From 0baab7db9e249bc32d79ff2fe42be85696a61bc1 Mon Sep 17 00:00:00 2001
From: Bruno Deanoz <mrx.anonymousorg@gmail.com>
Date: Wed, 24 Jun 2026 13:45:44 +0200
Subject: [PATCH] fix Live Call echo: proper AEC pipeline for Android voice
 communication
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The model was talking to itself because the mic picked up speaker output.
Three fixes that together enable Android's hardware echo cancellation:

1. AudioManager.MODE_IN_COMMUNICATION — tells the platform this is a
   voice call so the audio routing + AEC pipeline activates
2. USAGE_VOICE_COMMUNICATION on AudioTrack — links playback to the
   capture path so AEC knows what to subtract from the mic signal
3. Explicit AcousticEchoCanceler on the AudioRecord session as fallback

Also enables speakerphone mode (isSpeakerphoneOn) since MODE_IN_COMMUNICATION
defaults to earpiece routing. Cleanup restores previous audio mode.
---
 .../java/dev/kaizen/app/live/AudioPipeline.kt | 29 +++++++++++++++++--
 1 file changed, 27 insertions(+), 2 deletions(-)

diff --git a/app/src/main/java/dev/kaizen/app/live/AudioPipeline.kt b/app/src/main/java/dev/kaizen/app/live/AudioPipeline.kt
index 064b276..a43cf48 100644
--- a/app/src/main/java/dev/kaizen/app/live/AudioPipeline.kt
+++ b/app/src/main/java/dev/kaizen/app/live/AudioPipeline.kt
@@ -5,16 +5,17 @@ import android.content.Context
 import android.content.pm.PackageManager
 import android.media.AudioAttributes
 import android.media.AudioFormat
+import android.media.AudioManager
 import android.media.AudioRecord
 import android.media.AudioTrack
 import android.media.MediaRecorder
+import android.media.audiofx.AcousticEchoCanceler
 import android.util.Base64
 import android.util.Log
 import androidx.core.content.ContextCompat
 import kotlinx.coroutines.*
 import java.nio.ByteBuffer
 import java.nio.ByteOrder
-import kotlin.math.abs
 import kotlin.math.sqrt
 
 /**
@@ -51,8 +52,11 @@ class AudioPipeline(private val context: Context) {
     var onChunk: ((base64Pcm: String) -> Unit)? = null
     var onAmplitude: ((Float) -> Unit)? = null
 
+    private val audioManager = context.getSystemService(Context.AUDIO_SERVICE) as AudioManager
+    private var previousAudioMode = AudioManager.MODE_NORMAL
     private var recorder: AudioRecord? = null
     private var track: AudioTrack? = null
+    private var aec: AcousticEchoCanceler? = null
     private var captureJob: Job? = null
 
     @Volatile private var ducking = false
@@ -87,6 +91,24 @@ class AudioPipeline(private val context: Context) {
         }
 
         recorder = rec
+
+        previousAudioMode = audioManager.mode
+        audioManager.mode = AudioManager.MODE_IN_COMMUNICATION
+        @Suppress("DEPRECATION")
+        audioManager.isSpeakerphoneOn = true
+
+        if (AcousticEchoCanceler.isAvailable()) {
+            try {
+                aec = AcousticEchoCanceler.create(rec.audioSessionId)
+                aec?.enabled = true
+                Log.i(TAG, "AEC enabled (session=${rec.audioSessionId})")
+            } catch (e: Exception) {
+                Log.w(TAG, "AEC init failed: ${e.message}")
+            }
+        } else {
+            Log.w(TAG, "AEC not available on this device")
+        }
+
         rec.startRecording()
         capturing = true
 
@@ -121,8 +143,11 @@ class AudioPipeline(private val context: Context) {
         try {
             recorder?.stop()
         } catch (_: IllegalStateException) {}
+        aec?.release()
+        aec = null
         recorder?.release()
         recorder = null
+        audioManager.mode = previousAudioMode
     }
 
     // ─── Playback ─────────────────────────────────────────────────────────
@@ -137,7 +162,7 @@ class AudioPipeline(private val context: Context) {
         val t = AudioTrack.Builder()
             .setAudioAttributes(
                 AudioAttributes.Builder()
-                    .setUsage(AudioAttributes.USAGE_ASSISTANT)
+                    .setUsage(AudioAttributes.USAGE_VOICE_COMMUNICATION)
                     .setContentType(AudioAttributes.CONTENT_TYPE_SPEECH)
                     .build()
             )