.resolveOptions = exports.PLUGIN_NAME = exports.DEFAULT_MICROPHONE_PERMISSION = exports.DEFAULT_CAMERA_PERMISSION = exports.withVoltkitIos = exports.setVoltkitInfoPlist = exports.withVoltkitAndroid = exports.setVoltkitManifest = exports.planAndroid = exports.blockedPermissions = exports.META_SCREEN_SHARE = exports.META_HW_AEC = exports.META_AUDIO_MODE = exports.CALL_SERVICE_CLASS = void 0;
+exports.resolveOptions = exports.PLUGIN_NAME = exports.DEFAULT_MICROPHONE_PERMISSION = exports.DEFAULT_CAMERA_PERMISSION = exports.withVoltkitIos = exports.setVoltkitInfoPlist = exports.withVoltkitAndroid = exports.setVoltkitManifest = exports.planAndroid = exports.blockedPermissions = exports.META_SCREEN_SHARE = exports.META_HW_AEC = exports.META_ECHO_SUPPRESSION = exports.META_AUDIO_MODE = exports.CALL_SERVICE_CLASS = void 0;
 /**
  * The Expo config plugin of @voltkit-rtc/react-native (plan §5.5). An app lists it in app.json:
  *
@@ -38,6 +38,7 @@ exports.default = (0, config_plugins_1.createRunOncePlugin)(withVoltkitRtc, opti
 var android_2 = require("./android");
 Object.defineProperty(exports, "CALL_SERVICE_CLASS", { enumerable: true, get: function () { return android_2.CALL_SERVICE_CLASS; } });
 Object.defineProperty(exports, "META_AUDIO_MODE", { enumerable: true, get: function () { return android_2.META_AUDIO_MODE; } });
+Object.defineProperty(exports, "META_ECHO_SUPPRESSION", { enumerable: true, get: function () { return android_2.META_ECHO_SUPPRESSION; } });
 Object.defineProperty(exports, "META_HW_AEC", { enumerable: true, get: function () { return android_2.META_HW_AEC; } });
 Object.defineProperty(exports, "META_SCREEN_SHARE", { enumerable: true, get: function () { return android_2.META_SCREEN_SHARE; } });
 Object.defineProperty(exports, "blockedPermissions", { enumerable: true, get: function () { return android_2.blockedPermissions; } });

package/plugin/build/options.d.ts

@@ -17,6 +17,8 @@
 export declare const PLUGIN_NAME = "@voltkit-rtc/react-native";
 /** How the app plays call audio (meta-data in.voltkit.rtc.AUDIO_MODE; RtcConfig.kt). */
 export type AudioMode = 'communication' | 'media';
+/** How hard WebRTC's echo canceller suppresses what is left of the echo (meta-data in.voltkit.rtc.ECHO_SUPPRESSION; RtcConfig.kt). */
+export type EchoSuppression = 'auto' | 'strong' | 'webrtc';
 export interface VoltkitRtcAndroidProps {
     /**
      * "communication" (default): call audio, echo cancellation, the earpiece, the loudspeaker and headsets. "media", for
@@ -51,8 +53,22 @@ export interface VoltkitRtcAndroidProps {
      * android.blockedPermissions blocks what it needs.
      */
     screenShare?: boolean;
-    /** The phone's own echo canceller and noise suppressor, where it has them (meta-data in.voltkit.rtc.HW_AEC; default true). */
+    /**
+     * The phone's own echo canceller and noise suppressor, where it has them (meta-data in.voltkit.rtc.HW_AEC). Default false
+     * from 0.1.6: WebRTC's own (AEC3, the one browsers use), because WebRTC switches its own off while it uses the phone's and
+     * some phones' cancellers do not cope with the loudspeaker. True for the phone's, as before 0.1.6.
+     */
     hardwareEchoCancellation?: boolean;
+    /**
+     * How hard WebRTC's echo canceller suppresses what is left of the echo (meta-data in.voltkit.rtc.ECHO_SUPPRESSION).
+     * "auto" (default from 0.1.7): "strong" on phones whose own canceller is Android's software one or absent, which are the
+     * phones that echo on the loudspeaker; "webrtc" on phones with a vendor's canceller. "strong": AEC3 tuned for phones
+     * whose loudspeaker leaks into their own microphone, everywhere (measured on one whose untuned canceller returned the
+     * far voice louder than it arrived; the price is that someone speaking into the phone while the room is loud is cut
+     * more often). "webrtc": WebRTC's own tuning everywhere, as 0.1.6 used. In force only with WebRTC's canceller
+     * (hardwareEchoCancellation false).
+     */
+    echoSuppression?: EchoSuppression;
 }
 export interface VoltkitRtcIosProps {
     /** Adds `audio` to UIBackgroundModes, so a call keeps its sound in the background (default true). */
@@ -75,6 +91,7 @@ export interface AndroidOptions {
     callServiceCamera: boolean;
     screenShare: boolean;
     hardwareEchoCancellation: boolean;
+    echoSuppression: EchoSuppression;
 }
 export interface IosOptions {
     /** Absent: keep the Info.plist's own text, or write the default. */

package/plugin/build/options.js

@@ -22,13 +22,15 @@ exports.PLUGIN_NAME = '@voltkit-rtc/react-native';
 exports.DEFAULT_CAMERA_PERMISSION = 'Allow $(PRODUCT_NAME) to use your camera in calls.';
 exports.DEFAULT_MICROPHONE_PERMISSION = 'Allow $(PRODUCT_NAME) to use your microphone in calls.';
 const AUDIO_MODES = ['communication', 'media'];
+const ECHO_SUPPRESSIONS = ['auto', 'strong', 'webrtc'];
 const ANDROID_DEFAULTS = {
     audioMode: 'communication',
     camera: true,
     callService: true,
     callServiceCamera: false,
     screenShare: false,
-    hardwareEchoCancellation: true,
+    hardwareEchoCancellation: false,
+    echoSuppression: 'auto',
 };
 const IOS_DEFAULTS = { backgroundAudio: true, voip: false };
 const TOP_KEYS = ['cameraPermission', 'microphonePermission', 'android', 'ios'];
@@ -112,8 +114,17 @@ function resolveOptions(props) {
         }
         audioMode = value;
     }
+    let echoSuppression = ANDROID_DEFAULTS.echoSuppression;
+    if (androidProps.echoSuppression !== undefined) {
+        const value = androidProps.echoSuppression;
+        if (!ECHO_SUPPRESSIONS.includes(value)) {
+            fail(`android.echoSuppression must be ${ECHO_SUPPRESSIONS.map((mode) => `"${mode}"`).join(' or ')}, not ${show(value)}`);
+        }
+        echoSuppression = value;
+    }
     const android = {
         audioMode,
+        echoSuppression,
         camera: bool(androidProps, 'camera', 'android', ANDROID_DEFAULTS.camera),
         callService: bool(androidProps, 'callService', 'android', ANDROID_DEFAULTS.callService),
         callServiceCamera: bool(androidProps, 'callServiceCamera', 'android', ANDROID_DEFAULTS.callServiceCamera),

package/src/native.ts

@@ -10,7 +10,7 @@ import { iosUnsupported, logError, nativeMissing } from './errors';
 import type { AudioDeviceType, DeviceKind, MediaDevice } from '@voltkit-rtc/client';
 
 /** The version of this package: what getDiagnostics() reports, and what the native side was built from. */
-export const SDK_VERSION = '0.1.5';
+export const SDK_VERSION = '0.1.7';
 
 /** The native module, or null where there is none (iOS, Expo Go, an app not rebuilt). */
 export function nativeModule(): Spec | null {