Include Arduino header when using Arduino

Co-authored-by: clydebarrow <2366188+clydebarrow@users.noreply.github.com>

CODEOWNERS

@@ -278,7 +278,7 @@ esphome/components/mdns/* @esphome/core
 esphome/components/media_player/* @jesserockz
 esphome/components/micro_wake_word/* @jesserockz @kahrendt
 esphome/components/micronova/* @jorre05
-esphome/components/microphone/* @jesserockz
+esphome/components/microphone/* @jesserockz @kahrendt
 esphome/components/mics_4514/* @jesserockz
 esphome/components/midea/* @dudanov
 esphome/components/midea_ir/* @dudanov
@@ -319,6 +319,7 @@ esphome/components/online_image/* @clydebarrow @guillempages
 esphome/components/opentherm/* @olegtarasov
 esphome/components/ota/* @esphome/core
 esphome/components/output/* @esphome/core
+esphome/components/packet_transport/* @clydebarrow
 esphome/components/pca6416a/* @Mat931
 esphome/components/pca9554/* @clydebarrow @hwstar
 esphome/components/pcf85063/* @brogon
@@ -328,6 +329,7 @@ esphome/components/pipsolar/* @andreashergert1984
 esphome/components/pm1006/* @habbie
 esphome/components/pm2005/* @andrewjswan
 esphome/components/pmsa003i/* @sjtrny
+esphome/components/pmsx003/* @ximex
 esphome/components/pmwcs3/* @SeByDocKy
 esphome/components/pn532/* @OttoWinter @jesserockz
 esphome/components/pn532_i2c/* @OttoWinter @jesserockz
@@ -427,6 +429,7 @@ esphome/components/sun/* @OttoWinter
 esphome/components/sun_gtil2/* @Mat931
 esphome/components/switch/* @esphome/core
 esphome/components/switch/binary_sensor/* @ssieb
+esphome/components/syslog/* @clydebarrow
 esphome/components/t6615/* @tylermenezes
 esphome/components/tc74/* @sethgirvan
 esphome/components/tca9548a/* @andreashergert1984
@@ -466,6 +469,7 @@ esphome/components/tuya/switch/* @jesserockz
 esphome/components/tuya/text_sensor/* @dentra
 esphome/components/uart/* @esphome/core
 esphome/components/uart/button/* @ssieb
+esphome/components/uart/packet_transport/* @clydebarrow
 esphome/components/udp/* @clydebarrow
 esphome/components/ufire_ec/* @pvizeli
 esphome/components/ufire_ise/* @pvizeli

esphome/components/adc/__init__.py

@@ -47,9 +47,10 @@ SAMPLING_MODES = {
 adc1_channel_t = cg.global_ns.enum("adc1_channel_t")
 adc2_channel_t = cg.global_ns.enum("adc2_channel_t")
 
-# From https://github.com/espressif/esp-idf/blob/master/components/driver/include/driver/adc_common.h
 # pin to adc1 channel mapping
+# https://github.com/espressif/esp-idf/blob/v4.4.8/components/driver/include/driver/adc.h
 ESP32_VARIANT_ADC1_PIN_TO_CHANNEL = {
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32/include/soc/adc_channel.h
     VARIANT_ESP32: {
         36: adc1_channel_t.ADC1_CHANNEL_0,
         37: adc1_channel_t.ADC1_CHANNEL_1,
@@ -60,6 +61,41 @@ ESP32_VARIANT_ADC1_PIN_TO_CHANNEL = {
         34: adc1_channel_t.ADC1_CHANNEL_6,
         35: adc1_channel_t.ADC1_CHANNEL_7,
     },
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32c2/include/soc/adc_channel.h
+    VARIANT_ESP32C2: {
+        0: adc1_channel_t.ADC1_CHANNEL_0,
+        1: adc1_channel_t.ADC1_CHANNEL_1,
+        2: adc1_channel_t.ADC1_CHANNEL_2,
+        3: adc1_channel_t.ADC1_CHANNEL_3,
+        4: adc1_channel_t.ADC1_CHANNEL_4,
+    },
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32c3/include/soc/adc_channel.h
+    VARIANT_ESP32C3: {
+        0: adc1_channel_t.ADC1_CHANNEL_0,
+        1: adc1_channel_t.ADC1_CHANNEL_1,
+        2: adc1_channel_t.ADC1_CHANNEL_2,
+        3: adc1_channel_t.ADC1_CHANNEL_3,
+        4: adc1_channel_t.ADC1_CHANNEL_4,
+    },
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32c6/include/soc/adc_channel.h
+    VARIANT_ESP32C6: {
+        0: adc1_channel_t.ADC1_CHANNEL_0,
+        1: adc1_channel_t.ADC1_CHANNEL_1,
+        2: adc1_channel_t.ADC1_CHANNEL_2,
+        3: adc1_channel_t.ADC1_CHANNEL_3,
+        4: adc1_channel_t.ADC1_CHANNEL_4,
+        5: adc1_channel_t.ADC1_CHANNEL_5,
+        6: adc1_channel_t.ADC1_CHANNEL_6,
+    },
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32h2/include/soc/adc_channel.h
+    VARIANT_ESP32H2: {
+        1: adc1_channel_t.ADC1_CHANNEL_0,
+        2: adc1_channel_t.ADC1_CHANNEL_1,
+        3: adc1_channel_t.ADC1_CHANNEL_2,
+        4: adc1_channel_t.ADC1_CHANNEL_3,
+        5: adc1_channel_t.ADC1_CHANNEL_4,
+    },
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32s2/include/soc/adc_channel.h
     VARIANT_ESP32S2: {
         1: adc1_channel_t.ADC1_CHANNEL_0,
         2: adc1_channel_t.ADC1_CHANNEL_1,
@@ -72,6 +108,7 @@ ESP32_VARIANT_ADC1_PIN_TO_CHANNEL = {
         9: adc1_channel_t.ADC1_CHANNEL_8,
         10: adc1_channel_t.ADC1_CHANNEL_9,
     },
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32s3/include/soc/adc_channel.h
     VARIANT_ESP32S3: {
         1: adc1_channel_t.ADC1_CHANNEL_0,
         2: adc1_channel_t.ADC1_CHANNEL_1,
@@ -84,40 +121,12 @@ ESP32_VARIANT_ADC1_PIN_TO_CHANNEL = {
         9: adc1_channel_t.ADC1_CHANNEL_8,
         10: adc1_channel_t.ADC1_CHANNEL_9,
     },
-    VARIANT_ESP32C3: {
-        0: adc1_channel_t.ADC1_CHANNEL_0,
-        1: adc1_channel_t.ADC1_CHANNEL_1,
-        2: adc1_channel_t.ADC1_CHANNEL_2,
-        3: adc1_channel_t.ADC1_CHANNEL_3,
-        4: adc1_channel_t.ADC1_CHANNEL_4,
-    },
-    VARIANT_ESP32C2: {
-        0: adc1_channel_t.ADC1_CHANNEL_0,
-        1: adc1_channel_t.ADC1_CHANNEL_1,
-        2: adc1_channel_t.ADC1_CHANNEL_2,
-        3: adc1_channel_t.ADC1_CHANNEL_3,
-        4: adc1_channel_t.ADC1_CHANNEL_4,
-    },
-    VARIANT_ESP32C6: {
-        0: adc1_channel_t.ADC1_CHANNEL_0,
-        1: adc1_channel_t.ADC1_CHANNEL_1,
-        2: adc1_channel_t.ADC1_CHANNEL_2,
-        3: adc1_channel_t.ADC1_CHANNEL_3,
-        4: adc1_channel_t.ADC1_CHANNEL_4,
-        5: adc1_channel_t.ADC1_CHANNEL_5,
-        6: adc1_channel_t.ADC1_CHANNEL_6,
-    },
-    VARIANT_ESP32H2: {
-        1: adc1_channel_t.ADC1_CHANNEL_0,
-        2: adc1_channel_t.ADC1_CHANNEL_1,
-        3: adc1_channel_t.ADC1_CHANNEL_2,
-        4: adc1_channel_t.ADC1_CHANNEL_3,
-        5: adc1_channel_t.ADC1_CHANNEL_4,
-    },
 }
 
+# pin to adc2 channel mapping
+# https://github.com/espressif/esp-idf/blob/v4.4.8/components/driver/include/driver/adc.h
 ESP32_VARIANT_ADC2_PIN_TO_CHANNEL = {
-    # TODO: add other variants
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32/include/soc/adc_channel.h
     VARIANT_ESP32: {
         4: adc2_channel_t.ADC2_CHANNEL_0,
         0: adc2_channel_t.ADC2_CHANNEL_1,
@@ -130,6 +139,19 @@ ESP32_VARIANT_ADC2_PIN_TO_CHANNEL = {
         25: adc2_channel_t.ADC2_CHANNEL_8,
         26: adc2_channel_t.ADC2_CHANNEL_9,
     },
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32c2/include/soc/adc_channel.h
+    VARIANT_ESP32C2: {
+        5: adc2_channel_t.ADC2_CHANNEL_0,
+    },
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32c3/include/soc/adc_channel.h
+    VARIANT_ESP32C3: {
+        5: adc2_channel_t.ADC2_CHANNEL_0,
+    },
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32c6/include/soc/adc_channel.h
+    VARIANT_ESP32C6: {},  # no ADC2
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32h2/include/soc/adc_channel.h
+    VARIANT_ESP32H2: {},  # no ADC2
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32s2/include/soc/adc_channel.h
     VARIANT_ESP32S2: {
         11: adc2_channel_t.ADC2_CHANNEL_0,
         12: adc2_channel_t.ADC2_CHANNEL_1,
@@ -142,6 +164,7 @@ ESP32_VARIANT_ADC2_PIN_TO_CHANNEL = {
         19: adc2_channel_t.ADC2_CHANNEL_8,
         20: adc2_channel_t.ADC2_CHANNEL_9,
     },
+    # https://github.com/espressif/esp-idf/blob/master/components/soc/esp32s3/include/soc/adc_channel.h
     VARIANT_ESP32S3: {
         11: adc2_channel_t.ADC2_CHANNEL_0,
         12: adc2_channel_t.ADC2_CHANNEL_1,
@@ -154,12 +177,6 @@ ESP32_VARIANT_ADC2_PIN_TO_CHANNEL = {
         19: adc2_channel_t.ADC2_CHANNEL_8,
         20: adc2_channel_t.ADC2_CHANNEL_9,
     },
-    VARIANT_ESP32C3: {
-        5: adc2_channel_t.ADC2_CHANNEL_0,
-    },
-    VARIANT_ESP32C2: {},
-    VARIANT_ESP32C6: {},
-    VARIANT_ESP32H2: {},
 }
 
 

esphome/components/api/api.proto

@@ -61,6 +61,7 @@ service APIConnection {
   rpc bluetooth_gatt_notify(BluetoothGATTNotifyRequest) returns (void) {}
   rpc subscribe_bluetooth_connections_free(SubscribeBluetoothConnectionsFreeRequest) returns (BluetoothConnectionsFreeResponse) {}
   rpc unsubscribe_bluetooth_le_advertisements(UnsubscribeBluetoothLEAdvertisementsRequest) returns (void) {}
+  rpc bluetooth_scanner_set_mode(BluetoothScannerSetModeRequest) returns (void) {}
 
   rpc subscribe_voice_assistant(SubscribeVoiceAssistantRequest) returns (void) {}
   rpc voice_assistant_get_configuration(VoiceAssistantConfigurationRequest) returns (VoiceAssistantConfigurationResponse) {}
@@ -1472,6 +1473,37 @@ message BluetoothDeviceClearCacheResponse {
   int32 error = 3;
 }
 
+enum BluetoothScannerState {
+  BLUETOOTH_SCANNER_STATE_IDLE = 0;
+  BLUETOOTH_SCANNER_STATE_STARTING = 1;
+  BLUETOOTH_SCANNER_STATE_RUNNING = 2;
+  BLUETOOTH_SCANNER_STATE_FAILED = 3;
+  BLUETOOTH_SCANNER_STATE_STOPPING = 4;
+  BLUETOOTH_SCANNER_STATE_STOPPED = 5;
+}
+
+enum BluetoothScannerMode {
+  BLUETOOTH_SCANNER_MODE_PASSIVE = 0;
+  BLUETOOTH_SCANNER_MODE_ACTIVE = 1;
+}
+
+message BluetoothScannerStateResponse {
+  option(id) = 126;
+  option(source) = SOURCE_SERVER;
+  option(ifdef) = "USE_BLUETOOTH_PROXY";
+
+  BluetoothScannerState state = 1;
+  BluetoothScannerMode mode = 2;
+}
+
+message BluetoothScannerSetModeRequest {
+  option(id) = 127;
+  option(source) = SOURCE_CLIENT;
+  option(ifdef) = "USE_BLUETOOTH_PROXY";
+
+  BluetoothScannerMode mode = 1;
+}
+
 // ==================== PUSH TO TALK ====================
 enum VoiceAssistantSubscribeFlag {
   VOICE_ASSISTANT_SUBSCRIBE_NONE = 0;

esphome/components/api/api_connection.cpp

@@ -1475,6 +1475,11 @@ BluetoothConnectionsFreeResponse APIConnection::subscribe_bluetooth_connections_
   resp.limit = bluetooth_proxy::global_bluetooth_proxy->get_bluetooth_connections_limit();
   return resp;
 }
+
+void APIConnection::bluetooth_scanner_set_mode(const BluetoothScannerSetModeRequest &msg) {
+  bluetooth_proxy::global_bluetooth_proxy->bluetooth_scanner_set_mode(
+      msg.mode == enums::BluetoothScannerMode::BLUETOOTH_SCANNER_MODE_ACTIVE);
+}
 #endif
 
 #ifdef USE_VOICE_ASSISTANT

esphome/components/api/api_connection.h

@@ -221,6 +221,7 @@ class APIConnection : public APIServerConnection {
   void bluetooth_gatt_notify(const BluetoothGATTNotifyRequest &msg) override;
   BluetoothConnectionsFreeResponse subscribe_bluetooth_connections_free(
       const SubscribeBluetoothConnectionsFreeRequest &msg) override;
+  void bluetooth_scanner_set_mode(const BluetoothScannerSetModeRequest &msg) override;
 
 #endif
 #ifdef USE_HOMEASSISTANT_TIME

esphome/components/api/api_pb2.cpp

@@ -422,6 +422,38 @@ const char *proto_enum_to_string<enums::BluetoothDeviceRequestType>(enums::Bluet
 }
 #endif
 #ifdef HAS_PROTO_MESSAGE_DUMP
+template<> const char *proto_enum_to_string<enums::BluetoothScannerState>(enums::BluetoothScannerState value) {
+  switch (value) {
+    case enums::BLUETOOTH_SCANNER_STATE_IDLE:
+      return "BLUETOOTH_SCANNER_STATE_IDLE";
+    case enums::BLUETOOTH_SCANNER_STATE_STARTING:
+      return "BLUETOOTH_SCANNER_STATE_STARTING";
+    case enums::BLUETOOTH_SCANNER_STATE_RUNNING:
+      return "BLUETOOTH_SCANNER_STATE_RUNNING";
+    case enums::BLUETOOTH_SCANNER_STATE_FAILED:
+      return "BLUETOOTH_SCANNER_STATE_FAILED";
+    case enums::BLUETOOTH_SCANNER_STATE_STOPPING:
+      return "BLUETOOTH_SCANNER_STATE_STOPPING";
+    case enums::BLUETOOTH_SCANNER_STATE_STOPPED:
+      return "BLUETOOTH_SCANNER_STATE_STOPPED";
+    default:
+      return "UNKNOWN";
+  }
+}
+#endif
+#ifdef HAS_PROTO_MESSAGE_DUMP
+template<> const char *proto_enum_to_string<enums::BluetoothScannerMode>(enums::BluetoothScannerMode value) {
+  switch (value) {
+    case enums::BLUETOOTH_SCANNER_MODE_PASSIVE:
+      return "BLUETOOTH_SCANNER_MODE_PASSIVE";
+    case enums::BLUETOOTH_SCANNER_MODE_ACTIVE:
+      return "BLUETOOTH_SCANNER_MODE_ACTIVE";
+    default:
+      return "UNKNOWN";
+  }
+}
+#endif
+#ifdef HAS_PROTO_MESSAGE_DUMP
 template<>
 const char *proto_enum_to_string<enums::VoiceAssistantSubscribeFlag>(enums::VoiceAssistantSubscribeFlag value) {
   switch (value) {
@@ -6775,6 +6807,61 @@ void BluetoothDeviceClearCacheResponse::dump_to(std::string &out) const {
   out.append("}");
 }
 #endif
+bool BluetoothScannerStateResponse::decode_varint(uint32_t field_id, ProtoVarInt value) {
+  switch (field_id) {
+    case 1: {
+      this->state = value.as_enum<enums::BluetoothScannerState>();
+      return true;
+    }
+    case 2: {
+      this->mode = value.as_enum<enums::BluetoothScannerMode>();
+      return true;
+    }
+    default:
+      return false;
+  }
+}
+void BluetoothScannerStateResponse::encode(ProtoWriteBuffer buffer) const {
+  buffer.encode_enum<enums::BluetoothScannerState>(1, this->state);
+  buffer.encode_enum<enums::BluetoothScannerMode>(2, this->mode);
+}
+#ifdef HAS_PROTO_MESSAGE_DUMP
+void BluetoothScannerStateResponse::dump_to(std::string &out) const {
+  __attribute__((unused)) char buffer[64];
+  out.append("BluetoothScannerStateResponse {\n");
+  out.append("  state: ");
+  out.append(proto_enum_to_string<enums::BluetoothScannerState>(this->state));
+  out.append("\n");
+
+  out.append("  mode: ");
+  out.append(proto_enum_to_string<enums::BluetoothScannerMode>(this->mode));
+  out.append("\n");
+  out.append("}");
+}
+#endif
+bool BluetoothScannerSetModeRequest::decode_varint(uint32_t field_id, ProtoVarInt value) {
+  switch (field_id) {
+    case 1: {
+      this->mode = value.as_enum<enums::BluetoothScannerMode>();
+      return true;
+    }
+    default:
+      return false;
+  }
+}
+void BluetoothScannerSetModeRequest::encode(ProtoWriteBuffer buffer) const {
+  buffer.encode_enum<enums::BluetoothScannerMode>(1, this->mode);
+}
+#ifdef HAS_PROTO_MESSAGE_DUMP
+void BluetoothScannerSetModeRequest::dump_to(std::string &out) const {
+  __attribute__((unused)) char buffer[64];
+  out.append("BluetoothScannerSetModeRequest {\n");
+  out.append("  mode: ");
+  out.append(proto_enum_to_string<enums::BluetoothScannerMode>(this->mode));
+  out.append("\n");
+  out.append("}");
+}
+#endif
 bool SubscribeVoiceAssistantRequest::decode_varint(uint32_t field_id, ProtoVarInt value) {
   switch (field_id) {
     case 1: {

esphome/components/api/api_pb2.h

@@ -169,6 +169,18 @@ enum BluetoothDeviceRequestType : uint32_t {
   BLUETOOTH_DEVICE_REQUEST_TYPE_CONNECT_V3_WITHOUT_CACHE = 5,
   BLUETOOTH_DEVICE_REQUEST_TYPE_CLEAR_CACHE = 6,
 };
+enum BluetoothScannerState : uint32_t {
+  BLUETOOTH_SCANNER_STATE_IDLE = 0,
+  BLUETOOTH_SCANNER_STATE_STARTING = 1,
+  BLUETOOTH_SCANNER_STATE_RUNNING = 2,
+  BLUETOOTH_SCANNER_STATE_FAILED = 3,
+  BLUETOOTH_SCANNER_STATE_STOPPING = 4,
+  BLUETOOTH_SCANNER_STATE_STOPPED = 5,
+};
+enum BluetoothScannerMode : uint32_t {
+  BLUETOOTH_SCANNER_MODE_PASSIVE = 0,
+  BLUETOOTH_SCANNER_MODE_ACTIVE = 1,
+};
 enum VoiceAssistantSubscribeFlag : uint32_t {
   VOICE_ASSISTANT_SUBSCRIBE_NONE = 0,
   VOICE_ASSISTANT_SUBSCRIBE_API_AUDIO = 1,
@@ -1742,6 +1754,29 @@ class BluetoothDeviceClearCacheResponse : public ProtoMessage {
  protected:
   bool decode_varint(uint32_t field_id, ProtoVarInt value) override;
 };
+class BluetoothScannerStateResponse : public ProtoMessage {
+ public:
+  enums::BluetoothScannerState state{};
+  enums::BluetoothScannerMode mode{};
+  void encode(ProtoWriteBuffer buffer) const override;
+#ifdef HAS_PROTO_MESSAGE_DUMP
+  void dump_to(std::string &out) const override;
+#endif
+
+ protected:
+  bool decode_varint(uint32_t field_id, ProtoVarInt value) override;
+};
+class BluetoothScannerSetModeRequest : public ProtoMessage {
+ public:
+  enums::BluetoothScannerMode mode{};
+  void encode(ProtoWriteBuffer buffer) const override;
+#ifdef HAS_PROTO_MESSAGE_DUMP
+  void dump_to(std::string &out) const override;
+#endif
+
+ protected:
+  bool decode_varint(uint32_t field_id, ProtoVarInt value) override;
+};
 class SubscribeVoiceAssistantRequest : public ProtoMessage {
  public:
   bool subscribe{false};

esphome/components/api/api_pb2_service.cpp

@@ -472,6 +472,16 @@ bool APIServerConnectionBase::send_bluetooth_device_clear_cache_response(const B
   return this->send_message_<BluetoothDeviceClearCacheResponse>(msg, 88);
 }
 #endif
+#ifdef USE_BLUETOOTH_PROXY
+bool APIServerConnectionBase::send_bluetooth_scanner_state_response(const BluetoothScannerStateResponse &msg) {
+#ifdef HAS_PROTO_MESSAGE_DUMP
+  ESP_LOGVV(TAG, "send_bluetooth_scanner_state_response: %s", msg.dump().c_str());
+#endif
+  return this->send_message_<BluetoothScannerStateResponse>(msg, 126);
+}
+#endif
+#ifdef USE_BLUETOOTH_PROXY
+#endif
 #ifdef USE_VOICE_ASSISTANT
 #endif
 #ifdef USE_VOICE_ASSISTANT
@@ -1212,6 +1222,17 @@ bool APIServerConnectionBase::read_message(uint32_t msg_size, uint32_t msg_type,
       ESP_LOGVV(TAG, "on_noise_encryption_set_key_request: %s", msg.dump().c_str());
 #endif
       this->on_noise_encryption_set_key_request(msg);
+#endif
+      break;
+    }
+    case 127: {
+#ifdef USE_BLUETOOTH_PROXY
+      BluetoothScannerSetModeRequest msg;
+      msg.decode(msg_data, msg_size);
+#ifdef HAS_PROTO_MESSAGE_DUMP
+      ESP_LOGVV(TAG, "on_bluetooth_scanner_set_mode_request: %s", msg.dump().c_str());
+#endif
+      this->on_bluetooth_scanner_set_mode_request(msg);
 #endif
       break;
     }
@@ -1705,6 +1726,19 @@ void APIServerConnection::on_unsubscribe_bluetooth_le_advertisements_request(
   this->unsubscribe_bluetooth_le_advertisements(msg);
 }
 #endif
+#ifdef USE_BLUETOOTH_PROXY
+void APIServerConnection::on_bluetooth_scanner_set_mode_request(const BluetoothScannerSetModeRequest &msg) {
+  if (!this->is_connection_setup()) {
+    this->on_no_setup_connection();
+    return;
+  }
+  if (!this->is_authenticated()) {
+    this->on_unauthenticated_access();
+    return;
+  }
+  this->bluetooth_scanner_set_mode(msg);
+}
+#endif
 #ifdef USE_VOICE_ASSISTANT
 void APIServerConnection::on_subscribe_voice_assistant_request(const SubscribeVoiceAssistantRequest &msg) {
   if (!this->is_connection_setup()) {

esphome/components/api/api_pb2_service.h

@@ -234,6 +234,12 @@ class APIServerConnectionBase : public ProtoService {
 #ifdef USE_BLUETOOTH_PROXY
   bool send_bluetooth_device_clear_cache_response(const BluetoothDeviceClearCacheResponse &msg);
 #endif
+#ifdef USE_BLUETOOTH_PROXY
+  bool send_bluetooth_scanner_state_response(const BluetoothScannerStateResponse &msg);
+#endif
+#ifdef USE_BLUETOOTH_PROXY
+  virtual void on_bluetooth_scanner_set_mode_request(const BluetoothScannerSetModeRequest &value){};
+#endif
 #ifdef USE_VOICE_ASSISTANT
   virtual void on_subscribe_voice_assistant_request(const SubscribeVoiceAssistantRequest &value){};
 #endif
@@ -440,6 +446,9 @@ class APIServerConnection : public APIServerConnectionBase {
 #ifdef USE_BLUETOOTH_PROXY
   virtual void unsubscribe_bluetooth_le_advertisements(const UnsubscribeBluetoothLEAdvertisementsRequest &msg) = 0;
 #endif
+#ifdef USE_BLUETOOTH_PROXY
+  virtual void bluetooth_scanner_set_mode(const BluetoothScannerSetModeRequest &msg) = 0;
+#endif
 #ifdef USE_VOICE_ASSISTANT
   virtual void subscribe_voice_assistant(const SubscribeVoiceAssistantRequest &msg) = 0;
 #endif
@@ -551,6 +560,9 @@ class APIServerConnection : public APIServerConnectionBase {
   void on_unsubscribe_bluetooth_le_advertisements_request(
       const UnsubscribeBluetoothLEAdvertisementsRequest &msg) override;
 #endif
+#ifdef USE_BLUETOOTH_PROXY
+  void on_bluetooth_scanner_set_mode_request(const BluetoothScannerSetModeRequest &msg) override;
+#endif
 #ifdef USE_VOICE_ASSISTANT
   void on_subscribe_voice_assistant_request(const SubscribeVoiceAssistantRequest &msg) override;
 #endif

esphome/components/api/client.py

@@ -29,9 +29,8 @@ async def async_run_logs(config: dict[str, Any], address: str) -> None:
     port: int = int(conf[CONF_PORT])
     password: str = conf[CONF_PASSWORD]
     noise_psk: str | None = None
-    if encryption_config := conf.get(CONF_ENCRYPTION):
-        if key := encryption_config.get(CONF_KEY):
-            noise_psk = key
+    if CONF_ENCRYPTION in conf:
+        noise_psk = conf[CONF_ENCRYPTION][CONF_KEY]
     _LOGGER.info("Starting log output from %s using esphome API", address)
     cli = APIClient(
         address,

esphome/components/as3935_i2c/as3935_i2c.h

@@ -1,10 +1,7 @@
 #pragma once
 
-#include "esphome/core/component.h"
 #include "esphome/components/as3935/as3935.h"
 #include "esphome/components/i2c/i2c.h"
-#include "esphome/components/sensor/sensor.h"
-#include "esphome/components/binary_sensor/binary_sensor.h"
 
 namespace esphome {
 namespace as3935_i2c {

esphome/components/as7341/as7341.h

@@ -7,7 +7,7 @@
 namespace esphome {
 namespace as7341 {
 
-static const uint8_t AS7341_CHIP_ID = 0X09;
+static const uint8_t AS7341_CHIP_ID = 0x09;
 
 static const uint8_t AS7341_CONFIG = 0x70;
 static const uint8_t AS7341_LED = 0x74;

esphome/components/audio/__init__.py

@@ -48,6 +48,12 @@ def set_stream_limits(
     min_sample_rate: int = _UNDEF,
     max_sample_rate: int = _UNDEF,
 ):
+    """Sets the limits for the audio stream that audio component can handle
+
+    When the component sinks audio (e.g., a speaker), these indicate the limits to the audio it can receive.
+    When the component sources audio (e.g., a microphone), these indicate the limits to the audio it can send.
+    """
+
     def set_limits_in_config(config):
         if min_bits_per_sample is not _UNDEF:
             config[CONF_MIN_BITS_PER_SAMPLE] = min_bits_per_sample
@@ -69,43 +75,87 @@ def final_validate_audio_schema(
     name: str,
     *,
     audio_device: str,
-    bits_per_sample: int,
-    channels: int,
-    sample_rate: int,
+    bits_per_sample: int = _UNDEF,
+    channels: int = _UNDEF,
+    sample_rate: int = _UNDEF,
+    enabled_channels: list[int] = _UNDEF,
+    audio_device_issue: bool = False,
 ):
+    """Validates audio compatibility when passed between different components.
+
+    The component derived from ``AUDIO_COMPONENT_SCHEMA`` should call ``set_stream_limits`` in a validator to specify its compatible settings
+
+      - If audio_device_issue is True, then the error message indicates the user should adjust the AUDIO_COMPONENT_SCHEMA derived component's configuration to match the values passed to this function
+      - If audio_device_issue is False, then the error message indicates the user should adjust the configuration of the component calling this function, as it falls out of the valid stream limits
+
+    Args:
+        name (str): Friendly name of the component calling this function with an audio component to validate
+        audio_device (str): The configuration parameter name that contains the ID of an AUDIO_COMPONENT_SCHEMA derived component to validate against
+        bits_per_sample (int, optional): The desired bits per sample
+        channels (int, optional): The desired number of channels
+        sample_rate (int, optional): The desired sample rate
+        enabled_channels (list[int], optional): The desired enabled channels
+        audio_device_issue (bool, optional): Format the error message to indicate the problem is in the configuration for the ``audio_device`` component. Defaults to False.
+    """
+
     def validate_audio_compatiblity(audio_config):
         audio_schema = {}
 
-        try:
-            cv.int_range(
-                min=audio_config.get(CONF_MIN_BITS_PER_SAMPLE),
-                max=audio_config.get(CONF_MAX_BITS_PER_SAMPLE),
-            )(bits_per_sample)
-        except cv.Invalid as exc:
-            raise cv.Invalid(
-                f"Invalid configuration for the {name} component. The {CONF_BITS_PER_SAMPLE} {str(exc)}"
-            ) from exc
+        if bits_per_sample is not _UNDEF:
+            try:
+                cv.int_range(
+                    min=audio_config.get(CONF_MIN_BITS_PER_SAMPLE),
+                    max=audio_config.get(CONF_MAX_BITS_PER_SAMPLE),
+                )(bits_per_sample)
+            except cv.Invalid as exc:
+                if audio_device_issue:
+                    error_string = f"Invalid configuration for the specified {audio_device}. The {name} component requires {bits_per_sample} bits per sample."
+                else:
+                    error_string = f"Invalid configuration for the {name} component. The {CONF_BITS_PER_SAMPLE} {str(exc)}"
+                raise cv.Invalid(error_string) from exc
 
-        try:
-            cv.int_range(
-                min=audio_config.get(CONF_MIN_CHANNELS),
-                max=audio_config.get(CONF_MAX_CHANNELS),
-            )(channels)
-        except cv.Invalid as exc:
-            raise cv.Invalid(
-                f"Invalid configuration for the {name} component. The {CONF_NUM_CHANNELS} {str(exc)}"
-            ) from exc
+        if channels is not _UNDEF:
+            try:
+                cv.int_range(
+                    min=audio_config.get(CONF_MIN_CHANNELS),
+                    max=audio_config.get(CONF_MAX_CHANNELS),
+                )(channels)
+            except cv.Invalid as exc:
+                if audio_device_issue:
+                    error_string = f"Invalid configuration for the specified {audio_device}. The {name} component requires {channels} channels."
+                else:
+                    error_string = f"Invalid configuration for the {name} component. The {CONF_NUM_CHANNELS} {str(exc)}"
+                raise cv.Invalid(error_string) from exc
 
-        try:
-            cv.int_range(
-                min=audio_config.get(CONF_MIN_SAMPLE_RATE),
-                max=audio_config.get(CONF_MAX_SAMPLE_RATE),
-            )(sample_rate)
-            return cv.Schema(audio_schema, extra=cv.ALLOW_EXTRA)(audio_config)
-        except cv.Invalid as exc:
-            raise cv.Invalid(
-                f"Invalid configuration for the {name} component. The {CONF_SAMPLE_RATE} {str(exc)}"
-            ) from exc
+        if sample_rate is not _UNDEF:
+            try:
+                cv.int_range(
+                    min=audio_config.get(CONF_MIN_SAMPLE_RATE),
+                    max=audio_config.get(CONF_MAX_SAMPLE_RATE),
+                )(sample_rate)
+            except cv.Invalid as exc:
+                if audio_device_issue:
+                    error_string = f"Invalid configuration for the specified {audio_device}. The {name} component requires a {sample_rate} sample rate."
+                else:
+                    error_string = f"Invalid configuration for the {name} component. The {CONF_SAMPLE_RATE} {str(exc)}"
+                raise cv.Invalid(error_string) from exc
+
+        if enabled_channels is not _UNDEF:
+            for channel in enabled_channels:
+                try:
+                    # Channels are 0-indexed
+                    cv.int_range(
+                        min=0,
+                        max=audio_config.get(CONF_MAX_CHANNELS) - 1,
+                    )(channel)
+                except cv.Invalid as exc:
+                    if audio_device_issue:
+                        error_string = f"Invalid configuration for the specified {audio_device}. The {name} component requires channel {channel}."
+                    else:
+                        error_string = f"Invalid configuration for the {name} component. Enabled channel {channel} {str(exc)}"
+                    raise cv.Invalid(error_string) from exc
+
+        return cv.Schema(audio_schema, extra=cv.ALLOW_EXTRA)(audio_config)
 
     return cv.Schema(
         {

esphome/components/audio/audio_resampler.cpp

@@ -4,6 +4,8 @@
 
 #include "esphome/core/hal.h"
 
+#include <cstring>
+
 namespace esphome {
 namespace audio {
 

esphome/components/audio/audio_resampler.h

@@ -6,6 +6,7 @@
 #include "audio_transfer_buffer.h"
 
 #include "esphome/core/defines.h"
+#include "esphome/core/helpers.h"
 #include "esphome/core/ring_buffer.h"
 
 #ifdef USE_SPEAKER

esphome/components/binary_sensor/binary_sensor.cpp

@@ -15,21 +15,17 @@ void BinarySensor::publish_state(bool state) {
   if (!this->publish_dedup_.next(state))
     return;
   if (this->filter_list_ == nullptr) {
-    this->send_state_internal(state, false);
+    this->send_state_internal(state);
   } else {
-    this->filter_list_->input(state, false);
+    this->filter_list_->input(state);
   }
 }
 void BinarySensor::publish_initial_state(bool state) {
-  if (!this->publish_dedup_.next(state))
-    return;
-  if (this->filter_list_ == nullptr) {
-    this->send_state_internal(state, true);
-  } else {
-    this->filter_list_->input(state, true);
-  }
+  this->has_state_ = false;
+  this->publish_state(state);
 }
-void BinarySensor::send_state_internal(bool state, bool is_initial) {
+void BinarySensor::send_state_internal(bool state) {
+  bool is_initial = !this->has_state_;
   if (is_initial) {
     ESP_LOGD(TAG, "'%s': Sending initial state %s", this->get_name().c_str(), ONOFF(state));
   } else {

esphome/components/binary_sensor/binary_sensor.h

@@ -67,7 +67,7 @@ class BinarySensor : public EntityBase, public EntityBase_DeviceClass {
 
   // ========== INTERNAL METHODS ==========
   // (In most use cases you won't need these)
-  void send_state_internal(bool state, bool is_initial);
+  void send_state_internal(bool state);
 
   /// Return whether this binary sensor has outputted a state.
   virtual bool has_state() const;

esphome/components/binary_sensor/filter.cpp

@@ -9,37 +9,37 @@ namespace binary_sensor {
 
 static const char *const TAG = "sensor.filter";
 
-void Filter::output(bool value, bool is_initial) {
+void Filter::output(bool value) {
   if (!this->dedup_.next(value))
     return;
 
   if (this->next_ == nullptr) {
-    this->parent_->send_state_internal(value, is_initial);
+    this->parent_->send_state_internal(value);
   } else {
-    this->next_->input(value, is_initial);
+    this->next_->input(value);
   }
 }
-void Filter::input(bool value, bool is_initial) {
-  auto b = this->new_value(value, is_initial);
+void Filter::input(bool value) {
+  auto b = this->new_value(value);
   if (b.has_value()) {
-    this->output(*b, is_initial);
+    this->output(*b);
   }
 }
 
-optional<bool> DelayedOnOffFilter::new_value(bool value, bool is_initial) {
+optional<bool> DelayedOnOffFilter::new_value(bool value) {
   if (value) {
-    this->set_timeout("ON_OFF", this->on_delay_.value(), [this, is_initial]() { this->output(true, is_initial); });
+    this->set_timeout("ON_OFF", this->on_delay_.value(), [this]() { this->output(true); });
   } else {
-    this->set_timeout("ON_OFF", this->off_delay_.value(), [this, is_initial]() { this->output(false, is_initial); });
+    this->set_timeout("ON_OFF", this->off_delay_.value(), [this]() { this->output(false); });
   }
   return {};
 }
 
 float DelayedOnOffFilter::get_setup_priority() const { return setup_priority::HARDWARE; }
 
-optional<bool> DelayedOnFilter::new_value(bool value, bool is_initial) {
+optional<bool> DelayedOnFilter::new_value(bool value) {
   if (value) {
-    this->set_timeout("ON", this->delay_.value(), [this, is_initial]() { this->output(true, is_initial); });
+    this->set_timeout("ON", this->delay_.value(), [this]() { this->output(true); });
     return {};
   } else {
     this->cancel_timeout("ON");
@@ -49,9 +49,9 @@ optional<bool> DelayedOnFilter::new_value(bool value, bool is_initial) {
 
 float DelayedOnFilter::get_setup_priority() const { return setup_priority::HARDWARE; }
 
-optional<bool> DelayedOffFilter::new_value(bool value, bool is_initial) {
+optional<bool> DelayedOffFilter::new_value(bool value) {
   if (!value) {
-    this->set_timeout("OFF", this->delay_.value(), [this, is_initial]() { this->output(false, is_initial); });
+    this->set_timeout("OFF", this->delay_.value(), [this]() { this->output(false); });
     return {};
   } else {
     this->cancel_timeout("OFF");
@@ -61,11 +61,11 @@ optional<bool> DelayedOffFilter::new_value(bool value, bool is_initial) {
 
 float DelayedOffFilter::get_setup_priority() const { return setup_priority::HARDWARE; }
 
-optional<bool> InvertFilter::new_value(bool value, bool is_initial) { return !value; }
+optional<bool> InvertFilter::new_value(bool value) { return !value; }
 
 AutorepeatFilter::AutorepeatFilter(std::vector<AutorepeatFilterTiming> timings) : timings_(std::move(timings)) {}
 
-optional<bool> AutorepeatFilter::new_value(bool value, bool is_initial) {
+optional<bool> AutorepeatFilter::new_value(bool value) {
   if (value) {
     // Ignore if already running
     if (this->active_timing_ != 0)
@@ -101,7 +101,7 @@ void AutorepeatFilter::next_timing_() {
 
 void AutorepeatFilter::next_value_(bool val) {
   const AutorepeatFilterTiming &timing = this->timings_[this->active_timing_ - 2];
-  this->output(val, false);  // This is at least the second one so not initial
+  this->output(val);
   this->set_timeout("ON_OFF", val ? timing.time_on : timing.time_off, [this, val]() { this->next_value_(!val); });
 }
 
@@ -109,18 +109,18 @@ float AutorepeatFilter::get_setup_priority() const { return setup_priority::HARD
 
 LambdaFilter::LambdaFilter(std::function<optional<bool>(bool)> f) : f_(std::move(f)) {}
 
-optional<bool> LambdaFilter::new_value(bool value, bool is_initial) { return this->f_(value); }
+optional<bool> LambdaFilter::new_value(bool value) { return this->f_(value); }
 
-optional<bool> SettleFilter::new_value(bool value, bool is_initial) {
+optional<bool> SettleFilter::new_value(bool value) {
   if (!this->steady_) {
-    this->set_timeout("SETTLE", this->delay_.value(), [this, value, is_initial]() {
+    this->set_timeout("SETTLE", this->delay_.value(), [this, value]() {
       this->steady_ = true;
-      this->output(value, is_initial);
+      this->output(value);
     });
     return {};
   } else {
     this->steady_ = false;
-    this->output(value, is_initial);
+    this->output(value);
     this->set_timeout("SETTLE", this->delay_.value(), [this]() { this->steady_ = true; });
     return value;
   }

esphome/components/binary_sensor/filter.h

@@ -14,11 +14,11 @@ class BinarySensor;
 
 class Filter {
  public:
-  virtual optional<bool> new_value(bool value, bool is_initial) = 0;
+  virtual optional<bool> new_value(bool value) = 0;
 
-  void input(bool value, bool is_initial);
+  void input(bool value);
 
-  void output(bool value, bool is_initial);
+  void output(bool value);
 
  protected:
   friend BinarySensor;
@@ -30,7 +30,7 @@ class Filter {
 
 class DelayedOnOffFilter : public Filter, public Component {
  public:
-  optional<bool> new_value(bool value, bool is_initial) override;
+  optional<bool> new_value(bool value) override;
 
   float get_setup_priority() const override;
 
@@ -44,7 +44,7 @@ class DelayedOnOffFilter : public Filter, public Component {
 
 class DelayedOnFilter : public Filter, public Component {
  public:
-  optional<bool> new_value(bool value, bool is_initial) override;
+  optional<bool> new_value(bool value) override;
 
   float get_setup_priority() const override;
 
@@ -56,7 +56,7 @@ class DelayedOnFilter : public Filter, public Component {
 
 class DelayedOffFilter : public Filter, public Component {
  public:
-  optional<bool> new_value(bool value, bool is_initial) override;
+  optional<bool> new_value(bool value) override;
 
   float get_setup_priority() const override;
 
@@ -68,7 +68,7 @@ class DelayedOffFilter : public Filter, public Component {
 
 class InvertFilter : public Filter {
  public:
-  optional<bool> new_value(bool value, bool is_initial) override;
+  optional<bool> new_value(bool value) override;
 };
 
 struct AutorepeatFilterTiming {
@@ -86,7 +86,7 @@ class AutorepeatFilter : public Filter, public Component {
  public:
   explicit AutorepeatFilter(std::vector<AutorepeatFilterTiming> timings);
 
-  optional<bool> new_value(bool value, bool is_initial) override;
+  optional<bool> new_value(bool value) override;
 
   float get_setup_priority() const override;
 
@@ -102,7 +102,7 @@ class LambdaFilter : public Filter {
  public:
   explicit LambdaFilter(std::function<optional<bool>(bool)> f);
 
-  optional<bool> new_value(bool value, bool is_initial) override;
+  optional<bool> new_value(bool value) override;
 
  protected:
   std::function<optional<bool>(bool)> f_;
@@ -110,7 +110,7 @@ class LambdaFilter : public Filter {
 
 class SettleFilter : public Filter, public Component {
  public:
-  optional<bool> new_value(bool value, bool is_initial) override;
+  optional<bool> new_value(bool value) override;
 
   float get_setup_priority() const override;
 

esphome/components/bl0906/constants.h

@@ -45,7 +45,7 @@ static const uint8_t BL0906_WRITE_COMMAND = 0xCA;
 static const uint8_t BL0906_V_RMS = 0x16;
 
 // Total power
-static const uint8_t BL0906_WATT_SUM = 0X2C;
+static const uint8_t BL0906_WATT_SUM = 0x2C;
 
 // Current1~6
 static const uint8_t BL0906_I_1_RMS = 0x0D;  // current_1
@@ -56,29 +56,29 @@ static const uint8_t BL0906_I_5_RMS = 0x13;
 static const uint8_t BL0906_I_6_RMS = 0x14;  // current_6
 
 // Power1~6
-static const uint8_t BL0906_WATT_1 = 0X23;  // power_1
-static const uint8_t BL0906_WATT_2 = 0X24;
-static const uint8_t BL0906_WATT_3 = 0X25;
-static const uint8_t BL0906_WATT_4 = 0X26;
-static const uint8_t BL0906_WATT_5 = 0X29;
-static const uint8_t BL0906_WATT_6 = 0X2A;  // power_6
+static const uint8_t BL0906_WATT_1 = 0x23;  // power_1
+static const uint8_t BL0906_WATT_2 = 0x24;
+static const uint8_t BL0906_WATT_3 = 0x25;
+static const uint8_t BL0906_WATT_4 = 0x26;
+static const uint8_t BL0906_WATT_5 = 0x29;
+static const uint8_t BL0906_WATT_6 = 0x2A;  // power_6
 
 // Active pulse count, unsigned
-static const uint8_t BL0906_CF_1_CNT = 0X30;  // Channel_1
-static const uint8_t BL0906_CF_2_CNT = 0X31;
-static const uint8_t BL0906_CF_3_CNT = 0X32;
-static const uint8_t BL0906_CF_4_CNT = 0X33;
-static const uint8_t BL0906_CF_5_CNT = 0X36;
-static const uint8_t BL0906_CF_6_CNT = 0X37;  // Channel_6
+static const uint8_t BL0906_CF_1_CNT = 0x30;  // Channel_1
+static const uint8_t BL0906_CF_2_CNT = 0x31;
+static const uint8_t BL0906_CF_3_CNT = 0x32;
+static const uint8_t BL0906_CF_4_CNT = 0x33;
+static const uint8_t BL0906_CF_5_CNT = 0x36;
+static const uint8_t BL0906_CF_6_CNT = 0x37;  // Channel_6
 
 // Total active pulse count, unsigned
-static const uint8_t BL0906_CF_SUM_CNT = 0X39;
+static const uint8_t BL0906_CF_SUM_CNT = 0x39;
 
 // Voltage frequency cycle
-static const uint8_t BL0906_FREQUENCY = 0X4E;
+static const uint8_t BL0906_FREQUENCY = 0x4E;
 
 // Internal temperature
-static const uint8_t BL0906_TEMPERATURE = 0X5E;
+static const uint8_t BL0906_TEMPERATURE = 0x5E;
 
 // Calibration register
 // RMS gain adjustment register

esphome/components/bluetooth_proxy/bluetooth_proxy.cpp

@@ -25,6 +25,22 @@ std::vector<uint64_t> get_128bit_uuid_vec(esp_bt_uuid_t uuid_source) {
 
 BluetoothProxy::BluetoothProxy() { global_bluetooth_proxy = this; }
 
+void BluetoothProxy::setup() {
+  this->parent_->add_scanner_state_callback([this](esp32_ble_tracker::ScannerState state) {
+    if (this->api_connection_ != nullptr) {
+      this->send_bluetooth_scanner_state_(state);
+    }
+  });
+}
+
+void BluetoothProxy::send_bluetooth_scanner_state_(esp32_ble_tracker::ScannerState state) {
+  api::BluetoothScannerStateResponse resp;
+  resp.state = static_cast<api::enums::BluetoothScannerState>(state);
+  resp.mode = this->parent_->get_scan_active() ? api::enums::BluetoothScannerMode::BLUETOOTH_SCANNER_MODE_ACTIVE
+                                               : api::enums::BluetoothScannerMode::BLUETOOTH_SCANNER_MODE_PASSIVE;
+  this->api_connection_->send_bluetooth_scanner_state_response(resp);
+}
+
 bool BluetoothProxy::parse_device(const esp32_ble_tracker::ESPBTDevice &device) {
   if (!api::global_api_server->is_connected() || this->api_connection_ == nullptr || this->raw_advertisements_)
     return false;
@@ -453,6 +469,8 @@ void BluetoothProxy::subscribe_api_connection(api::APIConnection *api_connection
   this->api_connection_ = api_connection;
   this->raw_advertisements_ = flags & BluetoothProxySubscriptionFlag::SUBSCRIPTION_RAW_ADVERTISEMENTS;
   this->parent_->recalculate_advertisement_parser_types();
+
+  this->send_bluetooth_scanner_state_(this->parent_->get_scanner_state());
 }
 
 void BluetoothProxy::unsubscribe_api_connection(api::APIConnection *api_connection) {
@@ -525,6 +543,17 @@ void BluetoothProxy::send_device_unpairing(uint64_t address, bool success, esp_e
   this->api_connection_->send_bluetooth_device_unpairing_response(call);
 }
 
+void BluetoothProxy::bluetooth_scanner_set_mode(bool active) {
+  if (this->parent_->get_scan_active() == active) {
+    return;
+  }
+  ESP_LOGD(TAG, "Setting scanner mode to %s", active ? "active" : "passive");
+  this->parent_->set_scan_active(active);
+  this->parent_->stop_scan();
+  this->parent_->set_scan_continuous(
+      true);  // Set this to true to automatically start scanning again when it has cleaned up.
+}
+
 BluetoothProxy *global_bluetooth_proxy = nullptr;  // NOLINT(cppcoreguidelines-avoid-non-const-global-variables)
 
 }  // namespace bluetooth_proxy

esphome/components/bluetooth_proxy/bluetooth_proxy.h

@@ -41,6 +41,7 @@ enum BluetoothProxyFeature : uint32_t {
   FEATURE_PAIRING = 1 << 3,
   FEATURE_CACHE_CLEARING = 1 << 4,
   FEATURE_RAW_ADVERTISEMENTS = 1 << 5,
+  FEATURE_STATE_AND_MODE = 1 << 6,
 };
 
 enum BluetoothProxySubscriptionFlag : uint32_t {
@@ -53,6 +54,7 @@ class BluetoothProxy : public esp32_ble_tracker::ESPBTDeviceListener, public Com
   bool parse_device(const esp32_ble_tracker::ESPBTDevice &device) override;
   bool parse_devices(esp_ble_gap_cb_param_t::ble_scan_result_evt_param *advertisements, size_t count) override;
   void dump_config() override;
+  void setup() override;
   void loop() override;
   esp32_ble_tracker::AdvertisementParserType get_advertisement_parser_type() override;
 
@@ -84,6 +86,8 @@ class BluetoothProxy : public esp32_ble_tracker::ESPBTDeviceListener, public Com
   void send_device_unpairing(uint64_t address, bool success, esp_err_t error = ESP_OK);
   void send_device_clear_cache(uint64_t address, bool success, esp_err_t error = ESP_OK);
 
+  void bluetooth_scanner_set_mode(bool active);
+
   static void uint64_to_bd_addr(uint64_t address, esp_bd_addr_t bd_addr) {
     bd_addr[0] = (address >> 40) & 0xff;
     bd_addr[1] = (address >> 32) & 0xff;
@@ -107,6 +111,7 @@ class BluetoothProxy : public esp32_ble_tracker::ESPBTDeviceListener, public Com
     uint32_t flags = 0;
     flags |= BluetoothProxyFeature::FEATURE_PASSIVE_SCAN;
     flags |= BluetoothProxyFeature::FEATURE_RAW_ADVERTISEMENTS;
+    flags |= BluetoothProxyFeature::FEATURE_STATE_AND_MODE;
     if (this->active_) {
       flags |= BluetoothProxyFeature::FEATURE_ACTIVE_CONNECTIONS;
       flags |= BluetoothProxyFeature::FEATURE_REMOTE_CACHING;
@@ -124,6 +129,7 @@ class BluetoothProxy : public esp32_ble_tracker::ESPBTDeviceListener, public Com
 
  protected:
   void send_api_packet_(const esp32_ble_tracker::ESPBTDevice &device);
+  void send_bluetooth_scanner_state_(esp32_ble_tracker::ScannerState state);
 
   BluetoothConnection *get_connection_(uint64_t address, bool reserve);
 

esphome/components/canbus/canbus.cpp

@@ -86,6 +86,9 @@ void Canbus::loop() {
       data.push_back(can_message.data[i]);
     }
 
+    this->callback_manager_(can_message.can_id, can_message.use_extended_id, can_message.remote_transmission_request,
+                            data);
+
     // fire all triggers
     for (auto *trigger : this->triggers_) {
       if ((trigger->can_id_ == (can_message.can_id & trigger->can_id_mask_)) &&

esphome/components/canbus/canbus.h

@@ -81,6 +81,20 @@ class Canbus : public Component {
   void set_bitrate(CanSpeed bit_rate) { this->bit_rate_ = bit_rate; }
 
   void add_trigger(CanbusTrigger *trigger);
+  /**
+   * Add a callback to be called when a CAN message is received. All received messages
+   * are passed to the callback without filtering.
+   *
+   * The callback function receives:
+   * - can_id of the received data
+   * - extended_id True if the can_id is an extended id
+   * - rtr If this is a remote transmission request
+   * - data The message data
+   */
+  void add_callback(
+      std::function<void(uint32_t can_id, bool extended_id, bool rtr, const std::vector<uint8_t> &data)> callback) {
+    this->callback_manager_.add(std::move(callback));
+  }
 
  protected:
   template<typename... Ts> friend class CanbusSendAction;
@@ -88,6 +102,8 @@ class Canbus : public Component {
   uint32_t can_id_;
   bool use_extended_id_;
   CanSpeed bit_rate_;
+  CallbackManager<void(uint32_t can_id, bool extended_id, bool rtr, const std::vector<uint8_t> &data)>
+      callback_manager_{};
 
   virtual bool setup_internal();
   virtual Error send_message(struct CanFrame *frame);

esphome/components/climate/climate_mode.h

@@ -20,7 +20,7 @@ enum ClimateMode : uint8_t {
   CLIMATE_MODE_FAN_ONLY = 4,
   /// The climate device is set to dry/humidity mode
   CLIMATE_MODE_DRY = 5,
-  /** The climate device is adjusting the temperatre dynamically.
+  /** The climate device is adjusting the temperature dynamically.
    * For example, the target temperature can be adjusted based on a schedule, or learned behavior.
    * The target temperature can't be adjusted when in this mode.
    */

esphome/components/climate/climate_traits.h

@@ -40,24 +40,24 @@ namespace climate {
  */
 class ClimateTraits {
  public:
-  bool get_supports_current_temperature() const { return supports_current_temperature_; }
+  bool get_supports_current_temperature() const { return this->supports_current_temperature_; }
   void set_supports_current_temperature(bool supports_current_temperature) {
-    supports_current_temperature_ = supports_current_temperature;
+    this->supports_current_temperature_ = supports_current_temperature;
   }
-  bool get_supports_current_humidity() const { return supports_current_humidity_; }
+  bool get_supports_current_humidity() const { return this->supports_current_humidity_; }
   void set_supports_current_humidity(bool supports_current_humidity) {
-    supports_current_humidity_ = supports_current_humidity;
+    this->supports_current_humidity_ = supports_current_humidity;
   }
-  bool get_supports_two_point_target_temperature() const { return supports_two_point_target_temperature_; }
+  bool get_supports_two_point_target_temperature() const { return this->supports_two_point_target_temperature_; }
   void set_supports_two_point_target_temperature(bool supports_two_point_target_temperature) {
-    supports_two_point_target_temperature_ = supports_two_point_target_temperature;
+    this->supports_two_point_target_temperature_ = supports_two_point_target_temperature;
   }
-  bool get_supports_target_humidity() const { return supports_target_humidity_; }
+  bool get_supports_target_humidity() const { return this->supports_target_humidity_; }
   void set_supports_target_humidity(bool supports_target_humidity) {
-    supports_target_humidity_ = supports_target_humidity;
+    this->supports_target_humidity_ = supports_target_humidity;
   }
-  void set_supported_modes(std::set<ClimateMode> modes) { supported_modes_ = std::move(modes); }
-  void add_supported_mode(ClimateMode mode) { supported_modes_.insert(mode); }
+  void set_supported_modes(std::set<ClimateMode> modes) { this->supported_modes_ = std::move(modes); }
+  void add_supported_mode(ClimateMode mode) { this->supported_modes_.insert(mode); }
   ESPDEPRECATED("This method is deprecated, use set_supported_modes() instead", "v1.20")
   void set_supports_auto_mode(bool supports_auto_mode) { set_mode_support_(CLIMATE_MODE_AUTO, supports_auto_mode); }
   ESPDEPRECATED("This method is deprecated, use set_supported_modes() instead", "v1.20")
@@ -72,15 +72,15 @@ class ClimateTraits {
   }
   ESPDEPRECATED("This method is deprecated, use set_supported_modes() instead", "v1.20")
   void set_supports_dry_mode(bool supports_dry_mode) { set_mode_support_(CLIMATE_MODE_DRY, supports_dry_mode); }
-  bool supports_mode(ClimateMode mode) const { return supported_modes_.count(mode); }
-  const std::set<ClimateMode> &get_supported_modes() const { return supported_modes_; }
+  bool supports_mode(ClimateMode mode) const { return this->supported_modes_.count(mode); }
+  const std::set<ClimateMode> &get_supported_modes() const { return this->supported_modes_; }
 
-  void set_supports_action(bool supports_action) { supports_action_ = supports_action; }
-  bool get_supports_action() const { return supports_action_; }
+  void set_supports_action(bool supports_action) { this->supports_action_ = supports_action; }
+  bool get_supports_action() const { return this->supports_action_; }
 
-  void set_supported_fan_modes(std::set<ClimateFanMode> modes) { supported_fan_modes_ = std::move(modes); }
-  void add_supported_fan_mode(ClimateFanMode mode) { supported_fan_modes_.insert(mode); }
-  void add_supported_custom_fan_mode(const std::string &mode) { supported_custom_fan_modes_.insert(mode); }
+  void set_supported_fan_modes(std::set<ClimateFanMode> modes) { this->supported_fan_modes_ = std::move(modes); }
+  void add_supported_fan_mode(ClimateFanMode mode) { this->supported_fan_modes_.insert(mode); }
+  void add_supported_custom_fan_mode(const std::string &mode) { this->supported_custom_fan_modes_.insert(mode); }
   ESPDEPRECATED("This method is deprecated, use set_supported_fan_modes() instead", "v1.20")
   void set_supports_fan_mode_on(bool supported) { set_fan_mode_support_(CLIMATE_FAN_ON, supported); }
   ESPDEPRECATED("This method is deprecated, use set_supported_fan_modes() instead", "v1.20")
@@ -99,35 +99,37 @@ class ClimateTraits {
   void set_supports_fan_mode_focus(bool supported) { set_fan_mode_support_(CLIMATE_FAN_FOCUS, supported); }
   ESPDEPRECATED("This method is deprecated, use set_supported_fan_modes() instead", "v1.20")
   void set_supports_fan_mode_diffuse(bool supported) { set_fan_mode_support_(CLIMATE_FAN_DIFFUSE, supported); }
-  bool supports_fan_mode(ClimateFanMode fan_mode) const { return supported_fan_modes_.count(fan_mode); }
-  bool get_supports_fan_modes() const { return !supported_fan_modes_.empty() || !supported_custom_fan_modes_.empty(); }
-  const std::set<ClimateFanMode> &get_supported_fan_modes() const { return supported_fan_modes_; }
+  bool supports_fan_mode(ClimateFanMode fan_mode) const { return this->supported_fan_modes_.count(fan_mode); }
+  bool get_supports_fan_modes() const {
+    return !this->supported_fan_modes_.empty() || !this->supported_custom_fan_modes_.empty();
+  }
+  const std::set<ClimateFanMode> &get_supported_fan_modes() const { return this->supported_fan_modes_; }
 
   void set_supported_custom_fan_modes(std::set<std::string> supported_custom_fan_modes) {
-    supported_custom_fan_modes_ = std::move(supported_custom_fan_modes);
+    this->supported_custom_fan_modes_ = std::move(supported_custom_fan_modes);
   }
-  const std::set<std::string> &get_supported_custom_fan_modes() const { return supported_custom_fan_modes_; }
+  const std::set<std::string> &get_supported_custom_fan_modes() const { return this->supported_custom_fan_modes_; }
   bool supports_custom_fan_mode(const std::string &custom_fan_mode) const {
-    return supported_custom_fan_modes_.count(custom_fan_mode);
+    return this->supported_custom_fan_modes_.count(custom_fan_mode);
   }
 
-  void set_supported_presets(std::set<ClimatePreset> presets) { supported_presets_ = std::move(presets); }
-  void add_supported_preset(ClimatePreset preset) { supported_presets_.insert(preset); }
-  void add_supported_custom_preset(const std::string &preset) { supported_custom_presets_.insert(preset); }
-  bool supports_preset(ClimatePreset preset) const { return supported_presets_.count(preset); }
-  bool get_supports_presets() const { return !supported_presets_.empty(); }
-  const std::set<climate::ClimatePreset> &get_supported_presets() const { return supported_presets_; }
+  void set_supported_presets(std::set<ClimatePreset> presets) { this->supported_presets_ = std::move(presets); }
+  void add_supported_preset(ClimatePreset preset) { this->supported_presets_.insert(preset); }
+  void add_supported_custom_preset(const std::string &preset) { this->supported_custom_presets_.insert(preset); }
+  bool supports_preset(ClimatePreset preset) const { return this->supported_presets_.count(preset); }
+  bool get_supports_presets() const { return !this->supported_presets_.empty(); }
+  const std::set<climate::ClimatePreset> &get_supported_presets() const { return this->supported_presets_; }
 
   void set_supported_custom_presets(std::set<std::string> supported_custom_presets) {
-    supported_custom_presets_ = std::move(supported_custom_presets);
+    this->supported_custom_presets_ = std::move(supported_custom_presets);
   }
-  const std::set<std::string> &get_supported_custom_presets() const { return supported_custom_presets_; }
+  const std::set<std::string> &get_supported_custom_presets() const { return this->supported_custom_presets_; }
   bool supports_custom_preset(const std::string &custom_preset) const {
-    return supported_custom_presets_.count(custom_preset);
+    return this->supported_custom_presets_.count(custom_preset);
   }
 
-  void set_supported_swing_modes(std::set<ClimateSwingMode> modes) { supported_swing_modes_ = std::move(modes); }
-  void add_supported_swing_mode(ClimateSwingMode mode) { supported_swing_modes_.insert(mode); }
+  void set_supported_swing_modes(std::set<ClimateSwingMode> modes) { this->supported_swing_modes_ = std::move(modes); }
+  void add_supported_swing_mode(ClimateSwingMode mode) { this->supported_swing_modes_.insert(mode); }
   ESPDEPRECATED("This method is deprecated, use set_supported_swing_modes() instead", "v1.20")
   void set_supports_swing_mode_off(bool supported) { set_swing_mode_support_(CLIMATE_SWING_OFF, supported); }
   ESPDEPRECATED("This method is deprecated, use set_supported_swing_modes() instead", "v1.20")
@@ -138,54 +140,58 @@ class ClimateTraits {
   void set_supports_swing_mode_horizontal(bool supported) {
     set_swing_mode_support_(CLIMATE_SWING_HORIZONTAL, supported);
   }
-  bool supports_swing_mode(ClimateSwingMode swing_mode) const { return supported_swing_modes_.count(swing_mode); }
-  bool get_supports_swing_modes() const { return !supported_swing_modes_.empty(); }
-  const std::set<ClimateSwingMode> &get_supported_swing_modes() const { return supported_swing_modes_; }
+  bool supports_swing_mode(ClimateSwingMode swing_mode) const { return this->supported_swing_modes_.count(swing_mode); }
+  bool get_supports_swing_modes() const { return !this->supported_swing_modes_.empty(); }
+  const std::set<ClimateSwingMode> &get_supported_swing_modes() const { return this->supported_swing_modes_; }
 
-  float get_visual_min_temperature() const { return visual_min_temperature_; }
-  void set_visual_min_temperature(float visual_min_temperature) { visual_min_temperature_ = visual_min_temperature; }
-  float get_visual_max_temperature() const { return visual_max_temperature_; }
-  void set_visual_max_temperature(float visual_max_temperature) { visual_max_temperature_ = visual_max_temperature; }
-  float get_visual_target_temperature_step() const { return visual_target_temperature_step_; }
-  float get_visual_current_temperature_step() const { return visual_current_temperature_step_; }
+  float get_visual_min_temperature() const { return this->visual_min_temperature_; }
+  void set_visual_min_temperature(float visual_min_temperature) {
+    this->visual_min_temperature_ = visual_min_temperature;
+  }
+  float get_visual_max_temperature() const { return this->visual_max_temperature_; }
+  void set_visual_max_temperature(float visual_max_temperature) {
+    this->visual_max_temperature_ = visual_max_temperature;
+  }
+  float get_visual_target_temperature_step() const { return this->visual_target_temperature_step_; }
+  float get_visual_current_temperature_step() const { return this->visual_current_temperature_step_; }
   void set_visual_target_temperature_step(float temperature_step) {
-    visual_target_temperature_step_ = temperature_step;
+    this->visual_target_temperature_step_ = temperature_step;
   }
   void set_visual_current_temperature_step(float temperature_step) {
-    visual_current_temperature_step_ = temperature_step;
+    this->visual_current_temperature_step_ = temperature_step;
   }
   void set_visual_temperature_step(float temperature_step) {
-    visual_target_temperature_step_ = temperature_step;
-    visual_current_temperature_step_ = temperature_step;
+    this->visual_target_temperature_step_ = temperature_step;
+    this->visual_current_temperature_step_ = temperature_step;
   }
   int8_t get_target_temperature_accuracy_decimals() const;
   int8_t get_current_temperature_accuracy_decimals() const;
 
-  float get_visual_min_humidity() const { return visual_min_humidity_; }
-  void set_visual_min_humidity(float visual_min_humidity) { visual_min_humidity_ = visual_min_humidity; }
-  float get_visual_max_humidity() const { return visual_max_humidity_; }
-  void set_visual_max_humidity(float visual_max_humidity) { visual_max_humidity_ = visual_max_humidity; }
+  float get_visual_min_humidity() const { return this->visual_min_humidity_; }
+  void set_visual_min_humidity(float visual_min_humidity) { this->visual_min_humidity_ = visual_min_humidity; }
+  float get_visual_max_humidity() const { return this->visual_max_humidity_; }
+  void set_visual_max_humidity(float visual_max_humidity) { this->visual_max_humidity_ = visual_max_humidity; }
 
  protected:
   void set_mode_support_(climate::ClimateMode mode, bool supported) {
     if (supported) {
-      supported_modes_.insert(mode);
+      this->supported_modes_.insert(mode);
     } else {
-      supported_modes_.erase(mode);
+      this->supported_modes_.erase(mode);
     }
   }
   void set_fan_mode_support_(climate::ClimateFanMode mode, bool supported) {
     if (supported) {
-      supported_fan_modes_.insert(mode);
+      this->supported_fan_modes_.insert(mode);
     } else {
-      supported_fan_modes_.erase(mode);
+      this->supported_fan_modes_.erase(mode);
     }
   }
   void set_swing_mode_support_(climate::ClimateSwingMode mode, bool supported) {
     if (supported) {
-      supported_swing_modes_.insert(mode);
+      this->supported_swing_modes_.insert(mode);
     } else {
-      supported_swing_modes_.erase(mode);
+      this->supported_swing_modes_.erase(mode);
     }
   }
 

esphome/components/climate_ir_lg/climate_ir_lg.cpp

@@ -32,7 +32,7 @@ const uint32_t FAN_MAX = 0x40;
 
 // Temperature
 const uint8_t TEMP_RANGE = TEMP_MAX - TEMP_MIN + 1;
-const uint32_t TEMP_MASK = 0XF00;
+const uint32_t TEMP_MASK = 0xF00;
 const uint32_t TEMP_SHIFT = 8;
 
 const uint16_t BITS = 28;
@@ -43,11 +43,11 @@ void LgIrClimate::transmit_state() {
   // ESP_LOGD(TAG, "climate_lg_ir mode_before_ code: 0x%02X", modeBefore_);
 
   // Set command
-  if (send_swing_cmd_) {
-    send_swing_cmd_ = false;
+  if (this->send_swing_cmd_) {
+    this->send_swing_cmd_ = false;
     remote_state |= COMMAND_SWING;
   } else {
-    bool climate_is_off = (mode_before_ == climate::CLIMATE_MODE_OFF);
+    bool climate_is_off = (this->mode_before_ == climate::CLIMATE_MODE_OFF);
     switch (this->mode) {
       case climate::CLIMATE_MODE_COOL:
         remote_state |= climate_is_off ? COMMAND_ON_COOL : COMMAND_COOL;
@@ -71,7 +71,7 @@ void LgIrClimate::transmit_state() {
     }
   }
 
-  mode_before_ = this->mode;
+  this->mode_before_ = this->mode;
 
   ESP_LOGD(TAG, "climate_lg_ir mode code: 0x%02X", this->mode);
 
@@ -102,7 +102,7 @@ void LgIrClimate::transmit_state() {
     remote_state |= ((temp - 15) << TEMP_SHIFT);
   }
 
-  transmit_(remote_state);
+  this->transmit_(remote_state);
   this->publish_state();
 }
 
@@ -187,7 +187,7 @@ bool LgIrClimate::on_receive(remote_base::RemoteReceiveData data) {
 }
 
 void LgIrClimate::transmit_(uint32_t value) {
-  calc_checksum_(value);
+  this->calc_checksum_(value);
   ESP_LOGD(TAG, "Sending climate_lg_ir code: 0x%02" PRIX32, value);
 
   auto transmit = this->transmitter_->transmit();

esphome/components/climate_ir_lg/climate_ir_lg.h

@@ -21,7 +21,7 @@ class LgIrClimate : public climate_ir::ClimateIR {
 
   /// Override control to change settings of the climate device.
   void control(const climate::ClimateCall &call) override {
-    send_swing_cmd_ = call.get_swing_mode().has_value();
+    this->send_swing_cmd_ = call.get_swing_mode().has_value();
     // swing resets after unit powered off
     if (call.get_mode().has_value() && *call.get_mode() == climate::CLIMATE_MODE_OFF)
       this->swing_mode = climate::CLIMATE_SWING_OFF;

esphome/components/daikin/daikin.cpp

@@ -65,7 +65,7 @@ void DaikinClimate::transmit_state() {
   transmit.perform();
 }
 
-uint8_t DaikinClimate::operation_mode_() {
+uint8_t DaikinClimate::operation_mode_() const {
   uint8_t operating_mode = DAIKIN_MODE_ON;
   switch (this->mode) {
     case climate::CLIMATE_MODE_COOL:
@@ -92,9 +92,12 @@ uint8_t DaikinClimate::operation_mode_() {
   return operating_mode;
 }
 
-uint16_t DaikinClimate::fan_speed_() {
+uint16_t DaikinClimate::fan_speed_() const {
   uint16_t fan_speed;
   switch (this->fan_mode.value()) {
+    case climate::CLIMATE_FAN_QUIET:
+      fan_speed = DAIKIN_FAN_SILENT << 8;
+      break;
     case climate::CLIMATE_FAN_LOW:
       fan_speed = DAIKIN_FAN_1 << 8;
       break;
@@ -126,12 +129,11 @@ uint16_t DaikinClimate::fan_speed_() {
   return fan_speed;
 }
 
-uint8_t DaikinClimate::temperature_() {
+uint8_t DaikinClimate::temperature_() const {
   // Force special temperatures depending on the mode
   switch (this->mode) {
     case climate::CLIMATE_MODE_FAN_ONLY:
       return 0x32;
-    case climate::CLIMATE_MODE_HEAT_COOL:
     case climate::CLIMATE_MODE_DRY:
       return 0xc0;
     default:
@@ -148,19 +150,25 @@ bool DaikinClimate::parse_state_frame_(const uint8_t frame[]) {
   if (frame[DAIKIN_STATE_FRAME_SIZE - 1] != checksum)
     return false;
   uint8_t mode = frame[5];
+  // Temperature is given in degrees celcius * 2
+  // only update for states that use the temperature
+  uint8_t temperature = frame[6];
   if (mode & DAIKIN_MODE_ON) {
     switch (mode & 0xF0) {
       case DAIKIN_MODE_COOL:
         this->mode = climate::CLIMATE_MODE_COOL;
+        this->target_temperature = static_cast<float>(temperature * 0.5f);
         break;
       case DAIKIN_MODE_DRY:
         this->mode = climate::CLIMATE_MODE_DRY;
         break;
       case DAIKIN_MODE_HEAT:
         this->mode = climate::CLIMATE_MODE_HEAT;
+        this->target_temperature = static_cast<float>(temperature * 0.5f);
         break;
       case DAIKIN_MODE_AUTO:
         this->mode = climate::CLIMATE_MODE_HEAT_COOL;
+        this->target_temperature = static_cast<float>(temperature * 0.5f);
         break;
       case DAIKIN_MODE_FAN:
         this->mode = climate::CLIMATE_MODE_FAN_ONLY;
@@ -169,10 +177,6 @@ bool DaikinClimate::parse_state_frame_(const uint8_t frame[]) {
   } else {
     this->mode = climate::CLIMATE_MODE_OFF;
   }
-  uint8_t temperature = frame[6];
-  if (!(temperature & 0xC0)) {
-    this->target_temperature = temperature >> 1;
-  }
   uint8_t fan_mode = frame[8];
   uint8_t swing_mode = frame[9];
   if (fan_mode & 0xF && swing_mode & 0xF) {
@@ -187,7 +191,6 @@ bool DaikinClimate::parse_state_frame_(const uint8_t frame[]) {
   switch (fan_mode & 0xF0) {
     case DAIKIN_FAN_1:
     case DAIKIN_FAN_2:
-    case DAIKIN_FAN_SILENT:
       this->fan_mode = climate::CLIMATE_FAN_LOW;
       break;
     case DAIKIN_FAN_3:
@@ -200,6 +203,9 @@ bool DaikinClimate::parse_state_frame_(const uint8_t frame[]) {
     case DAIKIN_FAN_AUTO:
       this->fan_mode = climate::CLIMATE_FAN_AUTO;
       break;
+    case DAIKIN_FAN_SILENT:
+      this->fan_mode = climate::CLIMATE_FAN_QUIET;
+      break;
   }
   this->publish_state();
   return true;

esphome/components/daikin/daikin.h

@@ -44,17 +44,17 @@ class DaikinClimate : public climate_ir::ClimateIR {
  public:
   DaikinClimate()
       : climate_ir::ClimateIR(DAIKIN_TEMP_MIN, DAIKIN_TEMP_MAX, 1.0f, true, true,
-                              {climate::CLIMATE_FAN_AUTO, climate::CLIMATE_FAN_LOW, climate::CLIMATE_FAN_MEDIUM,
-                               climate::CLIMATE_FAN_HIGH},
+                              {climate::CLIMATE_FAN_QUIET, climate::CLIMATE_FAN_AUTO, climate::CLIMATE_FAN_LOW,
+                               climate::CLIMATE_FAN_MEDIUM, climate::CLIMATE_FAN_HIGH},
                               {climate::CLIMATE_SWING_OFF, climate::CLIMATE_SWING_VERTICAL,
                                climate::CLIMATE_SWING_HORIZONTAL, climate::CLIMATE_SWING_BOTH}) {}
 
  protected:
   // Transmit via IR the state of this climate controller.
   void transmit_state() override;
-  uint8_t operation_mode_();
-  uint16_t fan_speed_();
-  uint8_t temperature_();
+  uint8_t operation_mode_() const;
+  uint16_t fan_speed_() const;
+  uint8_t temperature_() const;
   // Handle received IR Buffer
   bool on_receive(remote_base::RemoteReceiveData data) override;
   bool parse_state_frame_(const uint8_t frame[]);

esphome/components/debug/debug_component.cpp

@@ -1,6 +1,7 @@
 #include "debug_component.h"
 
 #include <algorithm>
+#include "esphome/core/application.h"
 #include "esphome/core/log.h"
 #include "esphome/core/hal.h"
 #include "esphome/core/helpers.h"
@@ -25,6 +26,7 @@ void DebugComponent::dump_config() {
 #ifdef USE_SENSOR
   LOG_SENSOR("  ", "Free space on heap", this->free_sensor_);
   LOG_SENSOR("  ", "Largest free heap block", this->block_sensor_);
+  LOG_SENSOR("  ", "CPU frequency", this->cpu_frequency_sensor_);
 #if defined(USE_ESP8266) && USE_ARDUINO_VERSION_CODE >= VERSION_CODE(2, 5, 2)
   LOG_SENSOR("  ", "Heap fragmentation", this->fragmentation_sensor_);
 #endif  // defined(USE_ESP8266) && USE_ARDUINO_VERSION_CODE >= VERSION_CODE(2, 5, 2)
@@ -86,6 +88,9 @@ void DebugComponent::update() {
     this->loop_time_sensor_->publish_state(this->max_loop_time_);
     this->max_loop_time_ = 0;
   }
+  if (this->cpu_frequency_sensor_ != nullptr) {
+    this->cpu_frequency_sensor_->publish_state(arch_get_cpu_freq_hz());
+  }
 
 #endif  // USE_SENSOR
   update_platform_();

esphome/components/debug/debug_component.h

@@ -34,8 +34,12 @@ class DebugComponent : public PollingComponent {
 #endif
   void set_loop_time_sensor(sensor::Sensor *loop_time_sensor) { loop_time_sensor_ = loop_time_sensor; }
 #ifdef USE_ESP32
+  void on_shutdown() override;
   void set_psram_sensor(sensor::Sensor *psram_sensor) { this->psram_sensor_ = psram_sensor; }
 #endif  // USE_ESP32
+  void set_cpu_frequency_sensor(sensor::Sensor *cpu_frequency_sensor) {
+    this->cpu_frequency_sensor_ = cpu_frequency_sensor;
+  }
 #endif  // USE_SENSOR
  protected:
   uint32_t free_heap_{};
@@ -53,6 +57,7 @@ class DebugComponent : public PollingComponent {
 #ifdef USE_ESP32
   sensor::Sensor *psram_sensor_{nullptr};
 #endif  // USE_ESP32
+  sensor::Sensor *cpu_frequency_sensor_{nullptr};
 #endif  // USE_SENSOR
 
 #ifdef USE_ESP32
@@ -75,6 +80,7 @@ class DebugComponent : public PollingComponent {
 #endif  // USE_TEXT_SENSOR
 
   std::string get_reset_reason_();
+  std::string get_wakeup_cause_();
   uint32_t get_free_heap_();
   void get_device_info_(std::string &device_info);
   void update_platform_();

esphome/components/debug/debug_esp32.cpp

@@ -1,25 +1,18 @@
 #include "debug_component.h"
+
 #ifdef USE_ESP32
+#include "esphome/core/application.h"
 #include "esphome/core/log.h"
+#include "esphome/core/hal.h"
+#include <esp_sleep.h>
 
 #include <esp_heap_caps.h>
 #include <esp_system.h>
 #include <esp_chip_info.h>
 #include <esp_partition.h>
 
-#if defined(USE_ESP32_VARIANT_ESP32)
-#include <esp32/rom/rtc.h>
-#elif defined(USE_ESP32_VARIANT_ESP32C3)
-#include <esp32c3/rom/rtc.h>
-#elif defined(USE_ESP32_VARIANT_ESP32C6)
-#include <esp32c6/rom/rtc.h>
-#elif defined(USE_ESP32_VARIANT_ESP32S2)
-#include <esp32s2/rom/rtc.h>
-#elif defined(USE_ESP32_VARIANT_ESP32S3)
-#include <esp32s3/rom/rtc.h>
-#elif defined(USE_ESP32_VARIANT_ESP32H2)
-#include <esp32h2/rom/rtc.h>
-#endif
+#include <map>
+
 #ifdef USE_ARDUINO
 #include <Esp.h>
 #endif
@@ -29,6 +22,90 @@ namespace debug {
 
 static const char *const TAG = "debug";
 
+// index by values returned by esp_reset_reason
+
+static const char *const RESET_REASONS[] = {
+    "unknown source",
+    "power-on event",
+    "external pin",
+    "software via esp_restart",
+    "exception/panic",
+    "interrupt watchdog",
+    "task watchdog",
+    "other watchdogs",
+    "exiting deep sleep mode",
+    "brownout",
+    "SDIO",
+    "USB peripheral",
+    "JTAG",
+    "efuse error",
+    "power glitch detected",
+    "CPU lock up",
+};
+
+static const char *const REBOOT_KEY = "reboot_source";
+static const size_t REBOOT_MAX_LEN = 24;
+
+// on shutdown, store the source of the reboot request
+void DebugComponent::on_shutdown() {
+  auto *component = App.get_current_component();
+  char buffer[REBOOT_MAX_LEN]{};
+  auto pref = global_preferences->make_preference(REBOOT_MAX_LEN, fnv1_hash(REBOOT_KEY + App.get_name()));
+  if (component != nullptr) {
+    strncpy(buffer, component->get_component_source(), REBOOT_MAX_LEN - 1);
+  }
+  ESP_LOGD(TAG, "Storing reboot source: %s", buffer);
+  pref.save(&buffer);
+  global_preferences->sync();
+}
+
+std::string DebugComponent::get_reset_reason_() {
+  std::string reset_reason;
+  unsigned reason = esp_reset_reason();
+  if (reason < sizeof(RESET_REASONS) / sizeof(RESET_REASONS[0])) {
+    reset_reason = RESET_REASONS[reason];
+    if (reason == ESP_RST_SW) {
+      auto pref = global_preferences->make_preference(REBOOT_MAX_LEN, fnv1_hash(REBOOT_KEY + App.get_name()));
+      char buffer[REBOOT_MAX_LEN]{};
+      if (pref.load(&buffer)) {
+        reset_reason = "Reboot request from " + std::string(buffer);
+      }
+    }
+  } else {
+    reset_reason = "unknown source";
+  }
+  ESP_LOGD(TAG, "Reset Reason: %s", reset_reason.c_str());
+  return reset_reason;
+}
+
+static const char *const WAKEUP_CAUSES[] = {
+    "undefined",
+    "undefined",
+    "external signal using RTC_IO",
+    "external signal using RTC_CNTL",
+    "timer",
+    "touchpad",
+    "ULP program",
+    "GPIO",
+    "UART",
+    "WIFI",
+    "COCPU int",
+    "COCPU crash",
+    "BT",
+};
+
+std::string DebugComponent::get_wakeup_cause_() {
+  const char *wake_reason;
+  unsigned reason = esp_sleep_get_wakeup_cause();
+  if (reason < sizeof(WAKEUP_CAUSES) / sizeof(WAKEUP_CAUSES[0])) {
+    wake_reason = WAKEUP_CAUSES[reason];
+  } else {
+    wake_reason = "unknown source";
+  }
+  ESP_LOGD(TAG, "Wakeup Reason: %s", wake_reason);
+  return wake_reason;
+}
+
 void DebugComponent::log_partition_info_() {
   ESP_LOGCONFIG(TAG, "Partition table:");
   ESP_LOGCONFIG(TAG, "  %-12s %-4s %-8s %-10s %-10s", "Name", "Type", "Subtype", "Address", "Size");
@@ -42,171 +119,16 @@ void DebugComponent::log_partition_info_() {
   esp_partition_iterator_release(it);
 }
 
-std::string DebugComponent::get_reset_reason_() {
-  std::string reset_reason;
-  switch (esp_reset_reason()) {
-    case ESP_RST_POWERON:
-      reset_reason = "Reset due to power-on event";
-      break;
-    case ESP_RST_EXT:
-      reset_reason = "Reset by external pin";
-      break;
-    case ESP_RST_SW:
-      reset_reason = "Software reset via esp_restart";
-      break;
-    case ESP_RST_PANIC:
-      reset_reason = "Software reset due to exception/panic";
-      break;
-    case ESP_RST_INT_WDT:
-      reset_reason = "Reset (software or hardware) due to interrupt watchdog";
-      break;
-    case ESP_RST_TASK_WDT:
-      reset_reason = "Reset due to task watchdog";
-      break;
-    case ESP_RST_WDT:
-      reset_reason = "Reset due to other watchdogs";
-      break;
-    case ESP_RST_DEEPSLEEP:
-      reset_reason = "Reset after exiting deep sleep mode";
-      break;
-    case ESP_RST_BROWNOUT:
-      reset_reason = "Brownout reset (software or hardware)";
-      break;
-    case ESP_RST_SDIO:
-      reset_reason = "Reset over SDIO";
-      break;
-#ifdef USE_ESP32_VARIANT_ESP32
-#if (ESP_IDF_VERSION >= ESP_IDF_VERSION_VAL(5, 1, 4))
-    case ESP_RST_USB:
-      reset_reason = "Reset by USB peripheral";
-      break;
-    case ESP_RST_JTAG:
-      reset_reason = "Reset by JTAG";
-      break;
-    case ESP_RST_EFUSE:
-      reset_reason = "Reset due to efuse error";
-      break;
-    case ESP_RST_PWR_GLITCH:
-      reset_reason = "Reset due to power glitch detected";
-      break;
-    case ESP_RST_CPU_LOCKUP:
-      reset_reason = "Reset due to CPU lock up (double exception)";
-      break;
-#endif        // ESP_IDF_VERSION >= ESP_IDF_VERSION_VAL(5, 1, 4)
-#endif        // USE_ESP32_VARIANT_ESP32
-    default:  // Includes ESP_RST_UNKNOWN
-      switch (rtc_get_reset_reason(0)) {
-        case POWERON_RESET:
-          reset_reason = "Power On Reset";
-          break;
-#if defined(USE_ESP32_VARIANT_ESP32)
-        case SW_RESET:
-#elif defined(USE_ESP32_VARIANT_ESP32C3) || defined(USE_ESP32_VARIANT_ESP32S2) || \
-    defined(USE_ESP32_VARIANT_ESP32S3) || defined(USE_ESP32_VARIANT_ESP32C6)
-        case RTC_SW_SYS_RESET:
-#endif
-          reset_reason = "Software Reset Digital Core";
-          break;
-#if defined(USE_ESP32_VARIANT_ESP32)
-        case OWDT_RESET:
-          reset_reason = "Watch Dog Reset Digital Core";
-          break;
-#endif
-        case DEEPSLEEP_RESET:
-          reset_reason = "Deep Sleep Reset Digital Core";
-          break;
-#if defined(USE_ESP32_VARIANT_ESP32)
-        case SDIO_RESET:
-          reset_reason = "SLC Module Reset Digital Core";
-          break;
-#endif
-        case TG0WDT_SYS_RESET:
-          reset_reason = "Timer Group 0 Watch Dog Reset Digital Core";
-          break;
-        case TG1WDT_SYS_RESET:
-          reset_reason = "Timer Group 1 Watch Dog Reset Digital Core";
-          break;
-        case RTCWDT_SYS_RESET:
-          reset_reason = "RTC Watch Dog Reset Digital Core";
-          break;
-#if !defined(USE_ESP32_VARIANT_ESP32C6) && !defined(USE_ESP32_VARIANT_ESP32H2)
-        case INTRUSION_RESET:
-          reset_reason = "Intrusion Reset CPU";
-          break;
-#endif
-#if defined(USE_ESP32_VARIANT_ESP32)
-        case TGWDT_CPU_RESET:
-          reset_reason = "Timer Group Reset CPU";
-          break;
-#elif defined(USE_ESP32_VARIANT_ESP32C3) || defined(USE_ESP32_VARIANT_ESP32S2) || \
-    defined(USE_ESP32_VARIANT_ESP32S3) || defined(USE_ESP32_VARIANT_ESP32C6)
-        case TG0WDT_CPU_RESET:
-          reset_reason = "Timer Group 0 Reset CPU";
-          break;
-#endif
-#if defined(USE_ESP32_VARIANT_ESP32)
-        case SW_CPU_RESET:
-#elif defined(USE_ESP32_VARIANT_ESP32C3) || defined(USE_ESP32_VARIANT_ESP32S2) || \
-    defined(USE_ESP32_VARIANT_ESP32S3) || defined(USE_ESP32_VARIANT_ESP32C6)
-        case RTC_SW_CPU_RESET:
-#endif
-          reset_reason = "Software Reset CPU";
-          break;
-        case RTCWDT_CPU_RESET:
-          reset_reason = "RTC Watch Dog Reset CPU";
-          break;
-#if defined(USE_ESP32_VARIANT_ESP32)
-        case EXT_CPU_RESET:
-          reset_reason = "External CPU Reset";
-          break;
-#endif
-        case RTCWDT_BROWN_OUT_RESET:
-          reset_reason = "Voltage Unstable Reset";
-          break;
-        case RTCWDT_RTC_RESET:
-          reset_reason = "RTC Watch Dog Reset Digital Core And RTC Module";
-          break;
-#if defined(USE_ESP32_VARIANT_ESP32C3) || defined(USE_ESP32_VARIANT_ESP32S2) || defined(USE_ESP32_VARIANT_ESP32S3) || \
-    defined(USE_ESP32_VARIANT_ESP32C6)
-        case TG1WDT_CPU_RESET:
-          reset_reason = "Timer Group 1 Reset CPU";
-          break;
-        case SUPER_WDT_RESET:
-          reset_reason = "Super Watchdog Reset Digital Core And RTC Module";
-          break;
-        case EFUSE_RESET:
-          reset_reason = "eFuse Reset Digital Core";
-          break;
-#endif
-#if defined(USE_ESP32_VARIANT_ESP32C3) || defined(USE_ESP32_VARIANT_ESP32S2) || defined(USE_ESP32_VARIANT_ESP32S3)
-        case GLITCH_RTC_RESET:
-          reset_reason = "Glitch Reset Digital Core And RTC Module";
-          break;
-#endif
-#if defined(USE_ESP32_VARIANT_ESP32C3) || defined(USE_ESP32_VARIANT_ESP32S3) || defined(USE_ESP32_VARIANT_ESP32C6)
-        case USB_UART_CHIP_RESET:
-          reset_reason = "USB UART Reset Digital Core";
-          break;
-        case USB_JTAG_CHIP_RESET:
-          reset_reason = "USB JTAG Reset Digital Core";
-          break;
-#endif
-#if defined(USE_ESP32_VARIANT_ESP32C3) || defined(USE_ESP32_VARIANT_ESP32S3)
-        case POWER_GLITCH_RESET:
-          reset_reason = "Power Glitch Reset Digital Core And RTC Module";
-          break;
-#endif
-        default:
-          reset_reason = "Unknown Reset Reason";
-      }
-      break;
-  }
-  ESP_LOGD(TAG, "Reset Reason: %s", reset_reason.c_str());
-  return reset_reason;
-}
-
 uint32_t DebugComponent::get_free_heap_() { return heap_caps_get_free_size(MALLOC_CAP_INTERNAL); }
 
+static const std::map<int, const char *> CHIP_FEATURES = {
+    {CHIP_FEATURE_BLE, "BLE"},
+    {CHIP_FEATURE_BT, "BT"},
+    {CHIP_FEATURE_EMB_FLASH, "EMB Flash"},
+    {CHIP_FEATURE_EMB_PSRAM, "EMB PSRAM"},
+    {CHIP_FEATURE_WIFI_BGN, "2.4GHz WiFi"},
+};
+
 void DebugComponent::get_device_info_(std::string &device_info) {
 #if defined(USE_ARDUINO)
   const char *flash_mode;
@@ -242,44 +164,16 @@ void DebugComponent::get_device_info_(std::string &device_info) {
 
   esp_chip_info_t info;
   esp_chip_info(&info);
-  const char *model;
-#if defined(USE_ESP32_VARIANT_ESP32)
-  model = "ESP32";
-#elif defined(USE_ESP32_VARIANT_ESP32C3)
-  model = "ESP32-C3";
-#elif defined(USE_ESP32_VARIANT_ESP32C6)
-  model = "ESP32-C6";
-#elif defined(USE_ESP32_VARIANT_ESP32S2)
-  model = "ESP32-S2";
-#elif defined(USE_ESP32_VARIANT_ESP32S3)
-  model = "ESP32-S3";
-#elif defined(USE_ESP32_VARIANT_ESP32H2)
-  model = "ESP32-H2";
-#else
-  model = "UNKNOWN";
-#endif
+  const char *model = ESPHOME_VARIANT;
   std::string features;
-  if (info.features & CHIP_FEATURE_EMB_FLASH) {
-    features += "EMB_FLASH,";
-    info.features &= ~CHIP_FEATURE_EMB_FLASH;
+  for (auto feature : CHIP_FEATURES) {
+    if (info.features & feature.first) {
+      features += feature.second;
+      features += ", ";
+      info.features &= ~feature.first;
+    }
   }
-  if (info.features & CHIP_FEATURE_WIFI_BGN) {
-    features += "WIFI_BGN,";
-    info.features &= ~CHIP_FEATURE_WIFI_BGN;
-  }
-  if (info.features & CHIP_FEATURE_BLE) {
-    features += "BLE,";
-    info.features &= ~CHIP_FEATURE_BLE;
-  }
-  if (info.features & CHIP_FEATURE_BT) {
-    features += "BT,";
-    info.features &= ~CHIP_FEATURE_BT;
-  }
-  if (info.features & CHIP_FEATURE_EMB_PSRAM) {
-    features += "EMB_PSRAM,";
-    info.features &= ~CHIP_FEATURE_EMB_PSRAM;
-  }
-  if (info.features)
+  if (info.features != 0)
     features += "Other:" + format_hex(info.features);
   ESP_LOGD(TAG, "Chip: Model=%s, Features=%s Cores=%u, Revision=%u", model, features.c_str(), info.cores,
            info.revision);
@@ -289,6 +183,8 @@ void DebugComponent::get_device_info_(std::string &device_info) {
   device_info += features;
   device_info += " Cores:" + to_string(info.cores);
   device_info += " Revision:" + to_string(info.revision);
+  device_info += str_sprintf("|CPU Frequency: %" PRIu32 " MHz", arch_get_cpu_freq_hz() / 1000000);
+  ESP_LOGD(TAG, "CPU Frequency: %" PRIu32 " MHz", arch_get_cpu_freq_hz() / 1000000);
 
   // Framework detection
   device_info += "|Framework: ";
@@ -315,48 +211,7 @@ void DebugComponent::get_device_info_(std::string &device_info) {
   device_info += "|Reset: ";
   device_info += get_reset_reason_();
 
-  const char *wakeup_reason;
-  switch (rtc_get_wakeup_cause()) {
-    case NO_SLEEP:
-      wakeup_reason = "No Sleep";
-      break;
-    case EXT_EVENT0_TRIG:
-      wakeup_reason = "External Event 0";
-      break;
-    case EXT_EVENT1_TRIG:
-      wakeup_reason = "External Event 1";
-      break;
-    case GPIO_TRIG:
-      wakeup_reason = "GPIO";
-      break;
-    case TIMER_EXPIRE:
-      wakeup_reason = "Wakeup Timer";
-      break;
-    case SDIO_TRIG:
-      wakeup_reason = "SDIO";
-      break;
-    case MAC_TRIG:
-      wakeup_reason = "MAC";
-      break;
-    case UART0_TRIG:
-      wakeup_reason = "UART0";
-      break;
-    case UART1_TRIG:
-      wakeup_reason = "UART1";
-      break;
-    case TOUCH_TRIG:
-      wakeup_reason = "Touch";
-      break;
-    case SAR_TRIG:
-      wakeup_reason = "SAR";
-      break;
-    case BT_TRIG:
-      wakeup_reason = "BT";
-      break;
-    default:
-      wakeup_reason = "Unknown";
-  }
-  ESP_LOGD(TAG, "Wakeup Reason: %s", wakeup_reason);
+  std::string wakeup_reason = this->get_wakeup_cause_();
   device_info += "|Wakeup: ";
   device_info += wakeup_reason;
 }

esphome/components/debug/sensor.py

@@ -1,5 +1,6 @@
 import esphome.codegen as cg
 from esphome.components import sensor
+from esphome.components.esp32 import CONF_CPU_FREQUENCY
 import esphome.config_validation as cv
 from esphome.const import (
     CONF_BLOCK,
@@ -10,6 +11,7 @@ from esphome.const import (
     ICON_COUNTER,
     ICON_TIMER,
     UNIT_BYTES,
+    UNIT_HERTZ,
     UNIT_MILLISECOND,
     UNIT_PERCENT,
 )
@@ -60,6 +62,14 @@ CONFIG_SCHEMA = {
             entity_category=ENTITY_CATEGORY_DIAGNOSTIC,
         ),
     ),
+    cv.Optional(CONF_CPU_FREQUENCY): cv.All(
+        sensor.sensor_schema(
+            unit_of_measurement=UNIT_HERTZ,
+            icon="mdi:speedometer",
+            accuracy_decimals=0,
+            entity_category=ENTITY_CATEGORY_DIAGNOSTIC,
+        ),
+    ),
 }
 
 
@@ -85,3 +95,7 @@ async def to_code(config):
     if psram_conf := config.get(CONF_PSRAM):
         sens = await sensor.new_sensor(psram_conf)
         cg.add(debug_component.set_psram_sensor(sens))
+
+    if cpu_freq_conf := config.get(CONF_CPU_FREQUENCY):
+        sens = await sensor.new_sensor(cpu_freq_conf)
+        cg.add(debug_component.set_cpu_frequency_sensor(sens))

esphome/components/deep_sleep/deep_sleep_esp32.cpp

@@ -31,9 +31,12 @@ void DeepSleepComponent::set_wakeup_pin_mode(WakeupPinMode wakeup_pin_mode) {
 #if !defined(USE_ESP32_VARIANT_ESP32C3) && !defined(USE_ESP32_VARIANT_ESP32C6)
 void DeepSleepComponent::set_ext1_wakeup(Ext1Wakeup ext1_wakeup) { this->ext1_wakeup_ = ext1_wakeup; }
 
+#if !defined(USE_ESP32_VARIANT_ESP32H2)
 void DeepSleepComponent::set_touch_wakeup(bool touch_wakeup) { this->touch_wakeup_ = touch_wakeup; }
 #endif
 
+#endif
+
 void DeepSleepComponent::set_run_duration(WakeupCauseToRunDuration wakeup_cause_to_run_duration) {
   wakeup_cause_to_run_duration_ = wakeup_cause_to_run_duration;
 }
@@ -65,7 +68,7 @@ bool DeepSleepComponent::prepare_to_sleep_() {
 }
 
 void DeepSleepComponent::deep_sleep_() {
-#if !defined(USE_ESP32_VARIANT_ESP32C3) && !defined(USE_ESP32_VARIANT_ESP32C6)
+#if !defined(USE_ESP32_VARIANT_ESP32C3) && !defined(USE_ESP32_VARIANT_ESP32C6) && !defined(USE_ESP32_VARIANT_ESP32H2)
   if (this->sleep_duration_.has_value())
     esp_sleep_enable_timer_wakeup(*this->sleep_duration_);
   if (this->wakeup_pin_ != nullptr) {
@@ -84,6 +87,15 @@ void DeepSleepComponent::deep_sleep_() {
     esp_sleep_pd_config(ESP_PD_DOMAIN_RTC_PERIPH, ESP_PD_OPTION_ON);
   }
 #endif
+
+#if defined(USE_ESP32_VARIANT_ESP32H2)
+  if (this->sleep_duration_.has_value())
+    esp_sleep_enable_timer_wakeup(*this->sleep_duration_);
+  if (this->ext1_wakeup_.has_value()) {
+    esp_sleep_enable_ext1_wakeup(this->ext1_wakeup_->mask, this->ext1_wakeup_->wakeup_mode);
+  }
+#endif
+
 #if defined(USE_ESP32_VARIANT_ESP32C3) || defined(USE_ESP32_VARIANT_ESP32C6)
   if (this->sleep_duration_.has_value())
     esp_sleep_enable_timer_wakeup(*this->sleep_duration_);

esphome/components/display/rect.cpp

@@ -69,21 +69,16 @@ bool Rect::inside(int16_t test_x, int16_t test_y, bool absolute) const {  // NOL
     return true;
   }
   if (absolute) {
-    return ((test_x >= this->x) && (test_x <= this->x2()) && (test_y >= this->y) && (test_y <= this->y2()));
-  } else {
-    return ((test_x >= 0) && (test_x <= this->w) && (test_y >= 0) && (test_y <= this->h));
+    return test_x >= this->x && test_x < this->x2() && test_y >= this->y && test_y < this->y2();
   }
+  return test_x >= 0 && test_x < this->w && test_y >= 0 && test_y < this->h;
 }
 
-bool Rect::inside(Rect rect, bool absolute) const {
+bool Rect::inside(Rect rect) const {
   if (!this->is_set() || !rect.is_set()) {
     return true;
   }
-  if (absolute) {
-    return ((rect.x <= this->x2()) && (rect.x2() >= this->x) && (rect.y <= this->y2()) && (rect.y2() >= this->y));
-  } else {
-    return ((rect.x <= this->w) && (rect.w >= 0) && (rect.y <= this->h) && (rect.h >= 0));
-  }
+  return this->x2() >= rect.x && this->x <= rect.x2() && this->y2() >= rect.y && this->y <= rect.y2();
 }
 
 void Rect::info(const std::string &prefix) {

esphome/components/display/rect.h

@@ -26,7 +26,7 @@ class Rect {
   void extend(Rect rect);
   void shrink(Rect rect);
 
-  bool inside(Rect rect, bool absolute = true) const;
+  bool inside(Rect rect) const;
   bool inside(int16_t test_x, int16_t test_y, bool absolute = true) const;
   bool equal(Rect rect) const;
   void info(const std::string &prefix = "rect info:");

esphome/components/ens160_base/ens160_base.cpp

@@ -187,7 +187,7 @@ void ENS160Component::update() {
       }
       return;
     case INVALID_OUTPUT:
-      ESP_LOGE(TAG, "ENS160 Invalid Status - No Invalid Output");
+      ESP_LOGE(TAG, "ENS160 Invalid Status - No valid output");
       this->status_set_warning();
       return;
   }

esphome/components/esp32/__init__.py

@@ -1,4 +1,5 @@
 from dataclasses import dataclass
+import itertools
 import logging
 import os
 from pathlib import Path
@@ -37,6 +38,7 @@ from esphome.const import (
     __version__,
 )
 from esphome.core import CORE, HexInt, TimePeriod
+from esphome.cpp_generator import RawExpression
 import esphome.final_validate as fv
 from esphome.helpers import copy_file_if_changed, mkdir_p, write_file_if_changed
 
@@ -54,6 +56,12 @@ from .const import (  # noqa
     KEY_SUBMODULES,
     KEY_VARIANT,
     VARIANT_ESP32,
+    VARIANT_ESP32C2,
+    VARIANT_ESP32C3,
+    VARIANT_ESP32C6,
+    VARIANT_ESP32H2,
+    VARIANT_ESP32S2,
+    VARIANT_ESP32S3,
     VARIANT_FRIENDLY,
     VARIANTS,
 )
@@ -70,7 +78,43 @@ CONF_RELEASE = "release"
 CONF_ENABLE_IDF_EXPERIMENTAL_FEATURES = "enable_idf_experimental_features"
 
 
+def get_cpu_frequencies(*frequencies):
+    return [str(x) + "MHZ" for x in frequencies]
+
+
+CPU_FREQUENCIES = {
+    VARIANT_ESP32: get_cpu_frequencies(80, 160, 240),
+    VARIANT_ESP32S2: get_cpu_frequencies(80, 160, 240),
+    VARIANT_ESP32S3: get_cpu_frequencies(80, 160, 240),
+    VARIANT_ESP32C2: get_cpu_frequencies(80, 120),
+    VARIANT_ESP32C3: get_cpu_frequencies(80, 160),
+    VARIANT_ESP32C6: get_cpu_frequencies(80, 120, 160),
+    VARIANT_ESP32H2: get_cpu_frequencies(16, 32, 48, 64, 96),
+}
+
+# Make sure not missed here if a new variant added.
+assert all(v in CPU_FREQUENCIES for v in VARIANTS)
+
+FULL_CPU_FREQUENCIES = set(itertools.chain.from_iterable(CPU_FREQUENCIES.values()))
+
+
 def set_core_data(config):
+    cpu_frequency = config.get(CONF_CPU_FREQUENCY, None)
+    variant = config[CONF_VARIANT]
+    # if not specified in config, set to 160MHz if supported, the fastest otherwise
+    if cpu_frequency is None:
+        choices = CPU_FREQUENCIES[variant]
+        if "160MHZ" in choices:
+            cpu_frequency = "160MHZ"
+        else:
+            cpu_frequency = choices[-1]
+        config[CONF_CPU_FREQUENCY] = cpu_frequency
+    elif cpu_frequency not in CPU_FREQUENCIES[variant]:
+        raise cv.Invalid(
+            f"Invalid CPU frequency '{cpu_frequency}' for {config[CONF_VARIANT]}",
+            path=[CONF_CPU_FREQUENCY],
+        )
+
     CORE.data[KEY_ESP32] = {}
     CORE.data[KEY_CORE][KEY_TARGET_PLATFORM] = PLATFORM_ESP32
     conf = config[CONF_FRAMEWORK]
@@ -83,6 +127,7 @@ def set_core_data(config):
     CORE.data[KEY_CORE][KEY_FRAMEWORK_VERSION] = cv.Version.parse(
         config[CONF_FRAMEWORK][CONF_VERSION]
     )
+
     CORE.data[KEY_ESP32][KEY_BOARD] = config[CONF_BOARD]
     CORE.data[KEY_ESP32][KEY_VARIANT] = config[CONF_VARIANT]
     CORE.data[KEY_ESP32][KEY_EXTRA_BUILD_FILES] = {}
@@ -553,11 +598,15 @@ FLASH_SIZES = [
 ]
 
 CONF_FLASH_SIZE = "flash_size"
+CONF_CPU_FREQUENCY = "cpu_frequency"
 CONF_PARTITIONS = "partitions"
 CONFIG_SCHEMA = cv.All(
     cv.Schema(
         {
             cv.Required(CONF_BOARD): cv.string_strict,
+            cv.Optional(CONF_CPU_FREQUENCY): cv.one_of(
+                *FULL_CPU_FREQUENCIES, upper=True
+            ),
             cv.Optional(CONF_FLASH_SIZE, default="4MB"): cv.one_of(
                 *FLASH_SIZES, upper=True
             ),
@@ -598,6 +647,7 @@ async def to_code(config):
         os.path.join(os.path.dirname(__file__), "post_build.py.script"),
     )
 
+    freq = config[CONF_CPU_FREQUENCY][:-3]
     if conf[CONF_TYPE] == FRAMEWORK_ESP_IDF:
         cg.add_platformio_option("framework", "espidf")
         cg.add_build_flag("-DUSE_ESP_IDF")
@@ -631,6 +681,9 @@ async def to_code(config):
         add_idf_sdkconfig_option("CONFIG_ESP_TASK_WDT_CHECK_IDLE_TASK_CPU0", False)
         add_idf_sdkconfig_option("CONFIG_ESP_TASK_WDT_CHECK_IDLE_TASK_CPU1", False)
 
+        # Set default CPU frequency
+        add_idf_sdkconfig_option(f"CONFIG_ESP_DEFAULT_CPU_FREQ_MHZ_{freq}", True)
+
         cg.add_platformio_option("board_build.partitions", "partitions.csv")
         if CONF_PARTITIONS in config:
             add_extra_build_file(
@@ -696,6 +749,7 @@ async def to_code(config):
                 f"VERSION_CODE({framework_ver.major}, {framework_ver.minor}, {framework_ver.patch})"
             ),
         )
+        cg.add(RawExpression(f"setCpuFrequencyMhz({freq})"))
 
 
 APP_PARTITION_SIZES = {

esphome/components/esp32/core.cpp

@@ -13,11 +13,13 @@
 #include <hal/cpu_hal.h>
 
 #ifdef USE_ARDUINO
-#include <esp32-hal.h>
-#endif
+#include <Esp.h>
+#else
+#include <esp_clk_tree.h>
 
 void setup();
 void loop();
+#endif
 
 namespace esphome {
 
@@ -59,9 +61,13 @@ uint32_t arch_get_cpu_cycle_count() { return esp_cpu_get_cycle_count(); }
 uint32_t arch_get_cpu_cycle_count() { return cpu_hal_get_cycle_count(); }
 #endif
 uint32_t arch_get_cpu_freq_hz() {
-  rtc_cpu_freq_config_t config;
-  rtc_clk_cpu_freq_get_config(&config);
-  return config.freq_mhz * 1000000U;
+  uint32_t freq = 0;
+#ifdef USE_ESP_IDF
+  esp_clk_tree_src_get_freq_hz(SOC_MOD_CLK_CPU, ESP_CLK_TREE_SRC_FREQ_PRECISION_CACHED, &freq);
+#elif defined(USE_ARDUINO)
+  freq = ESP.getCpuFreqMHz() * 1000000;
+#endif
+  return freq;
 }
 
 #ifdef USE_ESP_IDF

esphome/components/esp32_ble/ble.cpp

@@ -2,10 +2,6 @@
 
 #include "ble.h"
 
-#ifdef USE_ESP32_VARIANT_ESP32C6
-#include "const_esp32c6.h"
-#endif  // USE_ESP32_VARIANT_ESP32C6
-
 #include "esphome/core/application.h"
 #include "esphome/core/log.h"
 
@@ -114,6 +110,7 @@ void ESP32BLE::advertising_init_() {
 
   this->advertising_->set_scan_response(true);
   this->advertising_->set_min_preferred_interval(0x06);
+  this->advertising_->set_appearance(this->appearance_);
 }
 
 bool ESP32BLE::ble_setup_() {
@@ -127,11 +124,7 @@ bool ESP32BLE::ble_setup_() {
   if (esp_bt_controller_get_status() != ESP_BT_CONTROLLER_STATUS_ENABLED) {
     // start bt controller
     if (esp_bt_controller_get_status() == ESP_BT_CONTROLLER_STATUS_IDLE) {
-#ifdef USE_ESP32_VARIANT_ESP32C6
-      esp_bt_controller_config_t cfg = BT_CONTROLLER_CONFIG;
-#else
       esp_bt_controller_config_t cfg = BT_CONTROLLER_INIT_CONFIG_DEFAULT();
-#endif
       err = esp_bt_controller_init(&cfg);
       if (err != ESP_OK) {
         ESP_LOGE(TAG, "esp_bt_controller_init failed: %s", esp_err_to_name(err));

esphome/components/esp32_ble/ble.h

@@ -95,6 +95,7 @@ class ESP32BLE : public Component {
   void advertising_start();
   void advertising_set_service_data(const std::vector<uint8_t> &data);
   void advertising_set_manufacturer_data(const std::vector<uint8_t> &data);
+  void advertising_set_appearance(uint16_t appearance) { this->appearance_ = appearance; }
   void advertising_add_service_uuid(ESPBTUUID uuid);
   void advertising_remove_service_uuid(ESPBTUUID uuid);
   void advertising_register_raw_advertisement_callback(std::function<void(bool)> &&callback);
@@ -128,11 +129,12 @@ class ESP32BLE : public Component {
   BLEComponentState state_{BLE_COMPONENT_STATE_OFF};
 
   Queue<BLEEvent> ble_events_;
-  BLEAdvertising *advertising_;
+  BLEAdvertising *advertising_{};
   esp_ble_io_cap_t io_cap_{ESP_IO_CAP_NONE};
-  uint32_t advertising_cycle_time_;
-  bool enable_on_boot_;
+  uint32_t advertising_cycle_time_{};
+  bool enable_on_boot_{};
   optional<std::string> name_;
+  uint16_t appearance_{0};
 };
 
 // NOLINTNEXTLINE(cppcoreguidelines-avoid-non-const-global-variables)

esphome/components/esp32_ble/ble_advertising.h

@@ -32,6 +32,7 @@ class BLEAdvertising {
   void set_scan_response(bool scan_response) { this->scan_response_ = scan_response; }
   void set_min_preferred_interval(uint16_t interval) { this->advertising_data_.min_interval = interval; }
   void set_manufacturer_data(const std::vector<uint8_t> &data);
+  void set_appearance(uint16_t appearance) { this->advertising_data_.appearance = appearance; }
   void set_service_data(const std::vector<uint8_t> &data);
   void register_raw_advertisement_callback(std::function<void(bool)> &&callback);
 

esphome/components/esp32_ble/const_esp32c6.h

@@ -1,74 +0,0 @@
-#pragma once
-
-#ifdef USE_ESP32_VARIANT_ESP32C6
-
-#include <esp_bt.h>
-
-namespace esphome {
-namespace esp32_ble {
-
-static const esp_bt_controller_config_t BT_CONTROLLER_CONFIG = {
-    .config_version = CONFIG_VERSION,
-    .ble_ll_resolv_list_size = CONFIG_BT_LE_LL_RESOLV_LIST_SIZE,
-    .ble_hci_evt_hi_buf_count = DEFAULT_BT_LE_HCI_EVT_HI_BUF_COUNT,
-    .ble_hci_evt_lo_buf_count = DEFAULT_BT_LE_HCI_EVT_LO_BUF_COUNT,
-    .ble_ll_sync_list_cnt = DEFAULT_BT_LE_MAX_PERIODIC_ADVERTISER_LIST,
-    .ble_ll_sync_cnt = DEFAULT_BT_LE_MAX_PERIODIC_SYNCS,
-    .ble_ll_rsp_dup_list_count = CONFIG_BT_LE_LL_DUP_SCAN_LIST_COUNT,
-    .ble_ll_adv_dup_list_count = CONFIG_BT_LE_LL_DUP_SCAN_LIST_COUNT,
-    .ble_ll_tx_pwr_dbm = BLE_LL_TX_PWR_DBM_N,
-    .rtc_freq = RTC_FREQ_N,
-    .ble_ll_sca = CONFIG_BT_LE_LL_SCA,
-    .ble_ll_scan_phy_number = BLE_LL_SCAN_PHY_NUMBER_N,
-    .ble_ll_conn_def_auth_pyld_tmo = BLE_LL_CONN_DEF_AUTH_PYLD_TMO_N,
-    .ble_ll_jitter_usecs = BLE_LL_JITTER_USECS_N,
-    .ble_ll_sched_max_adv_pdu_usecs = BLE_LL_SCHED_MAX_ADV_PDU_USECS_N,
-    .ble_ll_sched_direct_adv_max_usecs = BLE_LL_SCHED_DIRECT_ADV_MAX_USECS_N,
-    .ble_ll_sched_adv_max_usecs = BLE_LL_SCHED_ADV_MAX_USECS_N,
-    .ble_scan_rsp_data_max_len = DEFAULT_BT_LE_SCAN_RSP_DATA_MAX_LEN_N,
-    .ble_ll_cfg_num_hci_cmd_pkts = BLE_LL_CFG_NUM_HCI_CMD_PKTS_N,
-    .ble_ll_ctrl_proc_timeout_ms = BLE_LL_CTRL_PROC_TIMEOUT_MS_N,
-    .nimble_max_connections = DEFAULT_BT_LE_MAX_CONNECTIONS,
-    .ble_whitelist_size = DEFAULT_BT_NIMBLE_WHITELIST_SIZE,  // NOLINT
-    .ble_acl_buf_size = DEFAULT_BT_LE_ACL_BUF_SIZE,
-    .ble_acl_buf_count = DEFAULT_BT_LE_ACL_BUF_COUNT,
-    .ble_hci_evt_buf_size = DEFAULT_BT_LE_HCI_EVT_BUF_SIZE,
-    .ble_multi_adv_instances = DEFAULT_BT_LE_MAX_EXT_ADV_INSTANCES,
-    .ble_ext_adv_max_size = DEFAULT_BT_LE_EXT_ADV_MAX_SIZE,
-    .controller_task_stack_size = NIMBLE_LL_STACK_SIZE,
-    .controller_task_prio = ESP_TASK_BT_CONTROLLER_PRIO,
-    .controller_run_cpu = 0,
-    .enable_qa_test = RUN_QA_TEST,
-    .enable_bqb_test = RUN_BQB_TEST,
-#if ESP_IDF_VERSION < ESP_IDF_VERSION_VAL(5, 3, 1)
-    // The following fields have been removed since ESP IDF version 5.3.1, see commit:
-    // https://github.com/espressif/esp-idf/commit/e761c1de8f9c0777829d597b4d5a33bb070a30a8
-    .enable_uart_hci = HCI_UART_EN,
-    .ble_hci_uart_port = DEFAULT_BT_LE_HCI_UART_PORT,
-    .ble_hci_uart_baud = DEFAULT_BT_LE_HCI_UART_BAUD,
-    .ble_hci_uart_data_bits = DEFAULT_BT_LE_HCI_UART_DATA_BITS,
-    .ble_hci_uart_stop_bits = DEFAULT_BT_LE_HCI_UART_STOP_BITS,
-    .ble_hci_uart_flow_ctrl = DEFAULT_BT_LE_HCI_UART_FLOW_CTRL,
-    .ble_hci_uart_uart_parity = DEFAULT_BT_LE_HCI_UART_PARITY,
-#endif
-    .enable_tx_cca = DEFAULT_BT_LE_TX_CCA_ENABLED,
-    .cca_rssi_thresh = 256 - DEFAULT_BT_LE_CCA_RSSI_THRESH,
-    .sleep_en = NIMBLE_SLEEP_ENABLE,
-    .coex_phy_coded_tx_rx_time_limit = DEFAULT_BT_LE_COEX_PHY_CODED_TX_RX_TLIM_EFF,
-    .dis_scan_backoff = NIMBLE_DISABLE_SCAN_BACKOFF,
-    .ble_scan_classify_filter_enable = 1,
-    .main_xtal_freq = CONFIG_XTAL_FREQ,
-    .version_num = (uint8_t) efuse_hal_chip_revision(),
-    .cpu_freq_mhz = CONFIG_ESP_DEFAULT_CPU_FREQ_MHZ,
-    .ignore_wl_for_direct_adv = 0,
-    .enable_pcl = DEFAULT_BT_LE_POWER_CONTROL_ENABLED,
-#if ESP_IDF_VERSION >= ESP_IDF_VERSION_VAL(5, 1, 3)
-    .csa2_select = DEFAULT_BT_LE_50_FEATURE_SUPPORT,
-#endif
-    .config_magic = CONFIG_MAGIC,
-};
-
-}  // namespace esp32_ble
-}  // namespace esphome
-
-#endif  // USE_ESP32_VARIANT_ESP32C6

esphome/components/esp32_ble_server/__init__.py

@@ -32,6 +32,7 @@ DEPENDENCIES = ["esp32"]
 DOMAIN = "esp32_ble_server"
 
 CONF_ADVERTISE = "advertise"
+CONF_APPEARANCE = "appearance"
 CONF_BROADCAST = "broadcast"
 CONF_CHARACTERISTICS = "characteristics"
 CONF_DESCRIPTION = "description"
@@ -421,6 +422,7 @@ CONFIG_SCHEMA = cv.Schema(
         cv.GenerateID(): cv.declare_id(BLEServer),
         cv.GenerateID(esp32_ble.CONF_BLE_ID): cv.use_id(esp32_ble.ESP32BLE),
         cv.Optional(CONF_MANUFACTURER): value_schema("string", templatable=False),
+        cv.Optional(CONF_APPEARANCE, default=0): cv.uint16_t,
         cv.Optional(CONF_MODEL): value_schema("string", templatable=False),
         cv.Optional(CONF_FIRMWARE_VERSION): value_schema("string", templatable=False),
         cv.Optional(CONF_MANUFACTURER_DATA): cv.Schema([cv.uint8_t]),
@@ -531,6 +533,7 @@ async def to_code(config):
     cg.add(parent.register_gatts_event_handler(var))
     cg.add(parent.register_ble_status_event_handler(var))
     cg.add(var.set_parent(parent))
+    cg.add(parent.advertising_set_appearance(config[CONF_APPEARANCE]))
     if CONF_MANUFACTURER_DATA in config:
         cg.add(var.set_manufacturer_data(config[CONF_MANUFACTURER_DATA]))
     for service_config in config[CONF_SERVICES]:

esphome/components/esp32_ble_tracker/__init__.py

@@ -17,6 +17,7 @@ from esphome.components.esp32_ble import (
 import esphome.config_validation as cv
 from esphome.const import (
     CONF_ACTIVE,
+    CONF_CONTINUOUS,
     CONF_DURATION,
     CONF_ID,
     CONF_INTERVAL,
@@ -42,7 +43,6 @@ CONF_MAX_CONNECTIONS = "max_connections"
 CONF_ESP32_BLE_ID = "esp32_ble_id"
 CONF_SCAN_PARAMETERS = "scan_parameters"
 CONF_WINDOW = "window"
-CONF_CONTINUOUS = "continuous"
 CONF_ON_SCAN_END = "on_scan_end"
 
 DEFAULT_MAX_CONNECTIONS = 3

esphome/components/esp32_ble_tracker/esp32_ble_tracker.cpp

@@ -245,7 +245,7 @@ void ESP32BLETracker::stop_scan_() {
     return;
   }
   this->cancel_timeout("scan");
-  this->scanner_state_ = ScannerState::STOPPING;
+  this->set_scanner_state_(ScannerState::STOPPING);
   esp_err_t err = esp_ble_gap_stop_scanning();
   if (err != ESP_OK) {
     ESP_LOGE(TAG, "esp_ble_gap_stop_scanning failed: %d", err);
@@ -272,7 +272,7 @@ void ESP32BLETracker::start_scan_(bool first) {
     }
     return;
   }
-  this->scanner_state_ = ScannerState::STARTING;
+  this->set_scanner_state_(ScannerState::STARTING);
   ESP_LOGD(TAG, "Starting scan, set scanner state to STARTING.");
   if (!first) {
     for (auto *listener : this->listeners_)
@@ -315,7 +315,7 @@ void ESP32BLETracker::end_of_scan_() {
 
   for (auto *listener : this->listeners_)
     listener->on_scan_end();
-  this->scanner_state_ = ScannerState::IDLE;
+  this->set_scanner_state_(ScannerState::IDLE);
 }
 
 void ESP32BLETracker::register_client(ESPBTClient *client) {
@@ -398,9 +398,9 @@ void ESP32BLETracker::gap_scan_start_complete_(const esp_ble_gap_cb_param_t::ble
   }
   if (param.status == ESP_BT_STATUS_SUCCESS) {
     this->scan_start_fail_count_ = 0;
-    this->scanner_state_ = ScannerState::RUNNING;
+    this->set_scanner_state_(ScannerState::RUNNING);
   } else {
-    this->scanner_state_ = ScannerState::FAILED;
+    this->set_scanner_state_(ScannerState::FAILED);
     if (this->scan_start_fail_count_ != std::numeric_limits<uint8_t>::max()) {
       this->scan_start_fail_count_++;
     }
@@ -422,7 +422,7 @@ void ESP32BLETracker::gap_scan_stop_complete_(const esp_ble_gap_cb_param_t::ble_
       ESP_LOGE(TAG, "Scan was stopped when stop complete.");
     }
   }
-  this->scanner_state_ = ScannerState::STOPPED;
+  this->set_scanner_state_(ScannerState::STOPPED);
 }
 
 void ESP32BLETracker::gap_scan_result_(const esp_ble_gap_cb_param_t::ble_scan_result_evt_param &param) {
@@ -449,7 +449,7 @@ void ESP32BLETracker::gap_scan_result_(const esp_ble_gap_cb_param_t::ble_scan_re
         ESP_LOGE(TAG, "Scan was stopped when scan completed.");
       }
     }
-    this->scanner_state_ = ScannerState::STOPPED;
+    this->set_scanner_state_(ScannerState::STOPPED);
   }
 }
 
@@ -460,6 +460,11 @@ void ESP32BLETracker::gattc_event_handler(esp_gattc_cb_event_t event, esp_gatt_i
   }
 }
 
+void ESP32BLETracker::set_scanner_state_(ScannerState state) {
+  this->scanner_state_ = state;
+  this->scanner_state_callbacks_.call(state);
+}
+
 ESPBLEiBeacon::ESPBLEiBeacon(const uint8_t *data) { memcpy(&this->beacon_data_, data, sizeof(beacon_data_)); }
 optional<ESPBLEiBeacon> ESPBLEiBeacon::from_manufacturer_data(const ServiceData &data) {
   if (!data.uuid.contains(0x4C, 0x00))

esphome/components/esp32_ble_tracker/esp32_ble_tracker.h

@@ -218,6 +218,7 @@ class ESP32BLETracker : public Component,
   void set_scan_interval(uint32_t scan_interval) { scan_interval_ = scan_interval; }
   void set_scan_window(uint32_t scan_window) { scan_window_ = scan_window; }
   void set_scan_active(bool scan_active) { scan_active_ = scan_active; }
+  bool get_scan_active() const { return scan_active_; }
   void set_scan_continuous(bool scan_continuous) { scan_continuous_ = scan_continuous; }
 
   /// Setup the FreeRTOS task and the Bluetooth stack.
@@ -241,6 +242,11 @@ class ESP32BLETracker : public Component,
   void gap_event_handler(esp_gap_ble_cb_event_t event, esp_ble_gap_cb_param_t *param) override;
   void ble_before_disabled_event_handler() override;
 
+  void add_scanner_state_callback(std::function<void(ScannerState)> &&callback) {
+    this->scanner_state_callbacks_.add(std::move(callback));
+  }
+  ScannerState get_scanner_state() const { return this->scanner_state_; }
+
  protected:
   void stop_scan_();
   /// Start a single scan by setting up the parameters and doing some esp-idf calls.
@@ -255,6 +261,8 @@ class ESP32BLETracker : public Component,
   void gap_scan_start_complete_(const esp_ble_gap_cb_param_t::ble_scan_start_cmpl_evt_param &param);
   /// Called when a `ESP_GAP_BLE_SCAN_STOP_COMPLETE_EVT` event is received.
   void gap_scan_stop_complete_(const esp_ble_gap_cb_param_t::ble_scan_stop_cmpl_evt_param &param);
+  /// Called to set the scanner state. Will also call callbacks to let listeners know when state is changed.
+  void set_scanner_state_(ScannerState state);
 
   int app_id_{0};
 
@@ -273,6 +281,7 @@ class ESP32BLETracker : public Component,
   bool scan_continuous_;
   bool scan_active_;
   ScannerState scanner_state_{ScannerState::IDLE};
+  CallbackManager<void(ScannerState)> scanner_state_callbacks_;
   bool ble_was_disabled_{true};
   bool raw_advertisements_{false};
   bool parse_advertisements_{false};

esphome/components/fastled_base/__init__.py

@@ -40,9 +40,6 @@ async def new_fastled_light(config):
     if CONF_MAX_REFRESH_RATE in config:
         cg.add(var.set_max_refresh_rate(config[CONF_MAX_REFRESH_RATE]))
 
+    cg.add_library("fastled/FastLED", "3.9.16")
     await light.register_light(var, config)
-    # https://github.com/FastLED/FastLED/blob/master/library.json
-    # 3.3.3 has an issue on ESP32 with RMT and fastled_clockless:
-    # https://github.com/esphome/issues/issues/1375
-    cg.add_library("fastled/FastLED", "3.3.2")
     return var

esphome/components/fastled_base/fastled_light.cpp

@@ -34,7 +34,7 @@ void FastLEDLightOutput::write_state(light::LightState *state) {
   this->mark_shown_();
 
   ESP_LOGVV(TAG, "Writing RGB values to bus...");
-  this->controller_->showLeds();
+  this->controller_->showLeds(this->state_parent_->current_values.get_brightness() * 255);
 }
 
 }  // namespace fastled_base

esphome/components/gpio_expander/cached_gpio.h

@@ -8,30 +8,45 @@ namespace esphome {
 namespace gpio_expander {
 
 /// @brief A class to cache the read state of a GPIO expander.
+///        This class caches reads between GPIO Pins which are on the same bank.
+///        This means that for reading whole Port (ex. 8 pins) component needs only one
+///        I2C/SPI read per main loop call. It assumes, that one bit in byte identifies one GPIO pin
+///        Template parameters:
+///           T - Type which represents internal register. Could be uint8_t or uint16_t. Adjust to
+///               match size of your internal GPIO bank register.
+///           N - Number of pins
 template<typename T, T N> class CachedGpioExpander {
  public:
   bool digital_read(T pin) {
-    if (!this->read_cache_invalidated_[pin]) {
-      this->read_cache_invalidated_[pin] = true;
-      return this->digital_read_cache(pin);
+    uint8_t bank = pin / (sizeof(T) * BITS_PER_BYTE);
+    if (this->read_cache_invalidated_[bank]) {
+      this->read_cache_invalidated_[bank] = false;
+      if (!this->digital_read_hw(pin))
+        return false;
     }
-    return this->digital_read_hw(pin);
+    return this->digital_read_cache(pin);
   }
 
   void digital_write(T pin, bool value) { this->digital_write_hw(pin, value); }
 
  protected:
+  /// @brief Call component low level function to read GPIO state from device
   virtual bool digital_read_hw(T pin) = 0;
+  /// @brief Call component read function from internal cache.
   virtual bool digital_read_cache(T pin) = 0;
+  /// @brief Call component low level function to write GPIO state to device
   virtual void digital_write_hw(T pin, bool value) = 0;
+  const uint8_t cache_byte_size_ = N / (sizeof(T) * BITS_PER_BYTE);
 
+  /// @brief Invalidate cache. This function should be called in component loop().
   void reset_pin_cache_() {
-    for (T i = 0; i < N; i++) {
-      this->read_cache_invalidated_[i] = false;
+    for (T i = 0; i < this->cache_byte_size_; i++) {
+      this->read_cache_invalidated_[i] = true;
     }
   }
 
-  std::array<bool, N> read_cache_invalidated_{};
+  static const uint8_t BITS_PER_BYTE = 8;
+  std::array<bool, N / (sizeof(T) * BITS_PER_BYTE)> read_cache_invalidated_{};
 };
 
 }  // namespace gpio_expander

esphome/components/graph/__init__.py

@@ -5,6 +5,7 @@ import esphome.config_validation as cv
 from esphome.const import (
     CONF_BORDER,
     CONF_COLOR,
+    CONF_CONTINUOUS,
     CONF_DIRECTION,
     CONF_DURATION,
     CONF_HEIGHT,
@@ -61,8 +62,6 @@ VALUE_POSITION_TYPE = {
     "BELOW": ValuePositionType.VALUE_POSITION_TYPE_BELOW,
 }
 
-CONF_CONTINUOUS = "continuous"
-
 GRAPH_TRACE_SCHEMA = cv.Schema(
     {
         cv.GenerateID(): cv.declare_id(GraphTrace),

esphome/components/gree/climate.py

@@ -18,6 +18,7 @@ MODELS = {
     "yac": Model.GREE_YAC,
     "yac1fb9": Model.GREE_YAC1FB9,
     "yx1ff": Model.GREE_YX1FF,
+    "yag": Model.GREE_YAG,
 }
 
 CONFIG_SCHEMA = climate_ir.CLIMATE_IR_WITH_RECEIVER_SCHEMA.extend(

esphome/components/gree/gree.cpp

@@ -22,13 +22,21 @@ void GreeClimate::transmit_state() {
   remote_state[0] = this->fan_speed_() | this->operation_mode_();
   remote_state[1] = this->temperature_();
 
-  if (this->model_ == GREE_YAN || this->model_ == GREE_YX1FF) {
+  if (this->model_ == GREE_YAN || this->model_ == GREE_YX1FF || this->model_ == GREE_YAG) {
     remote_state[2] = 0x60;
     remote_state[3] = 0x50;
     remote_state[4] = this->vertical_swing_();
   }
 
-  if (this->model_ == GREE_YAC) {
+  if (this->model_ == GREE_YAG) {
+    remote_state[5] = 0x40;
+
+    if (this->vertical_swing_() == GREE_VDIR_SWING || this->horizontal_swing_() == GREE_HDIR_SWING) {
+      remote_state[0] |= (1 << 6);
+    }
+  }
+
+  if (this->model_ == GREE_YAC || this->model_ == GREE_YAG) {
     remote_state[4] |= (this->horizontal_swing_() << 4);
   }
 
@@ -57,6 +65,12 @@ void GreeClimate::transmit_state() {
   // Calculate the checksum
   if (this->model_ == GREE_YAN || this->model_ == GREE_YX1FF) {
     remote_state[7] = ((remote_state[0] << 4) + (remote_state[1] << 4) + 0xC0);
+  } else if (this->model_ == GREE_YAG) {
+    remote_state[7] =
+        ((((remote_state[0] & 0x0F) + (remote_state[1] & 0x0F) + (remote_state[2] & 0x0F) + (remote_state[3] & 0x0F) +
+           ((remote_state[4] & 0xF0) >> 4) + ((remote_state[5] & 0xF0) >> 4) + ((remote_state[6] & 0xF0) >> 4) + 0x0A) &
+          0x0F)
+         << 4);
   } else {
     remote_state[7] =
         ((((remote_state[0] & 0x0F) + (remote_state[1] & 0x0F) + (remote_state[2] & 0x0F) + (remote_state[3] & 0x0F) +

esphome/components/gree/gree.h

@@ -58,7 +58,7 @@ const uint8_t GREE_VDIR_MIDDLE = 0x04;
 const uint8_t GREE_VDIR_MDOWN = 0x05;
 const uint8_t GREE_VDIR_DOWN = 0x06;
 
-// Only available on YAC
+// Only available on YAC/YAG
 // Horizontal air directions. Note that these cannot be set on all heat pumps
 const uint8_t GREE_HDIR_AUTO = 0x00;
 const uint8_t GREE_HDIR_MANUAL = 0x00;
@@ -78,7 +78,7 @@ const uint8_t GREE_PRESET_SLEEP = 0x01;
 const uint8_t GREE_PRESET_SLEEP_BIT = 0x80;
 
 // Model codes
-enum Model { GREE_GENERIC, GREE_YAN, GREE_YAA, GREE_YAC, GREE_YAC1FB9, GREE_YX1FF };
+enum Model { GREE_GENERIC, GREE_YAN, GREE_YAA, GREE_YAC, GREE_YAC1FB9, GREE_YX1FF, GREE_YAG };
 
 class GreeClimate : public climate_ir::ClimateIR {
  public:

esphome/components/hlw8012/hlw8012.cpp

@@ -69,7 +69,7 @@ void HLW8012Component::update() {
 
   float power = cf_hz * this->power_multiplier_;
 
-  if (this->change_mode_at_ != 0) {
+  if (this->change_mode_at_ != 0 || this->change_mode_every_ == 0) {
     // Only read cf1 after one cycle. Apparently it's quite unstable after being changed.
     if (this->current_mode_) {
       float current = cf1_hz * this->current_multiplier_;

esphome/components/hm3301/hm3301.h

@@ -8,7 +8,7 @@
 namespace esphome {
 namespace hm3301 {
 
-static const uint8_t SELECT_COMM_CMD = 0X88;
+static const uint8_t SELECT_COMM_CMD = 0x88;
 
 class HM3301Component : public PollingComponent, public i2c::I2CDevice {
  public:

esphome/components/http_request/__init__.py

@@ -295,8 +295,8 @@ async def http_request_action_to_code(config, action_id, template_arg, args):
             for key in json_:
                 template_ = await cg.templatable(json_[key], args, cg.std_string)
                 cg.add(var.add_json(key, template_))
-    for key in config.get(CONF_REQUEST_HEADERS, []):
-        template_ = await cg.templatable(key, args, cg.std_string)
+    for key, value in config.get(CONF_REQUEST_HEADERS, {}).items():
+        template_ = await cg.templatable(value, args, cg.const_char_ptr)
         cg.add(var.add_request_header(key, template_))
 
     for value in config.get(CONF_COLLECT_HEADERS, []):

esphome/components/i2c/i2c.h

@@ -139,6 +139,10 @@ class I2CDevice {
   /// @param address of the device
   void set_i2c_address(uint8_t address) { address_ = address; }
 
+  /// @brief Returns the I2C address of the object.
+  /// @return the I2C address
+  uint8_t get_i2c_address() const { return this->address_; }
+
   /// @brief we store the pointer to the I2CBus to use
   /// @param bus pointer to the I2CBus object
   void set_i2c_bus(I2CBus *bus) { bus_ = bus; }

esphome/components/i2c/i2c_bus_esp_idf.cpp

@@ -67,7 +67,7 @@ void IDFI2CBus::setup() {
       ESP_LOGV(TAG, "i2c_timeout set to %" PRIu32 " ticks (%" PRIu32 " us)", timeout_ * 80, timeout_);
     }
   }
-  err = i2c_driver_install(port_, I2C_MODE_MASTER, 0, 0, ESP_INTR_FLAG_IRAM);
+  err = i2c_driver_install(port_, I2C_MODE_MASTER, 0, 0, 0);
   if (err != ESP_OK) {
     ESP_LOGW(TAG, "i2c_driver_install failed: %s", esp_err_to_name(err));
     this->mark_failed();

esphome/components/i2s_audio/__init__.py

@@ -39,6 +39,7 @@ CONF_SECONDARY = "secondary"
 
 CONF_USE_APLL = "use_apll"
 CONF_BITS_PER_CHANNEL = "bits_per_channel"
+CONF_MCLK_MULTIPLE = "mclk_multiple"
 CONF_MONO = "mono"
 CONF_LEFT = "left"
 CONF_RIGHT = "right"
@@ -122,8 +123,25 @@ I2S_SLOT_BIT_WIDTH = {
     32: i2s_slot_bit_width_t.I2S_SLOT_BIT_WIDTH_32BIT,
 }
 
+i2s_mclk_multiple_t = cg.global_ns.enum("i2s_mclk_multiple_t")
+I2S_MCLK_MULTIPLE = {
+    128: i2s_mclk_multiple_t.I2S_MCLK_MULTIPLE_128,
+    256: i2s_mclk_multiple_t.I2S_MCLK_MULTIPLE_256,
+    384: i2s_mclk_multiple_t.I2S_MCLK_MULTIPLE_384,
+    512: i2s_mclk_multiple_t.I2S_MCLK_MULTIPLE_512,
+}
+
 _validate_bits = cv.float_with_unit("bits", "bit")
 
+
+def validate_mclk_divisible_by_3(config):
+    if config[CONF_BITS_PER_SAMPLE] == 24 and config[CONF_MCLK_MULTIPLE] % 3 != 0:
+        raise cv.Invalid(
+            f"{CONF_MCLK_MULTIPLE} must be divisible by 3 when bits per sample is 24"
+        )
+    return config
+
+
 _use_legacy_driver = None
 
 
@@ -155,6 +173,7 @@ def i2s_audio_component_schema(
                 cv.Any(cv.float_with_unit("bits", "bit"), "default"),
                 cv.one_of(*I2S_BITS_PER_CHANNEL),
             ),
+            cv.Optional(CONF_MCLK_MULTIPLE, default=256): cv.one_of(*I2S_MCLK_MULTIPLE),
         }
     )
 
@@ -182,11 +201,10 @@ async def register_i2s_audio_component(var, config):
             slot_mask = CONF_BOTH
         cg.add(var.set_slot_mode(I2S_SLOT_MODE[slot_mode]))
         cg.add(var.set_std_slot_mask(I2S_STD_SLOT_MASK[slot_mask]))
-        cg.add(
-            var.set_slot_bit_width(I2S_SLOT_BIT_WIDTH[config[CONF_BITS_PER_CHANNEL]])
-        )
+        cg.add(var.set_slot_bit_width(I2S_SLOT_BIT_WIDTH[config[CONF_BITS_PER_SAMPLE]]))
     cg.add(var.set_sample_rate(config[CONF_SAMPLE_RATE]))
     cg.add(var.set_use_apll(config[CONF_USE_APLL]))
+    cg.add(var.set_mclk_multiple(I2S_MCLK_MULTIPLE[config[CONF_MCLK_MULTIPLE]]))
 
 
 def validate_use_legacy(value):

esphome/components/i2s_audio/i2s_audio.h

@@ -31,6 +31,7 @@ class I2SAudioBase : public Parented<I2SAudioComponent> {
 #endif
   void set_sample_rate(uint32_t sample_rate) { this->sample_rate_ = sample_rate; }
   void set_use_apll(uint32_t use_apll) { this->use_apll_ = use_apll; }
+  void set_mclk_multiple(i2s_mclk_multiple_t mclk_multiple) { this->mclk_multiple_ = mclk_multiple; }
 
  protected:
 #ifdef USE_I2S_LEGACY
@@ -46,6 +47,7 @@ class I2SAudioBase : public Parented<I2SAudioComponent> {
 #endif
   uint32_t sample_rate_;
   bool use_apll_;
+  i2s_mclk_multiple_t mclk_multiple_;
 };
 
 class I2SAudioIn : public I2SAudioBase {};

esphome/components/i2s_audio/microphone/__init__.py

@@ -1,13 +1,20 @@
 from esphome import pins
 import esphome.codegen as cg
-from esphome.components import esp32, microphone
+from esphome.components import audio, esp32, microphone
 from esphome.components.adc import ESP32_VARIANT_ADC1_PIN_TO_CHANNEL, validate_adc_pin
 import esphome.config_validation as cv
-from esphome.const import CONF_ID, CONF_NUMBER
+from esphome.const import (
+    CONF_BITS_PER_SAMPLE,
+    CONF_CHANNEL,
+    CONF_ID,
+    CONF_NUM_CHANNELS,
+    CONF_NUMBER,
+    CONF_SAMPLE_RATE,
+)
 
 from .. import (
-    CONF_CHANNEL,
     CONF_I2S_DIN_PIN,
+    CONF_LEFT,
     CONF_MONO,
     CONF_RIGHT,
     I2SAudioIn,
@@ -15,6 +22,7 @@ from .. import (
     i2s_audio_ns,
     register_i2s_audio_component,
     use_legacy,
+    validate_mclk_divisible_by_3,
 )
 
 CODEOWNERS = ["@jesserockz"]
@@ -32,7 +40,7 @@ INTERNAL_ADC_VARIANTS = [esp32.const.VARIANT_ESP32]
 PDM_VARIANTS = [esp32.const.VARIANT_ESP32, esp32.const.VARIANT_ESP32S3]
 
 
-def validate_esp32_variant(config):
+def _validate_esp32_variant(config):
     variant = esp32.get_esp32_variant()
     if config[CONF_ADC_TYPE] == "external":
         if config[CONF_PDM]:
@@ -46,12 +54,34 @@ def validate_esp32_variant(config):
     raise NotImplementedError
 
 
-def validate_channel(config):
+def _validate_channel(config):
     if config[CONF_CHANNEL] == CONF_MONO:
         raise cv.Invalid(f"I2S microphone does not support {CONF_MONO}.")
     return config
 
 
+def _set_num_channels_from_config(config):
+    if config[CONF_CHANNEL] in (CONF_LEFT, CONF_RIGHT):
+        config[CONF_NUM_CHANNELS] = 1
+    else:
+        config[CONF_NUM_CHANNELS] = 2
+
+    return config
+
+
+def _set_stream_limits(config):
+    audio.set_stream_limits(
+        min_bits_per_sample=config.get(CONF_BITS_PER_SAMPLE),
+        max_bits_per_sample=config.get(CONF_BITS_PER_SAMPLE),
+        min_channels=config.get(CONF_NUM_CHANNELS),
+        max_channels=config.get(CONF_NUM_CHANNELS),
+        min_sample_rate=config.get(CONF_SAMPLE_RATE),
+        max_sample_rate=config.get(CONF_SAMPLE_RATE),
+    )(config)
+
+    return config
+
+
 BASE_SCHEMA = microphone.MICROPHONE_SCHEMA.extend(
     i2s_audio_component_schema(
         I2SAudioMicrophone,
@@ -79,8 +109,11 @@ CONFIG_SCHEMA = cv.All(
         },
         key=CONF_ADC_TYPE,
     ),
-    validate_esp32_variant,
-    validate_channel,
+    _validate_esp32_variant,
+    _validate_channel,
+    _set_num_channels_from_config,
+    _set_stream_limits,
+    validate_mclk_divisible_by_3,
 )
 
 

esphome/components/i2s_audio/microphone/i2s_audio_microphone.cpp

@@ -15,10 +15,25 @@
 namespace esphome {
 namespace i2s_audio {
 
-static const size_t BUFFER_SIZE = 512;
+static const UBaseType_t MAX_LISTENERS = 16;
+
+static const uint32_t READ_DURATION_MS = 16;
+
+static const size_t TASK_STACK_SIZE = 4096;
+static const ssize_t TASK_PRIORITY = 23;
 
 static const char *const TAG = "i2s_audio.microphone";
 
+enum MicrophoneEventGroupBits : uint32_t {
+  COMMAND_STOP = (1 << 0),  // stops the microphone task
+  TASK_STARTING = (1 << 10),
+  TASK_RUNNING = (1 << 11),
+  TASK_STOPPING = (1 << 12),
+  TASK_STOPPED = (1 << 13),
+
+  ALL_BITS = 0x00FFFFFF,  // All valid FreeRTOS event group bits
+};
+
 void I2SAudioMicrophone::setup() {
   ESP_LOGCONFIG(TAG, "Setting up I2S Audio Microphone...");
 #ifdef USE_I2S_LEGACY
@@ -41,21 +56,64 @@ void I2SAudioMicrophone::setup() {
       }
     }
   }
+
+  this->active_listeners_semaphore_ = xSemaphoreCreateCounting(MAX_LISTENERS, MAX_LISTENERS);
+  if (this->active_listeners_semaphore_ == nullptr) {
+    ESP_LOGE(TAG, "Failed to create semaphore");
+    this->mark_failed();
+    return;
+  }
+
+  this->event_group_ = xEventGroupCreate();
+  if (this->event_group_ == nullptr) {
+    ESP_LOGE(TAG, "Failed to create event group");
+    this->mark_failed();
+    return;
+  }
 }
 
 void I2SAudioMicrophone::start() {
   if (this->is_failed())
     return;
-  if (this->state_ == microphone::STATE_RUNNING)
-    return;  // Already running
-  this->state_ = microphone::STATE_STARTING;
+
+  xSemaphoreTake(this->active_listeners_semaphore_, 0);
 }
-void I2SAudioMicrophone::start_() {
+
+bool I2SAudioMicrophone::start_driver_() {
   if (!this->parent_->try_lock()) {
-    return;  // Waiting for another i2s to return lock
+    return false;  // Waiting for another i2s to return lock
   }
   esp_err_t err;
 
+  uint8_t channel_count = 1;
+#ifdef USE_I2S_LEGACY
+  uint8_t bits_per_sample = this->bits_per_sample_;
+
+  if (this->channel_ == I2S_CHANNEL_FMT_RIGHT_LEFT) {
+    channel_count = 2;
+  }
+#else
+  if (this->slot_bit_width_ == I2S_SLOT_BIT_WIDTH_AUTO) {
+    this->slot_bit_width_ = I2S_SLOT_BIT_WIDTH_16BIT;
+  }
+  uint8_t bits_per_sample = this->slot_bit_width_;
+
+  if (this->slot_mode_ == I2S_SLOT_MODE_STEREO) {
+    channel_count = 2;
+  }
+#endif
+
+#ifdef USE_ESP32_VARIANT_ESP32
+  // ESP32 reads audio aligned to a multiple of 2 bytes. For example, if configured for 24 bits per sample, then it will
+  // produce 32 bits per sample, where the actual data is in the most significant bits. Other ESP32 variants produce 24
+  // bits per sample in this situation.
+  if (bits_per_sample < 16) {
+    bits_per_sample = 16;
+  } else if ((bits_per_sample > 16) && (bits_per_sample <= 32)) {
+    bits_per_sample = 32;
+  }
+#endif
+
 #ifdef USE_I2S_LEGACY
   i2s_driver_config_t config = {
       .mode = (i2s_mode_t) (this->i2s_mode_ | I2S_MODE_RX),
@@ -65,11 +123,11 @@ void I2SAudioMicrophone::start_() {
       .communication_format = I2S_COMM_FORMAT_STAND_I2S,
       .intr_alloc_flags = ESP_INTR_FLAG_LEVEL1,
       .dma_buf_count = 4,
-      .dma_buf_len = 256,
+      .dma_buf_len = 240,  // Must be divisible by 3 to support 24 bits per sample on old driver and newer variants
       .use_apll = this->use_apll_,
       .tx_desc_auto_clear = false,
       .fixed_mclk = 0,
-      .mclk_multiple = I2S_MCLK_MULTIPLE_256,
+      .mclk_multiple = this->mclk_multiple_,
       .bits_per_chan = this->bits_per_channel_,
   };
 
@@ -80,20 +138,20 @@ void I2SAudioMicrophone::start_() {
     if (err != ESP_OK) {
       ESP_LOGW(TAG, "Error installing I2S driver: %s", esp_err_to_name(err));
       this->status_set_error();
-      return;
+      return false;
     }
 
     err = i2s_set_adc_mode(ADC_UNIT_1, this->adc_channel_);
     if (err != ESP_OK) {
       ESP_LOGW(TAG, "Error setting ADC mode: %s", esp_err_to_name(err));
       this->status_set_error();
-      return;
+      return false;
     }
     err = i2s_adc_enable(this->parent_->get_port());
     if (err != ESP_OK) {
       ESP_LOGW(TAG, "Error enabling ADC: %s", esp_err_to_name(err));
       this->status_set_error();
-      return;
+      return false;
     }
 
   } else
@@ -106,7 +164,7 @@ void I2SAudioMicrophone::start_() {
     if (err != ESP_OK) {
       ESP_LOGW(TAG, "Error installing I2S driver: %s", esp_err_to_name(err));
       this->status_set_error();
-      return;
+      return false;
     }
 
     i2s_pin_config_t pin_config = this->parent_->get_pin_config();
@@ -116,7 +174,7 @@ void I2SAudioMicrophone::start_() {
     if (err != ESP_OK) {
       ESP_LOGW(TAG, "Error setting I2S pin: %s", esp_err_to_name(err));
       this->status_set_error();
-      return;
+      return false;
     }
   }
 #else
@@ -132,7 +190,7 @@ void I2SAudioMicrophone::start_() {
   if (err != ESP_OK) {
     ESP_LOGW(TAG, "Error creating new I2S channel: %s", esp_err_to_name(err));
     this->status_set_error();
-    return;
+    return false;
   }
 
   i2s_clock_src_t clk_src = I2S_CLK_SRC_DEFAULT;
@@ -144,10 +202,12 @@ void I2SAudioMicrophone::start_() {
   i2s_std_gpio_config_t pin_config = this->parent_->get_pin_config();
 #if SOC_I2S_SUPPORTS_PDM_RX
   if (this->pdm_) {
+    bits_per_sample = 16;  // PDM mics are always 16 bits per sample with the IDF 5 driver
+
     i2s_pdm_rx_clk_config_t clk_cfg = {
         .sample_rate_hz = this->sample_rate_,
         .clk_src = clk_src,
-        .mclk_multiple = I2S_MCLK_MULTIPLE_256,
+        .mclk_multiple = this->mclk_multiple_,
         .dn_sample_mode = I2S_PDM_DSR_8S,
     };
 
@@ -185,15 +245,10 @@ void I2SAudioMicrophone::start_() {
     i2s_std_clk_config_t clk_cfg = {
         .sample_rate_hz = this->sample_rate_,
         .clk_src = clk_src,
-        .mclk_multiple = I2S_MCLK_MULTIPLE_256,
+        .mclk_multiple = this->mclk_multiple_,
     };
-    i2s_data_bit_width_t data_bit_width;
-    if (this->slot_bit_width_ != I2S_SLOT_BIT_WIDTH_8BIT) {
-      data_bit_width = I2S_DATA_BIT_WIDTH_16BIT;
-    } else {
-      data_bit_width = I2S_DATA_BIT_WIDTH_8BIT;
-    }
-    i2s_std_slot_config_t std_slot_cfg = I2S_STD_PHILIPS_SLOT_DEFAULT_CONFIG(data_bit_width, this->slot_mode_);
+    i2s_std_slot_config_t std_slot_cfg =
+        I2S_STD_PHILIPS_SLOT_DEFAULT_CONFIG((i2s_data_bit_width_t) this->slot_bit_width_, this->slot_mode_);
     std_slot_cfg.slot_bit_width = this->slot_bit_width_;
     std_slot_cfg.slot_mask = this->std_slot_mask_;
 
@@ -210,7 +265,7 @@ void I2SAudioMicrophone::start_() {
   if (err != ESP_OK) {
     ESP_LOGW(TAG, "Error initializing I2S channel: %s", esp_err_to_name(err));
     this->status_set_error();
-    return;
+    return false;
   }
 
   /* Before reading data, start the RX channel first */
@@ -218,26 +273,25 @@ void I2SAudioMicrophone::start_() {
   if (err != ESP_OK) {
     ESP_LOGW(TAG, "Error enabling I2S Microphone: %s", esp_err_to_name(err));
     this->status_set_error();
-    return;
+    return false;
   }
 #endif
 
-  this->state_ = microphone::STATE_RUNNING;
-  this->high_freq_.start();
+  this->audio_stream_info_ = audio::AudioStreamInfo(bits_per_sample, channel_count, this->sample_rate_);
+
   this->status_clear_error();
+
+  return true;
 }
 
 void I2SAudioMicrophone::stop() {
   if (this->state_ == microphone::STATE_STOPPED || this->is_failed())
     return;
-  if (this->state_ == microphone::STATE_STARTING) {
-    this->state_ = microphone::STATE_STOPPED;
-    return;
-  }
-  this->state_ = microphone::STATE_STOPPING;
+
+  xSemaphoreGive(this->active_listeners_semaphore_);
 }
 
-void I2SAudioMicrophone::stop_() {
+void I2SAudioMicrophone::stop_driver_() {
   esp_err_t err;
 #ifdef USE_I2S_LEGACY
 #if SOC_I2S_SUPPORTS_ADC
@@ -279,12 +333,52 @@ void I2SAudioMicrophone::stop_() {
   }
 #endif
   this->parent_->unlock();
-  this->state_ = microphone::STATE_STOPPED;
-  this->high_freq_.stop();
   this->status_clear_error();
 }
 
-size_t I2SAudioMicrophone::read(int16_t *buf, size_t len, TickType_t ticks_to_wait) {
+void I2SAudioMicrophone::mic_task(void *params) {
+  I2SAudioMicrophone *this_microphone = (I2SAudioMicrophone *) params;
+
+  xEventGroupSetBits(this_microphone->event_group_, MicrophoneEventGroupBits::TASK_STARTING);
+
+  uint8_t start_counter = 0;
+  bool started = this_microphone->start_driver_();
+  while (!started && start_counter < 10) {
+    // Attempt to load the driver again in 100 ms. Doesn't slow down main loop since its in a task.
+    vTaskDelay(pdMS_TO_TICKS(100));
+    ++start_counter;
+    started = this_microphone->start_driver_();
+  }
+
+  if (started) {
+    xEventGroupSetBits(this_microphone->event_group_, MicrophoneEventGroupBits::TASK_RUNNING);
+    const size_t bytes_to_read = this_microphone->audio_stream_info_.ms_to_bytes(READ_DURATION_MS);
+    std::vector<uint8_t> samples;
+    samples.reserve(bytes_to_read);
+
+    while (!(xEventGroupGetBits(this_microphone->event_group_) & COMMAND_STOP)) {
+      if (this_microphone->data_callbacks_.size() > 0) {
+        samples.resize(bytes_to_read);
+        size_t bytes_read = this_microphone->read_(samples.data(), bytes_to_read, 2 * pdMS_TO_TICKS(READ_DURATION_MS));
+        samples.resize(bytes_read);
+        this_microphone->data_callbacks_.call(samples);
+      } else {
+        delay(READ_DURATION_MS);
+      }
+    }
+  }
+
+  xEventGroupSetBits(this_microphone->event_group_, MicrophoneEventGroupBits::TASK_STOPPING);
+  this_microphone->stop_driver_();
+
+  xEventGroupSetBits(this_microphone->event_group_, MicrophoneEventGroupBits::TASK_STOPPED);
+  while (true) {
+    // Continuously delay until the loop method delete the task
+    delay(10);
+  }
+}
+
+size_t I2SAudioMicrophone::read_(uint8_t *buf, size_t len, TickType_t ticks_to_wait) {
   size_t bytes_read = 0;
 #ifdef USE_I2S_LEGACY
   esp_err_t err = i2s_read(this->parent_->get_port(), buf, len, &bytes_read, ticks_to_wait);
@@ -303,38 +397,7 @@ size_t I2SAudioMicrophone::read(int16_t *buf, size_t len, TickType_t ticks_to_wa
     return 0;
   }
   this->status_clear_warning();
-  // ESP-IDF I2S implementation right-extends 8-bit data to 16 bits,
-  // and 24-bit data to 32 bits.
-#ifdef USE_I2S_LEGACY
-  switch (this->bits_per_sample_) {
-    case I2S_BITS_PER_SAMPLE_8BIT:
-    case I2S_BITS_PER_SAMPLE_16BIT:
-      return bytes_read;
-    case I2S_BITS_PER_SAMPLE_24BIT:
-    case I2S_BITS_PER_SAMPLE_32BIT: {
-      size_t samples_read = bytes_read / sizeof(int32_t);
-      for (size_t i = 0; i < samples_read; i++) {
-        int32_t temp = reinterpret_cast<int32_t *>(buf)[i] >> 14;
-        buf[i] = clamp<int16_t>(temp, INT16_MIN, INT16_MAX);
-      }
-      return samples_read * sizeof(int16_t);
-    }
-    default:
-      ESP_LOGE(TAG, "Unsupported bits per sample: %d", this->bits_per_sample_);
-      return 0;
-  }
-#else
-#ifndef USE_ESP32_VARIANT_ESP32
-  // For newer ESP32 variants 8 bit data needs to be extended to 16 bit.
-  if (this->slot_bit_width_ == I2S_SLOT_BIT_WIDTH_8BIT) {
-    size_t samples_read = bytes_read / sizeof(int8_t);
-    for (size_t i = samples_read - 1; i >= 0; i--) {
-      int16_t temp = static_cast<int16_t>(reinterpret_cast<int8_t *>(buf)[i]) << 8;
-      buf[i] = temp;
-    }
-    return samples_read * sizeof(int16_t);
-  }
-#else
+#if defined(USE_ESP32_VARIANT_ESP32) and not defined(USE_I2S_LEGACY)
   // For ESP32 8/16 bit standard mono mode samples need to be switched.
   if (this->slot_mode_ == I2S_SLOT_MODE_MONO && this->slot_bit_width_ <= 16 && !this->pdm_) {
     size_t samples_read = bytes_read / sizeof(int16_t);
@@ -346,31 +409,62 @@ size_t I2SAudioMicrophone::read(int16_t *buf, size_t len, TickType_t ticks_to_wa
   }
 #endif
   return bytes_read;
-#endif
-}
-
-void I2SAudioMicrophone::read_() {
-  std::vector<int16_t> samples;
-  samples.resize(BUFFER_SIZE);
-  size_t bytes_read = this->read(samples.data(), BUFFER_SIZE * sizeof(int16_t), 0);
-  samples.resize(bytes_read / sizeof(int16_t));
-  this->data_callbacks_.call(samples);
 }
 
 void I2SAudioMicrophone::loop() {
+  uint32_t event_group_bits = xEventGroupGetBits(this->event_group_);
+
+  if (event_group_bits & MicrophoneEventGroupBits::TASK_STARTING) {
+    ESP_LOGD(TAG, "Task has started, attempting to setup I2S audio driver");
+    xEventGroupClearBits(this->event_group_, MicrophoneEventGroupBits::TASK_STARTING);
+  }
+
+  if (event_group_bits & MicrophoneEventGroupBits::TASK_RUNNING) {
+    ESP_LOGD(TAG, "Task is running and reading data");
+
+    xEventGroupClearBits(this->event_group_, MicrophoneEventGroupBits::TASK_RUNNING);
+    this->state_ = microphone::STATE_RUNNING;
+  }
+
+  if (event_group_bits & MicrophoneEventGroupBits::TASK_STOPPING) {
+    ESP_LOGD(TAG, "Task is stopping, attempting to unload the I2S audio driver");
+    xEventGroupClearBits(this->event_group_, MicrophoneEventGroupBits::TASK_STOPPING);
+  }
+
+  if ((event_group_bits & MicrophoneEventGroupBits::TASK_STOPPED)) {
+    ESP_LOGD(TAG, "Task is finished, freeing resources");
+    vTaskDelete(this->task_handle_);
+    this->task_handle_ = nullptr;
+    xEventGroupClearBits(this->event_group_, ALL_BITS);
+    this->state_ = microphone::STATE_STOPPED;
+  }
+
+  if ((uxSemaphoreGetCount(this->active_listeners_semaphore_) < MAX_LISTENERS) &&
+      (this->state_ == microphone::STATE_STOPPED)) {
+    this->state_ = microphone::STATE_STARTING;
+  }
+  if ((uxSemaphoreGetCount(this->active_listeners_semaphore_) == MAX_LISTENERS) &&
+      (this->state_ == microphone::STATE_RUNNING)) {
+    this->state_ = microphone::STATE_STOPPING;
+  }
+
   switch (this->state_) {
-    case microphone::STATE_STOPPED:
-      break;
     case microphone::STATE_STARTING:
-      this->start_();
-      break;
-    case microphone::STATE_RUNNING:
-      if (this->data_callbacks_.size() > 0) {
-        this->read_();
+      if ((this->task_handle_ == nullptr) && !this->status_has_error()) {
+        xTaskCreate(I2SAudioMicrophone::mic_task, "mic_task", TASK_STACK_SIZE, (void *) this, TASK_PRIORITY,
+                    &this->task_handle_);
+
+        if (this->task_handle_ == nullptr) {
+          this->status_momentary_error("Task failed to start, attempting again in 1 second", 1000);
+        }
       }
       break;
+    case microphone::STATE_RUNNING:
+      break;
     case microphone::STATE_STOPPING:
-      this->stop_();
+      xEventGroupSetBits(this->event_group_, MicrophoneEventGroupBits::COMMAND_STOP);
+      break;
+    case microphone::STATE_STOPPED:
       break;
   }
 }

esphome/components/i2s_audio/microphone/i2s_audio_microphone.h

@@ -7,6 +7,9 @@
 #include "esphome/components/microphone/microphone.h"
 #include "esphome/core/component.h"
 
+#include <freertos/event_groups.h>
+#include <freertos/semphr.h>
+
 namespace esphome {
 namespace i2s_audio {
 
@@ -25,9 +28,6 @@ class I2SAudioMicrophone : public I2SAudioIn, public microphone::Microphone, pub
 
   void set_pdm(bool pdm) { this->pdm_ = pdm; }
 
-  size_t read(int16_t *buf, size_t len, TickType_t ticks_to_wait);
-  size_t read(int16_t *buf, size_t len) override { return this->read(buf, len, pdMS_TO_TICKS(100)); }
-
 #ifdef USE_I2S_LEGACY
 #if SOC_I2S_SUPPORTS_ADC
   void set_adc_channel(adc1_channel_t channel) {
@@ -38,9 +38,17 @@ class I2SAudioMicrophone : public I2SAudioIn, public microphone::Microphone, pub
 #endif
 
  protected:
-  void start_();
-  void stop_();
-  void read_();
+  bool start_driver_();
+  void stop_driver_();
+
+  size_t read_(uint8_t *buf, size_t len, TickType_t ticks_to_wait);
+
+  static void mic_task(void *params);
+
+  SemaphoreHandle_t active_listeners_semaphore_{nullptr};
+  EventGroupHandle_t event_group_{nullptr};
+
+  TaskHandle_t task_handle_{nullptr};
 
 #ifdef USE_I2S_LEGACY
   int8_t din_pin_{I2S_PIN_NO_CHANGE};
@@ -53,8 +61,6 @@ class I2SAudioMicrophone : public I2SAudioIn, public microphone::Microphone, pub
   i2s_chan_handle_t rx_handle_;
 #endif
   bool pdm_{false};
-
-  HighFrequencyLoopRequester high_freq_;
 };
 
 }  // namespace i2s_audio

esphome/components/i2s_audio/speaker/__init__.py

@@ -27,6 +27,7 @@ from .. import (
     i2s_audio_ns,
     register_i2s_audio_component,
     use_legacy,
+    validate_mclk_divisible_by_3,
 )
 
 AUTO_LOAD = ["audio"]
@@ -155,6 +156,7 @@ CONFIG_SCHEMA = cv.All(
     _validate_esp32_variant,
     _set_num_channels_from_config,
     _set_stream_limits,
+    validate_mclk_divisible_by_3,
 )
 
 

esphome/components/i2s_audio/speaker/i2s_audio_speaker.cpp

@@ -545,7 +545,7 @@ esp_err_t I2SAudioSpeaker::start_i2s_driver_(audio::AudioStreamInfo &audio_strea
     .use_apll = this->use_apll_,
     .tx_desc_auto_clear = true,
     .fixed_mclk = I2S_PIN_NO_CHANGE,
-    .mclk_multiple = I2S_MCLK_MULTIPLE_256,
+    .mclk_multiple = this->mclk_multiple_,
     .bits_per_chan = this->bits_per_channel_,
 #if SOC_I2S_SUPPORTS_TDM
     .chan_mask = (i2s_channel_t) (I2S_TDM_ACTIVE_CH0 | I2S_TDM_ACTIVE_CH1),
@@ -614,7 +614,7 @@ esp_err_t I2SAudioSpeaker::start_i2s_driver_(audio::AudioStreamInfo &audio_strea
   i2s_std_clk_config_t clk_cfg = {
       .sample_rate_hz = audio_stream_info.get_sample_rate(),
       .clk_src = clk_src,
-      .mclk_multiple = I2S_MCLK_MULTIPLE_256,
+      .mclk_multiple = this->mclk_multiple_,
   };
 
   i2s_slot_mode_t slot_mode = this->slot_mode_;

esphome/components/ili9xxx/ili9xxx_init.h

@@ -388,7 +388,7 @@ static const uint8_t PROGMEM INITCMD_GC9D01N[] = {
   0x8D, 1, 0xFF,
   0x8E, 1, 0xFF,
   0x8F, 1, 0xFF,
-  0X3A, 1, 0x05,    // COLMOD: Pixel Format Set (3Ah) MCU interface, 16 bits / pixel
+  0x3A, 1, 0x05,    // COLMOD: Pixel Format Set (3Ah) MCU interface, 16 bits / pixel
   0xEC, 1, 0x01,    // Inversion (ECh) DINV=1+2H1V column for Dual Gate (BFh=0)
                     // According to datasheet Inversion (ECh) value 0x01 isn't valid, but Lilygo uses it everywhere
   0x74, 7, 0x02, 0x0E, 0x00, 0x00, 0x00, 0x00, 0x00,

esphome/components/image/__init__.py

@@ -286,9 +286,18 @@ CONF_TRANSPARENCY = "transparency"
 IMAGE_DOWNLOAD_TIMEOUT = 30  # seconds
 
 SOURCE_LOCAL = "local"
-SOURCE_MDI = "mdi"
 SOURCE_WEB = "web"
 
+SOURCE_MDI = "mdi"
+SOURCE_MDIL = "mdil"
+SOURCE_MEMORY = "memory"
+
+MDI_SOURCES = {
+    SOURCE_MDI: "https://raw.githubusercontent.com/Templarian/MaterialDesign/master/svg/",
+    SOURCE_MDIL: "https://raw.githubusercontent.com/Pictogrammers/MaterialDesignLight/refs/heads/master/svg/",
+    SOURCE_MEMORY: "https://raw.githubusercontent.com/Pictogrammers/Memory/refs/heads/main/src/svg/",
+}
+
 Image_ = image_ns.class_("Image")
 
 INSTANCE_TYPE = Image_
@@ -313,12 +322,12 @@ def download_file(url, path):
     return str(path)
 
 
-def download_mdi(value):
+def download_gh_svg(value, source):
     mdi_id = value[CONF_ICON] if isinstance(value, dict) else value
-    base_dir = external_files.compute_local_file_dir(DOMAIN) / "mdi"
+    base_dir = external_files.compute_local_file_dir(DOMAIN) / source
     path = base_dir / f"{mdi_id}.svg"
 
-    url = f"https://raw.githubusercontent.com/Templarian/MaterialDesign/master/svg/{mdi_id}.svg"
+    url = MDI_SOURCES[source] + mdi_id + ".svg"
     return download_file(url, path)
 
 
@@ -353,12 +362,12 @@ def validate_cairosvg_installed():
 
 def validate_file_shorthand(value):
     value = cv.string_strict(value)
-    if value.startswith("mdi:"):
-        match = re.search(r"mdi:([a-zA-Z0-9\-]+)", value)
+    parts = value.strip().split(":")
+    if len(parts) == 2 and parts[0] in MDI_SOURCES:
+        match = re.match(r"[a-zA-Z0-9\-]+", parts[1])
         if match is None:
-            raise cv.Invalid("Could not parse mdi icon name.")
-        icon = match.group(1)
-        return download_mdi(icon)
+            raise cv.Invalid(f"Could not parse mdi icon name from '{value}'.")
+        return download_gh_svg(parts[1], parts[0])
 
     if value.startswith("http://") or value.startswith("https://"):
         return download_image(value)
@@ -374,12 +383,20 @@ LOCAL_SCHEMA = cv.All(
     local_path,
 )
 
-MDI_SCHEMA = cv.All(
-    {
-        cv.Required(CONF_ICON): cv.string,
-    },
-    download_mdi,
-)
+
+def mdi_schema(source):
+    def validate_mdi(value):
+        return download_gh_svg(value, source)
+
+    return cv.All(
+        cv.Schema(
+            {
+                cv.Required(CONF_ICON): cv.string,
+            }
+        ),
+        validate_mdi,
+    )
+
 
 WEB_SCHEMA = cv.All(
     {
@@ -388,12 +405,13 @@ WEB_SCHEMA = cv.All(
     download_image,
 )
 
+
 TYPED_FILE_SCHEMA = cv.typed_schema(
     {
         SOURCE_LOCAL: LOCAL_SCHEMA,
-        SOURCE_MDI: MDI_SCHEMA,
         SOURCE_WEB: WEB_SCHEMA,
-    },
+    }
+    | {source: mdi_schema(source) for source in MDI_SOURCES},
     key=CONF_SOURCE,
 )
 

esphome/components/image/image.cpp

@@ -6,10 +6,27 @@ namespace esphome {
 namespace image {
 
 void Image::draw(int x, int y, display::Display *display, Color color_on, Color color_off) {
+  int img_x0 = 0;
+  int img_y0 = 0;
+  int w = width_;
+  int h = height_;
+
+  auto clipping = display->get_clipping();
+  if (clipping.is_set()) {
+    if (clipping.x > x)
+      img_x0 += clipping.x - x;
+    if (clipping.y > y)
+      img_y0 += clipping.y - y;
+    if (w > clipping.x2() - x)
+      w = clipping.x2() - x;
+    if (h > clipping.y2() - y)
+      h = clipping.y2() - y;
+  }
+
   switch (type_) {
     case IMAGE_TYPE_BINARY: {
-      for (int img_x = 0; img_x < width_; img_x++) {
-        for (int img_y = 0; img_y < height_; img_y++) {
+      for (int img_x = img_x0; img_x < w; img_x++) {
+        for (int img_y = img_y0; img_y < h; img_y++) {
           if (this->get_binary_pixel_(img_x, img_y)) {
             display->draw_pixel_at(x + img_x, y + img_y, color_on);
           } else if (!this->transparency_) {
@@ -20,8 +37,8 @@ void Image::draw(int x, int y, display::Display *display, Color color_on, Color
       break;
     }
     case IMAGE_TYPE_GRAYSCALE:
-      for (int img_x = 0; img_x < width_; img_x++) {
-        for (int img_y = 0; img_y < height_; img_y++) {
+      for (int img_x = img_x0; img_x < w; img_x++) {
+        for (int img_y = img_y0; img_y < h; img_y++) {
           const uint32_t pos = (img_x + img_y * this->width_);
           const uint8_t gray = progmem_read_byte(this->data_start_ + pos);
           Color color = Color(gray, gray, gray, 0xFF);
@@ -47,8 +64,8 @@ void Image::draw(int x, int y, display::Display *display, Color color_on, Color
       }
       break;
     case IMAGE_TYPE_RGB565:
-      for (int img_x = 0; img_x < width_; img_x++) {
-        for (int img_y = 0; img_y < height_; img_y++) {
+      for (int img_x = img_x0; img_x < w; img_x++) {
+        for (int img_y = img_y0; img_y < h; img_y++) {
           auto color = this->get_rgb565_pixel_(img_x, img_y);
           if (color.w >= 0x80) {
             display->draw_pixel_at(x + img_x, y + img_y, color);
@@ -57,8 +74,8 @@ void Image::draw(int x, int y, display::Display *display, Color color_on, Color
       }
       break;
     case IMAGE_TYPE_RGB:
-      for (int img_x = 0; img_x < width_; img_x++) {
-        for (int img_y = 0; img_y < height_; img_y++) {
+      for (int img_x = img_x0; img_x < w; img_x++) {
+        for (int img_y = img_y0; img_y < h; img_y++) {
           auto color = this->get_rgb_pixel_(img_x, img_y);
           if (color.w >= 0x80) {
             display->draw_pixel_at(x + img_x, y + img_y, color);

esphome/components/ld2410/ld2410.h

@@ -129,7 +129,7 @@ enum PeriodicDataStructure : uint8_t {
   LIGHT_SENSOR = 37,
   OUT_PIN_SENSOR = 38,
 };
-enum PeriodicDataValue : uint8_t { HEAD = 0XAA, END = 0x55, CHECK = 0x00 };
+enum PeriodicDataValue : uint8_t { HEAD = 0xAA, END = 0x55, CHECK = 0x00 };
 
 enum AckDataStructure : uint8_t { COMMAND = 6, COMMAND_STATUS = 7 };
 

esphome/components/ld2450/ld2450.h

@@ -105,7 +105,7 @@ enum PeriodicDataStructure : uint8_t {
   TARGET_RESOLUTION = 10,
 };
 
-enum PeriodicDataValue : uint8_t { HEAD = 0XAA, END = 0x55, CHECK = 0x00 };
+enum PeriodicDataValue : uint8_t { HEAD = 0xAA, END = 0x55, CHECK = 0x00 };
 
 enum AckDataStructure : uint8_t { COMMAND = 6, COMMAND_STATUS = 7 };
 

esphome/components/lvgl/__init__.py

@@ -18,13 +18,13 @@ from esphome.const import (
     CONF_TRIGGER_ID,
     CONF_TYPE,
 )
-from esphome.core import CORE, ID
+from esphome.core import CORE, ID, Lambda
 from esphome.cpp_generator import MockObj
 from esphome.final_validate import full_config
 from esphome.helpers import write_file_if_changed
 
 from . import defines as df, helpers, lv_validation as lvalid
-from .automation import disp_update, focused_widgets, update_to_code
+from .automation import disp_update, focused_widgets, refreshed_widgets, update_to_code
 from .defines import add_define
 from .encoders import (
     ENCODERS_CONFIG,
@@ -240,6 +240,13 @@ def final_validation(configs):
                     "A non adjustable arc may not be focused",
                     path,
                 )
+        for w in refreshed_widgets:
+            path = global_config.get_path_for_id(w)
+            widget_conf = global_config.get_config_for_path(path[:-1])
+            if not any(isinstance(v, Lambda) for v in widget_conf.values()):
+                raise cv.Invalid(
+                    f"Widget '{w}' does not have any templated properties to refresh",
+                )
 
 
 async def to_code(configs):

esphome/components/lvgl/automation.py

@@ -35,7 +35,13 @@ from .lvcode import (
     lv_obj,
     lvgl_comp,
 )
-from .schemas import DISP_BG_SCHEMA, LIST_ACTION_SCHEMA, LVGL_SCHEMA, base_update_schema
+from .schemas import (
+    ALL_STYLES,
+    DISP_BG_SCHEMA,
+    LIST_ACTION_SCHEMA,
+    LVGL_SCHEMA,
+    base_update_schema,
+)
 from .types import (
     LV_STATE,
     LvglAction,
@@ -57,6 +63,7 @@ from .widgets import (
 
 # Record widgets that are used in a focused action here
 focused_widgets = set()
+refreshed_widgets = set()
 
 
 async def action_to_code(
@@ -361,3 +368,45 @@ async def obj_update_to_code(config, action_id, template_arg, args):
     return await action_to_code(
         widgets, do_update, action_id, template_arg, args, config
     )
+
+
+def validate_refresh_config(config):
+    for w in config:
+        refreshed_widgets.add(w[CONF_ID])
+    return config
+
+
+@automation.register_action(
+    "lvgl.widget.refresh",
+    ObjUpdateAction,
+    cv.All(
+        cv.ensure_list(
+            cv.maybe_simple_value(
+                {
+                    cv.Required(CONF_ID): cv.use_id(lv_obj_t),
+                },
+                key=CONF_ID,
+            )
+        ),
+        validate_refresh_config,
+    ),
+)
+async def obj_refresh_to_code(config, action_id, template_arg, args):
+    widget = await get_widgets(config)
+
+    async def do_refresh(widget: Widget):
+        # only update style properties that might have changed, i.e. are templated
+        config = {k: v for k, v in widget.config.items() if isinstance(v, Lambda)}
+        await set_obj_properties(widget, config)
+        # must pass all widget-specific options here, even if not templated, but only do so if at least one is
+        # templated. First filter out common style properties.
+        config = {k: v for k, v in widget.config.items() if k not in ALL_STYLES}
+        if any(isinstance(v, Lambda) for v in config.values()):
+            await widget.type.to_code(widget, config)
+            if (
+                widget.type.w_type.value_property is not None
+                and widget.type.w_type.value_property in config
+            ):
+                lv.event_send(widget.obj, UPDATE_EVENT, nullptr)
+
+    return await action_to_code(widget, do_refresh, action_id, template_arg, args)

esphome/components/lvgl/lv_validation.py

@@ -16,7 +16,7 @@ from esphome.const import (
 )
 from esphome.core import CORE, ID, Lambda
 from esphome.cpp_generator import MockObj
-from esphome.cpp_types import ESPTime, uint32
+from esphome.cpp_types import ESPTime, int32, uint32
 from esphome.helpers import cpp_string_escape
 from esphome.schema_extractors import SCHEMA_EXTRACT, schema_extractor
 
@@ -263,6 +263,15 @@ def pixels_validator(value):
 pixels = LValidator(pixels_validator, uint32, retmapper=literal)
 
 
+def padding_validator(value):
+    if isinstance(value, str) and value.lower().endswith("px"):
+        value = value[:-2]
+    return cv.int_(value)
+
+
+padding = LValidator(padding_validator, int32, retmapper=literal)
+
+
 def zoom_validator(value):
     value = cv.float_range(0.1, 10.0)(value)
     return value

esphome/components/lvgl/schemas.py

@@ -156,13 +156,13 @@ STYLE_PROPS = {
     "opa_layered": lvalid.opacity,
     "outline_color": lvalid.lv_color,
     "outline_opa": lvalid.opacity,
-    "outline_pad": lvalid.pixels,
+    "outline_pad": lvalid.padding,
     "outline_width": lvalid.pixels,
-    "pad_all": lvalid.pixels,
-    "pad_bottom": lvalid.pixels,
-    "pad_left": lvalid.pixels,
-    "pad_right": lvalid.pixels,
-    "pad_top": lvalid.pixels,
+    "pad_all": lvalid.padding,
+    "pad_bottom": lvalid.padding,
+    "pad_left": lvalid.padding,
+    "pad_right": lvalid.padding,
+    "pad_top": lvalid.padding,
     "shadow_color": lvalid.lv_color,
     "shadow_ofs_x": lvalid.lv_int,
     "shadow_ofs_y": lvalid.lv_int,
@@ -226,8 +226,8 @@ FULL_STYLE_SCHEMA = STYLE_SCHEMA.extend(
     {
         cv.Optional(df.CONF_GRID_CELL_X_ALIGN): grid_alignments,
         cv.Optional(df.CONF_GRID_CELL_Y_ALIGN): grid_alignments,
-        cv.Optional(df.CONF_PAD_ROW): lvalid.pixels,
-        cv.Optional(df.CONF_PAD_COLUMN): lvalid.pixels,
+        cv.Optional(df.CONF_PAD_ROW): lvalid.padding,
+        cv.Optional(df.CONF_PAD_COLUMN): lvalid.padding,
     }
 )
 
@@ -370,8 +370,8 @@ LAYOUT_SCHEMA = {
                 cv.Required(df.CONF_GRID_COLUMNS): [grid_spec],
                 cv.Optional(df.CONF_GRID_COLUMN_ALIGN): grid_alignments,
                 cv.Optional(df.CONF_GRID_ROW_ALIGN): grid_alignments,
-                cv.Optional(df.CONF_PAD_ROW): lvalid.pixels,
-                cv.Optional(df.CONF_PAD_COLUMN): lvalid.pixels,
+                cv.Optional(df.CONF_PAD_ROW): lvalid.padding,
+                cv.Optional(df.CONF_PAD_COLUMN): lvalid.padding,
             },
             df.TYPE_FLEX: {
                 cv.Optional(
@@ -380,8 +380,8 @@ LAYOUT_SCHEMA = {
                 cv.Optional(df.CONF_FLEX_ALIGN_MAIN, default="start"): flex_alignments,
                 cv.Optional(df.CONF_FLEX_ALIGN_CROSS, default="start"): flex_alignments,
                 cv.Optional(df.CONF_FLEX_ALIGN_TRACK, default="start"): flex_alignments,
-                cv.Optional(df.CONF_PAD_ROW): lvalid.pixels,
-                cv.Optional(df.CONF_PAD_COLUMN): lvalid.pixels,
+                cv.Optional(df.CONF_PAD_ROW): lvalid.padding,
+                cv.Optional(df.CONF_PAD_COLUMN): lvalid.padding,
             },
         },
         lower=True,
@@ -427,8 +427,8 @@ ALL_STYLES = {
     **STYLE_PROPS,
     **GRID_CELL_SCHEMA,
     **FLEX_OBJ_SCHEMA,
-    cv.Optional(df.CONF_PAD_ROW): lvalid.pixels,
-    cv.Optional(df.CONF_PAD_COLUMN): lvalid.pixels,
+    cv.Optional(df.CONF_PAD_ROW): lvalid.padding,
+    cv.Optional(df.CONF_PAD_COLUMN): lvalid.padding,
 }
 
 

esphome/components/lvgl/widgets/buttonmatrix.py

@@ -19,7 +19,7 @@ from ..defines import (
     CONF_SELECTED,
 )
 from ..helpers import lvgl_components_required
-from ..lv_validation import key_code, lv_bool, pixels
+from ..lv_validation import key_code, lv_bool, padding
 from ..lvcode import lv, lv_add, lv_expr
 from ..schemas import automation_schema
 from ..types import (
@@ -59,8 +59,8 @@ BUTTONMATRIX_BUTTON_SCHEMA = cv.Schema(
 BUTTONMATRIX_SCHEMA = cv.Schema(
     {
         cv.Optional(CONF_ONE_CHECKED, default=False): lv_bool,
-        cv.Optional(CONF_PAD_ROW): pixels,
-        cv.Optional(CONF_PAD_COLUMN): pixels,
+        cv.Optional(CONF_PAD_ROW): padding,
+        cv.Optional(CONF_PAD_COLUMN): padding,
         cv.GenerateID(CONF_BUTTON_TEXT_LIST_ID): cv.declare_id(char_ptr),
         cv.Required(CONF_ROWS): cv.ensure_list(
             cv.Schema(

esphome/components/lvgl/widgets/checkbox.py

@@ -2,7 +2,7 @@ from esphome.config_validation import Optional
 from esphome.const import CONF_TEXT
 
 from ..defines import CONF_INDICATOR, CONF_MAIN, CONF_PAD_COLUMN
-from ..lv_validation import lv_text, pixels
+from ..lv_validation import lv_text, padding
 from ..lvcode import lv
 from ..schemas import TEXT_SCHEMA
 from ..types import LvBoolean
@@ -19,7 +19,7 @@ class CheckboxType(WidgetType):
             (CONF_MAIN, CONF_INDICATOR),
             TEXT_SCHEMA.extend(
                 {
-                    Optional(CONF_PAD_COLUMN): pixels,
+                    Optional(CONF_PAD_COLUMN): padding,
                 }
             ),
         )

esphome/components/lvgl/widgets/img.py

@@ -10,7 +10,7 @@ from ..defines import (
     CONF_ZOOM,
     LvConstant,
 )
-from ..lv_validation import angle, lv_bool, lv_image, size, zoom
+from ..lv_validation import lv_angle, lv_bool, lv_image, size, zoom
 from ..lvcode import lv
 from ..types import lv_img_t
 from . import Widget, WidgetType
@@ -22,7 +22,7 @@ BASE_IMG_SCHEMA = cv.Schema(
     {
         cv.Optional(CONF_PIVOT_X): size,
         cv.Optional(CONF_PIVOT_Y): size,
-        cv.Optional(CONF_ANGLE): angle,
+        cv.Optional(CONF_ANGLE): lv_angle,
         cv.Optional(CONF_ZOOM): zoom,
         cv.Optional(CONF_OFFSET_X): size,
         cv.Optional(CONF_OFFSET_Y): size,
@@ -66,17 +66,19 @@ class ImgType(WidgetType):
         if (pivot_x := config.get(CONF_PIVOT_X)) and (
             pivot_y := config.get(CONF_PIVOT_Y)
         ):
-            lv.img_set_pivot(w.obj, pivot_x, pivot_y)
+            lv.img_set_pivot(
+                w.obj, await size.process(pivot_x), await size.process(pivot_y)
+            )
         if (cf_angle := config.get(CONF_ANGLE)) is not None:
-            lv.img_set_angle(w.obj, cf_angle)
+            lv.img_set_angle(w.obj, await lv_angle.process(cf_angle))
         if (img_zoom := config.get(CONF_ZOOM)) is not None:
-            lv.img_set_zoom(w.obj, img_zoom)
+            lv.img_set_zoom(w.obj, await zoom.process(img_zoom))
         if (offset := config.get(CONF_OFFSET_X)) is not None:
-            lv.img_set_offset_x(w.obj, offset)
+            lv.img_set_offset_x(w.obj, await size.process(offset))
         if (offset := config.get(CONF_OFFSET_Y)) is not None:
-            lv.img_set_offset_y(w.obj, offset)
+            lv.img_set_offset_y(w.obj, await size.process(offset))
         if CONF_ANTIALIAS in config:
-            lv.img_set_antialias(w.obj, config[CONF_ANTIALIAS])
+            lv.img_set_antialias(w.obj, await lv_bool.process(config[CONF_ANTIALIAS]))
         if mode := config.get(CONF_MODE):
             await w.set_property("size_mode", mode)
 

esphome/components/max7219digit/max7219digit.cpp

@@ -4,6 +4,8 @@
 #include "esphome/core/hal.h"
 #include "max7219font.h"
 
+#include <algorithm>
+
 namespace esphome {
 namespace max7219digit {
 
@@ -61,45 +63,42 @@ void MAX7219Component::dump_config() {
 }
 
 void MAX7219Component::loop() {
-  uint32_t now = millis();
-
+  const uint32_t now = millis();
+  const uint32_t millis_since_last_scroll = now - this->last_scroll_;
+  const size_t first_line_size = this->max_displaybuffer_[0].size();
   // check if the buffer has shrunk past the current position since last update
-  if ((this->max_displaybuffer_[0].size() >= this->old_buffer_size_ + 3) ||
-      (this->max_displaybuffer_[0].size() <= this->old_buffer_size_ - 3)) {
+  if ((first_line_size >= this->old_buffer_size_ + 3) || (first_line_size <= this->old_buffer_size_ - 3)) {
+    ESP_LOGV(TAG, "Buffer size changed %d to %d", this->old_buffer_size_, first_line_size);
     this->stepsleft_ = 0;
     this->display();
-    this->old_buffer_size_ = this->max_displaybuffer_[0].size();
+    this->old_buffer_size_ = first_line_size;
   }
 
-  // Reset the counter back to 0 when full string has been displayed.
-  if (this->stepsleft_ > this->max_displaybuffer_[0].size())
-    this->stepsleft_ = 0;
-
-  // Return if there is no need to scroll or scroll is off
-  if (!this->scroll_ || (this->max_displaybuffer_[0].size() <= (size_t) get_width_internal())) {
+  if (!this->scroll_ || (first_line_size <= (size_t) get_width_internal())) {
+    ESP_LOGVV(TAG, "Return if there is no need to scroll or scroll is off.");
     this->display();
     return;
   }
 
-  if ((this->stepsleft_ == 0) && (now - this->last_scroll_ < this->scroll_delay_)) {
+  if ((this->stepsleft_ == 0) && (millis_since_last_scroll < this->scroll_delay_)) {
+    ESP_LOGVV(TAG, "At first step. Waiting for scroll delay");
     this->display();
     return;
   }
 
-  // Dwell time at end of string in case of stop at end
   if (this->scroll_mode_ == ScrollMode::STOP) {
-    if (this->stepsleft_ >= this->max_displaybuffer_[0].size() - (size_t) get_width_internal() + 1) {
-      if (now - this->last_scroll_ >= this->scroll_dwell_) {
-        this->stepsleft_ = 0;
-        this->last_scroll_ = now;
-        this->display();
+    if (this->stepsleft_ + get_width_internal() == first_line_size + 1) {
+      if (millis_since_last_scroll < this->scroll_dwell_) {
+        ESP_LOGVV(TAG, "Dwell time at end of string in case of stop at end. Step %d, since last scroll %d, dwell %d.",
+                  this->stepsleft_, millis_since_last_scroll, this->scroll_dwell_);
+        return;
       }
-      return;
+      ESP_LOGV(TAG, "Dwell time passed. Continue scrolling.");
     }
   }
 
-  // Actual call to scroll left action
-  if (now - this->last_scroll_ >= this->scroll_speed_) {
+  if (millis_since_last_scroll >= this->scroll_speed_) {
+    ESP_LOGVV(TAG, "Call to scroll left action");
     this->last_scroll_ = now;
     this->scroll_left();
     this->display();
@@ -227,19 +226,20 @@ void MAX7219Component::scroll(bool on_off) { this->set_scroll(on_off); }
 
 void MAX7219Component::scroll_left() {
   for (int chip_line = 0; chip_line < this->num_chip_lines_; chip_line++) {
+    auto scroll = [&](std::vector<uint8_t> &line, uint16_t steps) {
+      std::rotate(line.begin(), std::next(line.begin(), steps), line.end());
+    };
     if (this->update_) {
       this->max_displaybuffer_[chip_line].push_back(this->bckgrnd_);
-      for (uint16_t i = 0; i < this->stepsleft_; i++) {
-        this->max_displaybuffer_[chip_line].push_back(this->max_displaybuffer_[chip_line].front());
-        this->max_displaybuffer_[chip_line].erase(this->max_displaybuffer_[chip_line].begin());
-      }
+      scroll(this->max_displaybuffer_[chip_line],
+             (this->stepsleft_ + 1) % (this->max_displaybuffer_[chip_line].size()));
     } else {
-      this->max_displaybuffer_[chip_line].push_back(this->max_displaybuffer_[chip_line].front());
-      this->max_displaybuffer_[chip_line].erase(this->max_displaybuffer_[chip_line].begin());
+      scroll(this->max_displaybuffer_[chip_line], 1);
     }
   }
   this->update_ = false;
   this->stepsleft_++;
+  this->stepsleft_ %= this->max_displaybuffer_[0].size();
 }
 
 void MAX7219Component::send_char(uint8_t chip, uint8_t data) {

esphome/components/mdns/__init__.py

@@ -35,8 +35,8 @@ SERVICE_SCHEMA = cv.Schema(
     {
         cv.Required(CONF_SERVICE): cv.string,
         cv.Required(CONF_PROTOCOL): cv.string,
-        cv.Optional(CONF_PORT, default=0): cv.Any(0, cv.port),
-        cv.Optional(CONF_TXT, default={}): {cv.string: cv.string},
+        cv.Optional(CONF_PORT, default=0): cv.templatable(cv.Any(0, cv.port)),
+        cv.Optional(CONF_TXT, default={}): {cv.string: cv.templatable(cv.string)},
     }
 )
 
@@ -102,12 +102,18 @@ async def to_code(config):
 
     for service in config[CONF_SERVICES]:
         txt = [
-            mdns_txt_record(txt_key, txt_value)
+            cg.StructInitializer(
+                MDNSTXTRecord,
+                ("key", txt_key),
+                ("value", await cg.templatable(txt_value, [], cg.std_string)),
+            )
             for txt_key, txt_value in service[CONF_TXT].items()
         ]
-
         exp = mdns_service(
-            service[CONF_SERVICE], service[CONF_PROTOCOL], service[CONF_PORT], txt
+            service[CONF_SERVICE],
+            service[CONF_PROTOCOL],
+            await cg.templatable(service[CONF_PORT], [], cg.uint16),
+            txt,
         )
 
         cg.add(var.add_extra_service(exp))

esphome/components/mdns/mdns_component.cpp

@@ -121,9 +121,11 @@ void MDNSComponent::dump_config() {
   ESP_LOGCONFIG(TAG, "  Hostname: %s", this->hostname_.c_str());
   ESP_LOGV(TAG, "  Services:");
   for (const auto &service : this->services_) {
-    ESP_LOGV(TAG, "  - %s, %s, %d", service.service_type.c_str(), service.proto.c_str(), service.port);
+    ESP_LOGV(TAG, "  - %s, %s, %d", service.service_type.c_str(), service.proto.c_str(),
+             const_cast<TemplatableValue<uint16_t> &>(service.port).value());
     for (const auto &record : service.txt_records) {
-      ESP_LOGV(TAG, "    TXT: %s = %s", record.key.c_str(), record.value.c_str());
+      ESP_LOGV(TAG, "    TXT: %s = %s", record.key.c_str(),
+               const_cast<TemplatableValue<std::string> &>(record.value).value().c_str());
     }
   }
 }

esphome/components/mdns/mdns_component.h

@@ -3,6 +3,7 @@
 #ifdef USE_MDNS
 #include <string>
 #include <vector>
+#include "esphome/core/automation.h"
 #include "esphome/core/component.h"
 
 namespace esphome {
@@ -10,7 +11,7 @@ namespace mdns {
 
 struct MDNSTXTRecord {
   std::string key;
-  std::string value;
+  TemplatableValue<std::string> value;
 };
 
 struct MDNSService {
@@ -20,7 +21,7 @@ struct MDNSService {
   // second label indicating protocol _including_ underscore character prefix
   // as defined in RFC6763 Section 7, like "_tcp" or "_udp"
   std::string proto;
-  uint16_t port;
+  TemplatableValue<uint16_t> port;
   std::vector<MDNSTXTRecord> txt_records;
 };
 

esphome/components/mdns/mdns_esp32.cpp

@@ -31,11 +31,12 @@ void MDNSComponent::setup() {
       mdns_txt_item_t it{};
       // dup strings to ensure the pointer is valid even after the record loop
       it.key = strdup(record.key.c_str());
-      it.value = strdup(record.value.c_str());
+      it.value = strdup(const_cast<TemplatableValue<std::string> &>(record.value).value().c_str());
       txt_records.push_back(it);
     }
-    err = mdns_service_add(nullptr, service.service_type.c_str(), service.proto.c_str(), service.port,
-                           txt_records.data(), txt_records.size());
+    uint16_t port = const_cast<TemplatableValue<uint16_t> &>(service.port).value();
+    err = mdns_service_add(nullptr, service.service_type.c_str(), service.proto.c_str(), port, txt_records.data(),
+                           txt_records.size());
 
     // free records
     for (const auto &it : txt_records) {

esphome/components/mdns/mdns_esp8266.cpp

@@ -29,9 +29,11 @@ void MDNSComponent::setup() {
     while (*service_type == '_') {
       service_type++;
     }
-    MDNS.addService(service_type, proto, service.port);
+    uint16_t port = const_cast<TemplatableValue<uint16_t> &>(service.port).value();
+    MDNS.addService(service_type, proto, port);
     for (const auto &record : service.txt_records) {
-      MDNS.addServiceTxt(service_type, proto, record.key.c_str(), record.value.c_str());
+      MDNS.addServiceTxt(service_type, proto, record.key.c_str(),
+                         const_cast<TemplatableValue<std::string> &>(record.value).value().c_str());
     }
   }
 }

esphome/components/mdns/mdns_libretiny.cpp

@@ -29,9 +29,11 @@ void MDNSComponent::setup() {
     while (*service_type == '_') {
       service_type++;
     }
-    MDNS.addService(service_type, proto, service.port);
+    uint16_t port_ = const_cast<TemplatableValue<uint16_t> &>(service.port).value();
+    MDNS.addService(service_type, proto, port_);
     for (const auto &record : service.txt_records) {
-      MDNS.addServiceTxt(service_type, proto, record.key.c_str(), record.value.c_str());
+      MDNS.addServiceTxt(service_type, proto, record.key.c_str(),
+                         const_cast<TemplatableValue<std::string> &>(record.value).value().c_str());
     }
   }
 }

esphome/components/mdns/mdns_rp2040.cpp

@@ -29,9 +29,11 @@ void MDNSComponent::setup() {
     while (*service_type == '_') {
       service_type++;
     }
-    MDNS.addService(service_type, proto, service.port);
+    uint16_t port = const_cast<TemplatableValue<uint16_t> &>(service.port).value();
+    MDNS.addService(service_type, proto, port);
     for (const auto &record : service.txt_records) {
-      MDNS.addServiceTxt(service_type, proto, record.key.c_str(), record.value.c_str());
+      MDNS.addServiceTxt(service_type, proto, record.key.c_str(),
+                         const_cast<TemplatableValue<std::string> &>(record.value).value().c_str());
     }
   }
 }

esphome/components/media_player/__init__.py

@@ -134,11 +134,13 @@ MEDIA_PLAYER_SCHEMA = cv.ENTITY_BASE_SCHEMA.extend(
 )
 
 
-MEDIA_PLAYER_ACTION_SCHEMA = cv.Schema(
-    {
-        cv.GenerateID(): cv.use_id(MediaPlayer),
-        cv.Optional(CONF_ANNOUNCEMENT, default=False): cv.templatable(cv.boolean),
-    }
+MEDIA_PLAYER_ACTION_SCHEMA = automation.maybe_simple_id(
+    cv.Schema(
+        {
+            cv.GenerateID(): cv.use_id(MediaPlayer),
+            cv.Optional(CONF_ANNOUNCEMENT, default=False): cv.templatable(cv.boolean),
+        }
+    )
 )
 
 MEDIA_PLAYER_CONDITION_SCHEMA = automation.maybe_simple_id(

esphome/components/micro_wake_word/__init__.py

@@ -12,6 +12,7 @@ import esphome.config_validation as cv
 from esphome.const import (
     CONF_FILE,
     CONF_ID,
+    CONF_INTERNAL,
     CONF_MICROPHONE,
     CONF_MODEL,
     CONF_PASSWORD,
@@ -40,6 +41,7 @@ CONF_ON_WAKE_WORD_DETECTED = "on_wake_word_detected"
 CONF_PROBABILITY_CUTOFF = "probability_cutoff"
 CONF_SLIDING_WINDOW_AVERAGE_SIZE = "sliding_window_average_size"
 CONF_SLIDING_WINDOW_SIZE = "sliding_window_size"
+CONF_STOP_AFTER_DETECTION = "stop_after_detection"
 CONF_TENSOR_ARENA_SIZE = "tensor_arena_size"
 CONF_VAD = "vad"
 
@@ -49,13 +51,20 @@ micro_wake_word_ns = cg.esphome_ns.namespace("micro_wake_word")
 
 MicroWakeWord = micro_wake_word_ns.class_("MicroWakeWord", cg.Component)
 
+DisableModelAction = micro_wake_word_ns.class_("DisableModelAction", automation.Action)
+EnableModelAction = micro_wake_word_ns.class_("EnableModelAction", automation.Action)
 StartAction = micro_wake_word_ns.class_("StartAction", automation.Action)
 StopAction = micro_wake_word_ns.class_("StopAction", automation.Action)
 
+ModelIsEnabledCondition = micro_wake_word_ns.class_(
+    "ModelIsEnabledCondition", automation.Condition
+)
 IsRunningCondition = micro_wake_word_ns.class_(
     "IsRunningCondition", automation.Condition
 )
 
+WakeWordModel = micro_wake_word_ns.class_("WakeWordModel")
+
 
 def _validate_json_filename(value):
     value = cv.string(value)
@@ -169,9 +178,10 @@ def _convert_manifest_v1_to_v2(v1_manifest):
 
     # Original Inception-based V1 manifest models require a minimum of 45672 bytes
     v2_manifest[KEY_MICRO][CONF_TENSOR_ARENA_SIZE] = 45672
-
     # Original Inception-based V1 manifest models use a 20 ms feature step size
     v2_manifest[KEY_MICRO][CONF_FEATURE_STEP_SIZE] = 20
+    # Original Inception-based V1 manifest models were trained only on TTS English samples
+    v2_manifest[KEY_TRAINED_LANGUAGES] = ["en"]
 
     return v2_manifest
 
@@ -296,14 +306,16 @@ MODEL_SOURCE_SCHEMA = cv.Any(
 
 MODEL_SCHEMA = cv.Schema(
     {
+        cv.GenerateID(CONF_ID): cv.declare_id(WakeWordModel),
         cv.Optional(CONF_MODEL): MODEL_SOURCE_SCHEMA,
         cv.Optional(CONF_PROBABILITY_CUTOFF): cv.percentage,
         cv.Optional(CONF_SLIDING_WINDOW_SIZE): cv.positive_int,
+        cv.Optional(CONF_INTERNAL, default=False): cv.boolean,
         cv.GenerateID(CONF_RAW_DATA_ID): cv.declare_id(cg.uint8),
     }
 )
 
-# Provide a default VAD model that could be overridden
+# Provides a default VAD model that could be overridden
 VAD_MODEL_SCHEMA = MODEL_SCHEMA.extend(
     cv.Schema(
         {
@@ -328,7 +340,14 @@ CONFIG_SCHEMA = cv.All(
     cv.Schema(
         {
             cv.GenerateID(): cv.declare_id(MicroWakeWord),
-            cv.GenerateID(CONF_MICROPHONE): cv.use_id(microphone.Microphone),
+            cv.Optional(
+                CONF_MICROPHONE, default={}
+            ): microphone.microphone_source_schema(
+                min_bits_per_sample=16,
+                max_bits_per_sample=16,
+                min_channels=1,
+                max_channels=1,
+            ),
             cv.Required(CONF_MODELS): cv.ensure_list(
                 cv.maybe_simple_value(MODEL_SCHEMA, key=CONF_MODEL)
             ),
@@ -336,6 +355,7 @@ CONFIG_SCHEMA = cv.All(
                 single=True
             ),
             cv.Optional(CONF_VAD): _maybe_empty_vad_schema,
+            cv.Optional(CONF_STOP_AFTER_DETECTION, default=True): cv.boolean,
             cv.Optional(CONF_MODEL): cv.invalid(
                 f"The {CONF_MODEL} parameter has moved to be a list element under the {CONF_MODELS} parameter."
             ),
@@ -404,39 +424,42 @@ def _feature_step_size_validate(config):
             raise cv.Invalid("Cannot load models with different features step sizes.")
 
 
-FINAL_VALIDATE_SCHEMA = _feature_step_size_validate
+FINAL_VALIDATE_SCHEMA = cv.All(
+    cv.Schema(
+        {
+            cv.Required(
+                CONF_MICROPHONE
+            ): microphone.final_validate_microphone_source_schema(
+                "micro_wake_word", sample_rate=16000
+            ),
+        },
+        extra=cv.ALLOW_EXTRA,
+    ),
+    _feature_step_size_validate,
+)
 
 
 async def to_code(config):
     var = cg.new_Pvariable(config[CONF_ID])
     await cg.register_component(var, config)
 
-    mic = await cg.get_variable(config[CONF_MICROPHONE])
-    cg.add(var.set_microphone(mic))
+    mic_source = await microphone.microphone_source_to_code(config[CONF_MICROPHONE])
+    cg.add(var.set_microphone_source(mic_source))
+
+    cg.add_define("USE_MICRO_WAKE_WORD")
+    cg.add_define("USE_OTA_STATE_CALLBACK")
 
     esp32.add_idf_component(
         name="esp-tflite-micro",
         repo="https://github.com/espressif/esp-tflite-micro",
-        ref="v1.3.1",
-    )
-    # add esp-nn dependency for tflite-micro to work around https://github.com/espressif/esp-nn/issues/17
-    # ...remove after switching to IDF 5.1.4+
-    esp32.add_idf_component(
-        name="esp-nn",
-        repo="https://github.com/espressif/esp-nn",
-        ref="v1.1.0",
+        ref="v1.3.3.1",
     )
 
     cg.add_build_flag("-DTF_LITE_STATIC_MEMORY")
     cg.add_build_flag("-DTF_LITE_DISABLE_X86_NEON")
     cg.add_build_flag("-DESP_NN")
 
-    if on_wake_word_detection_config := config.get(CONF_ON_WAKE_WORD_DETECTED):
-        await automation.build_automation(
-            var.get_wake_word_detected_trigger(),
-            [(cg.std_string, "wake_word")],
-            on_wake_word_detection_config,
-        )
+    cg.add_library("kahrendt/ESPMicroSpeechFeatures", "1.1.0")
 
     if vad_model := config.get(CONF_VAD):
         cg.add_define("USE_MICRO_WAKE_WORD_VAD")
@@ -444,7 +467,7 @@ async def to_code(config):
         # Use the general model loading code for the VAD codegen
         config[CONF_MODELS].append(vad_model)
 
-    for model_parameters in config[CONF_MODELS]:
+    for i, model_parameters in enumerate(config[CONF_MODELS]):
         model_config = model_parameters.get(CONF_MODEL)
         data = []
         manifest, data = _model_config_to_manifest_data(model_config)
@@ -455,6 +478,8 @@ async def to_code(config):
         probability_cutoff = model_parameters.get(
             CONF_PROBABILITY_CUTOFF, manifest[KEY_MICRO][CONF_PROBABILITY_CUTOFF]
         )
+        quantized_probability_cutoff = int(probability_cutoff * 255)
+
         sliding_window_size = model_parameters.get(
             CONF_SLIDING_WINDOW_SIZE,
             manifest[KEY_MICRO][CONF_SLIDING_WINDOW_SIZE],
@@ -464,24 +489,40 @@ async def to_code(config):
             cg.add(
                 var.add_vad_model(
                     prog_arr,
-                    probability_cutoff,
+                    quantized_probability_cutoff,
                     sliding_window_size,
                     manifest[KEY_MICRO][CONF_TENSOR_ARENA_SIZE],
                 )
             )
         else:
-            cg.add(
-                var.add_wake_word_model(
-                    prog_arr,
-                    probability_cutoff,
-                    sliding_window_size,
-                    manifest[KEY_WAKE_WORD],
-                    manifest[KEY_MICRO][CONF_TENSOR_ARENA_SIZE],
-                )
+            # Only enable the first wake word by default. After first boot, the enable state is saved/loaded to the flash
+            default_enabled = i == 0
+            wake_word_model = cg.new_Pvariable(
+                model_parameters[CONF_ID],
+                str(model_parameters[CONF_ID]),
+                prog_arr,
+                quantized_probability_cutoff,
+                sliding_window_size,
+                manifest[KEY_WAKE_WORD],
+                manifest[KEY_MICRO][CONF_TENSOR_ARENA_SIZE],
+                default_enabled,
+                model_parameters[CONF_INTERNAL],
             )
 
+            for lang in manifest[KEY_TRAINED_LANGUAGES]:
+                cg.add(wake_word_model.add_trained_language(lang))
+
+            cg.add(var.add_wake_word_model(wake_word_model))
+
     cg.add(var.set_features_step_size(manifest[KEY_MICRO][CONF_FEATURE_STEP_SIZE]))
-    cg.add_library("kahrendt/ESPMicroSpeechFeatures", "1.1.0")
+    cg.add(var.set_stop_after_detection(config[CONF_STOP_AFTER_DETECTION]))
+
+    if on_wake_word_detection_config := config.get(CONF_ON_WAKE_WORD_DETECTED):
+        await automation.build_automation(
+            var.get_wake_word_detected_trigger(),
+            [(cg.std_string, "wake_word")],
+            on_wake_word_detection_config,
+        )
 
 
 MICRO_WAKE_WORD_ACTION_SCHEMA = cv.Schema({cv.GenerateID(): cv.use_id(MicroWakeWord)})
@@ -496,3 +537,30 @@ async def micro_wake_word_action_to_code(config, action_id, template_arg, args):
     var = cg.new_Pvariable(action_id, template_arg)
     await cg.register_parented(var, config[CONF_ID])
     return var
+
+
+MICRO_WAKE_WORLD_MODEL_ACTION_SCHEMA = automation.maybe_simple_id(
+    {
+        cv.Required(CONF_ID): cv.use_id(WakeWordModel),
+    }
+)
+
+
+@register_action(
+    "micro_wake_word.enable_model",
+    EnableModelAction,
+    MICRO_WAKE_WORLD_MODEL_ACTION_SCHEMA,
+)
+@register_action(
+    "micro_wake_word.disable_model",
+    DisableModelAction,
+    MICRO_WAKE_WORLD_MODEL_ACTION_SCHEMA,
+)
+@register_condition(
+    "micro_wake_word.model_is_enabled",
+    ModelIsEnabledCondition,
+    MICRO_WAKE_WORLD_MODEL_ACTION_SCHEMA,
+)
+async def model_action(config, action_id, template_arg, args):
+    parent = await cg.get_variable(config[CONF_ID])
+    return cg.new_Pvariable(action_id, template_arg, parent)

esphome/components/micro_wake_word/automation.h

@@ -0,0 +1,54 @@
+#pragma once
+
+#include "micro_wake_word.h"
+#include "streaming_model.h"
+
+#ifdef USE_ESP_IDF
+namespace esphome {
+namespace micro_wake_word {
+
+template<typename... Ts> class StartAction : public Action<Ts...>, public Parented<MicroWakeWord> {
+ public:
+  void play(Ts... x) override { this->parent_->start(); }
+};
+
+template<typename... Ts> class StopAction : public Action<Ts...>, public Parented<MicroWakeWord> {
+ public:
+  void play(Ts... x) override { this->parent_->stop(); }
+};
+
+template<typename... Ts> class IsRunningCondition : public Condition<Ts...>, public Parented<MicroWakeWord> {
+ public:
+  bool check(Ts... x) override { return this->parent_->is_running(); }
+};
+
+template<typename... Ts> class EnableModelAction : public Action<Ts...> {
+ public:
+  explicit EnableModelAction(WakeWordModel *wake_word_model) : wake_word_model_(wake_word_model) {}
+  void play(Ts... x) override { this->wake_word_model_->enable(); }
+
+ protected:
+  WakeWordModel *wake_word_model_;
+};
+
+template<typename... Ts> class DisableModelAction : public Action<Ts...> {
+ public:
+  explicit DisableModelAction(WakeWordModel *wake_word_model) : wake_word_model_(wake_word_model) {}
+  void play(Ts... x) override { this->wake_word_model_->disable(); }
+
+ protected:
+  WakeWordModel *wake_word_model_;
+};
+
+template<typename... Ts> class ModelIsEnabledCondition : public Condition<Ts...> {
+ public:
+  explicit ModelIsEnabledCondition(WakeWordModel *wake_word_model) : wake_word_model_(wake_word_model) {}
+  bool check(Ts... x) override { return this->wake_word_model_->is_enabled(); }
+
+ protected:
+  WakeWordModel *wake_word_model_;
+};
+
+}  // namespace micro_wake_word
+}  // namespace esphome
+#endif

esphome/components/micro_wake_word/micro_wake_word.cpp

@@ -1,5 +1,4 @@
 #include "micro_wake_word.h"
-#include "streaming_model.h"
 
 #ifdef USE_ESP_IDF
 
@@ -7,41 +6,57 @@
 #include "esphome/core/helpers.h"
 #include "esphome/core/log.h"
 
-#include <frontend.h>
-#include <frontend_util.h>
+#include "esphome/components/audio/audio_transfer_buffer.h"
 
-#include <tensorflow/lite/core/c/common.h>
-#include <tensorflow/lite/micro/micro_interpreter.h>
-#include <tensorflow/lite/micro/micro_mutable_op_resolver.h>
-
-#include <cmath>
+#ifdef USE_OTA
+#include "esphome/components/ota/ota_backend.h"
+#endif
 
 namespace esphome {
 namespace micro_wake_word {
 
 static const char *const TAG = "micro_wake_word";
 
-static const size_t SAMPLE_RATE_HZ = 16000;  // 16 kHz
-static const size_t BUFFER_LENGTH = 64;      // 0.064 seconds
-static const size_t BUFFER_SIZE = SAMPLE_RATE_HZ / 1000 * BUFFER_LENGTH;
-static const size_t INPUT_BUFFER_SIZE = 16 * SAMPLE_RATE_HZ / 1000;  // 16ms * 16kHz / 1000ms
+static const ssize_t DETECTION_QUEUE_LENGTH = 5;
+
+static const size_t DATA_TIMEOUT_MS = 50;
+
+static const uint32_t RING_BUFFER_DURATION_MS = 120;
+static const uint32_t RING_BUFFER_SAMPLES = RING_BUFFER_DURATION_MS * (AUDIO_SAMPLE_FREQUENCY / 1000);
+static const size_t RING_BUFFER_SIZE = RING_BUFFER_SAMPLES * sizeof(int16_t);
+
+static const uint32_t INFERENCE_TASK_STACK_SIZE = 3072;
+static const UBaseType_t INFERENCE_TASK_PRIORITY = 3;
+
+enum EventGroupBits : uint32_t {
+  COMMAND_STOP = (1 << 0),  // Signals the inference task should stop
+
+  TASK_STARTING = (1 << 3),
+  TASK_RUNNING = (1 << 4),
+  TASK_STOPPING = (1 << 5),
+  TASK_STOPPED = (1 << 6),
+
+  ERROR_MEMORY = (1 << 9),
+  ERROR_INFERENCE = (1 << 10),
+
+  WARNING_FULL_RING_BUFFER = (1 << 13),
+
+  ERROR_BITS = ERROR_MEMORY | ERROR_INFERENCE,
+  ALL_BITS = 0xfffff,  // 24 total bits available in an event group
+};
 
 float MicroWakeWord::get_setup_priority() const { return setup_priority::AFTER_CONNECTION; }
 
 static const LogString *micro_wake_word_state_to_string(State state) {
   switch (state) {
-    case State::IDLE:
-      return LOG_STR("IDLE");
-    case State::START_MICROPHONE:
-      return LOG_STR("START_MICROPHONE");
-    case State::STARTING_MICROPHONE:
-      return LOG_STR("STARTING_MICROPHONE");
+    case State::STARTING:
+      return LOG_STR("STARTING");
     case State::DETECTING_WAKE_WORD:
       return LOG_STR("DETECTING_WAKE_WORD");
-    case State::STOP_MICROPHONE:
-      return LOG_STR("STOP_MICROPHONE");
-    case State::STOPPING_MICROPHONE:
-      return LOG_STR("STOPPING_MICROPHONE");
+    case State::STOPPING:
+      return LOG_STR("STOPPING");
+    case State::STOPPED:
+      return LOG_STR("STOPPED");
     default:
       return LOG_STR("UNKNOWN");
   }
@@ -51,7 +66,7 @@ void MicroWakeWord::dump_config() {
   ESP_LOGCONFIG(TAG, "microWakeWord:");
   ESP_LOGCONFIG(TAG, "  models:");
   for (auto &model : this->wake_word_models_) {
-    model.log_model_config();
+    model->log_model_config();
   }
 #ifdef USE_MICRO_WAKE_WORD_VAD
   this->vad_model_->log_model_config();
@@ -61,108 +76,266 @@ void MicroWakeWord::dump_config() {
 void MicroWakeWord::setup() {
   ESP_LOGCONFIG(TAG, "Setting up microWakeWord...");
 
-  this->microphone_->add_data_callback([this](const std::vector<int16_t> &data) {
-    if (this->state_ != State::DETECTING_WAKE_WORD) {
-      return;
-    }
-    std::shared_ptr<RingBuffer> temp_ring_buffer = this->ring_buffer_;
-    if (this->ring_buffer_.use_count() == 2) {
-      // mWW still owns the ring buffer and temp_ring_buffer does as well, proceed to copy audio into ring buffer
+  this->frontend_config_.window.size_ms = FEATURE_DURATION_MS;
+  this->frontend_config_.window.step_size_ms = this->features_step_size_;
+  this->frontend_config_.filterbank.num_channels = PREPROCESSOR_FEATURE_SIZE;
+  this->frontend_config_.filterbank.lower_band_limit = FILTERBANK_LOWER_BAND_LIMIT;
+  this->frontend_config_.filterbank.upper_band_limit = FILTERBANK_UPPER_BAND_LIMIT;
+  this->frontend_config_.noise_reduction.smoothing_bits = NOISE_REDUCTION_SMOOTHING_BITS;
+  this->frontend_config_.noise_reduction.even_smoothing = NOISE_REDUCTION_EVEN_SMOOTHING;
+  this->frontend_config_.noise_reduction.odd_smoothing = NOISE_REDUCTION_ODD_SMOOTHING;
+  this->frontend_config_.noise_reduction.min_signal_remaining = NOISE_REDUCTION_MIN_SIGNAL_REMAINING;
+  this->frontend_config_.pcan_gain_control.enable_pcan = PCAN_GAIN_CONTROL_ENABLE_PCAN;
+  this->frontend_config_.pcan_gain_control.strength = PCAN_GAIN_CONTROL_STRENGTH;
+  this->frontend_config_.pcan_gain_control.offset = PCAN_GAIN_CONTROL_OFFSET;
+  this->frontend_config_.pcan_gain_control.gain_bits = PCAN_GAIN_CONTROL_GAIN_BITS;
+  this->frontend_config_.log_scale.enable_log = LOG_SCALE_ENABLE_LOG;
+  this->frontend_config_.log_scale.scale_shift = LOG_SCALE_SCALE_SHIFT;
 
-      size_t bytes_free = temp_ring_buffer->free();
-
-      if (bytes_free < data.size() * sizeof(int16_t)) {
-        ESP_LOGW(
-            TAG,
-            "Not enough free bytes in ring buffer to store incoming audio data (free bytes=%d, incoming bytes=%d). "
-            "Resetting the ring buffer. Wake word detection accuracy will be reduced.",
-            bytes_free, data.size());
-
-        temp_ring_buffer->reset();
-      }
-      temp_ring_buffer->write((void *) data.data(), data.size() * sizeof(int16_t));
-    }
-  });
-
-  if (!this->register_streaming_ops_(this->streaming_op_resolver_)) {
+  this->event_group_ = xEventGroupCreate();
+  if (this->event_group_ == nullptr) {
+    ESP_LOGE(TAG, "Failed to create event group");
     this->mark_failed();
     return;
   }
 
+  this->detection_queue_ = xQueueCreate(DETECTION_QUEUE_LENGTH, sizeof(DetectionEvent));
+  if (this->detection_queue_ == nullptr) {
+    ESP_LOGE(TAG, "Failed to create detection event queue");
+    this->mark_failed();
+    return;
+  }
+
+  this->microphone_source_->add_data_callback([this](const std::vector<uint8_t> &data) {
+    if (this->state_ == State::STOPPED) {
+      return;
+    }
+    std::shared_ptr<RingBuffer> temp_ring_buffer = this->ring_buffer_.lock();
+    if (this->ring_buffer_.use_count() > 1) {
+      size_t bytes_free = temp_ring_buffer->free();
+
+      if (bytes_free < data.size()) {
+        xEventGroupSetBits(this->event_group_, EventGroupBits::WARNING_FULL_RING_BUFFER);
+        temp_ring_buffer->reset();
+      }
+      temp_ring_buffer->write((void *) data.data(), data.size());
+    }
+  });
+
+#ifdef USE_OTA
+  ota::get_global_ota_callback()->add_on_state_callback(
+      [this](ota::OTAState state, float progress, uint8_t error, ota::OTAComponent *comp) {
+        if (state == ota::OTA_STARTED) {
+          this->suspend_task_();
+        } else if (state == ota::OTA_ERROR) {
+          this->resume_task_();
+        }
+      });
+#endif
   ESP_LOGCONFIG(TAG, "Micro Wake Word initialized");
-
-  this->frontend_config_.window.size_ms = FEATURE_DURATION_MS;
-  this->frontend_config_.window.step_size_ms = this->features_step_size_;
-  this->frontend_config_.filterbank.num_channels = PREPROCESSOR_FEATURE_SIZE;
-  this->frontend_config_.filterbank.lower_band_limit = 125.0;
-  this->frontend_config_.filterbank.upper_band_limit = 7500.0;
-  this->frontend_config_.noise_reduction.smoothing_bits = 10;
-  this->frontend_config_.noise_reduction.even_smoothing = 0.025;
-  this->frontend_config_.noise_reduction.odd_smoothing = 0.06;
-  this->frontend_config_.noise_reduction.min_signal_remaining = 0.05;
-  this->frontend_config_.pcan_gain_control.enable_pcan = 1;
-  this->frontend_config_.pcan_gain_control.strength = 0.95;
-  this->frontend_config_.pcan_gain_control.offset = 80.0;
-  this->frontend_config_.pcan_gain_control.gain_bits = 21;
-  this->frontend_config_.log_scale.enable_log = 1;
-  this->frontend_config_.log_scale.scale_shift = 6;
 }
 
-void MicroWakeWord::add_wake_word_model(const uint8_t *model_start, float probability_cutoff,
-                                        size_t sliding_window_average_size, const std::string &wake_word,
-                                        size_t tensor_arena_size) {
-  this->wake_word_models_.emplace_back(model_start, probability_cutoff, sliding_window_average_size, wake_word,
-                                       tensor_arena_size);
+void MicroWakeWord::inference_task(void *params) {
+  MicroWakeWord *this_mww = (MicroWakeWord *) params;
+
+  xEventGroupSetBits(this_mww->event_group_, EventGroupBits::TASK_STARTING);
+
+  {  // Ensures any C++ objects fall out of scope to deallocate before deleting the task
+    const size_t new_samples_to_read = this_mww->features_step_size_ * (AUDIO_SAMPLE_FREQUENCY / 1000);
+    std::unique_ptr<audio::AudioSourceTransferBuffer> audio_buffer;
+    int8_t features_buffer[PREPROCESSOR_FEATURE_SIZE];
+
+    if (!(xEventGroupGetBits(this_mww->event_group_) & ERROR_BITS)) {
+      // Allocate audio transfer buffer
+      audio_buffer = audio::AudioSourceTransferBuffer::create(new_samples_to_read * sizeof(int16_t));
+
+      if (audio_buffer == nullptr) {
+        xEventGroupSetBits(this_mww->event_group_, EventGroupBits::ERROR_MEMORY);
+      }
+    }
+
+    if (!(xEventGroupGetBits(this_mww->event_group_) & ERROR_BITS)) {
+      // Allocate ring buffer
+      std::shared_ptr<RingBuffer> temp_ring_buffer = RingBuffer::create(RING_BUFFER_SIZE);
+      if (temp_ring_buffer.use_count() == 0) {
+        xEventGroupSetBits(this_mww->event_group_, EventGroupBits::ERROR_MEMORY);
+      }
+      audio_buffer->set_source(temp_ring_buffer);
+      this_mww->ring_buffer_ = temp_ring_buffer;
+    }
+
+    if (!(xEventGroupGetBits(this_mww->event_group_) & ERROR_BITS)) {
+      this_mww->microphone_source_->start();
+      xEventGroupSetBits(this_mww->event_group_, EventGroupBits::TASK_RUNNING);
+
+      while (!(xEventGroupGetBits(this_mww->event_group_) & COMMAND_STOP)) {
+        audio_buffer->transfer_data_from_source(pdMS_TO_TICKS(DATA_TIMEOUT_MS));
+
+        if (audio_buffer->available() < new_samples_to_read * sizeof(int16_t)) {
+          // Insufficient data to generate new spectrogram features, read more next iteration
+          continue;
+        }
+
+        // Generate new spectrogram features
+        size_t processed_samples = this_mww->generate_features_(
+            (int16_t *) audio_buffer->get_buffer_start(), audio_buffer->available() / sizeof(int16_t), features_buffer);
+        audio_buffer->decrease_buffer_length(processed_samples * sizeof(int16_t));
+
+        // Run inference using the new spectorgram features
+        if (!this_mww->update_model_probabilities_(features_buffer)) {
+          xEventGroupSetBits(this_mww->event_group_, EventGroupBits::ERROR_INFERENCE);
+          break;
+        }
+
+        // Process each model's probabilities and possibly send a Detection Event to the queue
+        this_mww->process_probabilities_();
+      }
+    }
+  }
+
+  xEventGroupSetBits(this_mww->event_group_, EventGroupBits::TASK_STOPPING);
+
+  this_mww->unload_models_();
+  this_mww->microphone_source_->stop();
+  FrontendFreeStateContents(&this_mww->frontend_state_);
+
+  xEventGroupSetBits(this_mww->event_group_, EventGroupBits::TASK_STOPPED);
+  while (true) {
+    // Continuously delay until the main loop deletes the task
+    delay(10);
+  }
 }
 
+std::vector<WakeWordModel *> MicroWakeWord::get_wake_words() {
+  std::vector<WakeWordModel *> external_wake_word_models;
+  for (auto *model : this->wake_word_models_) {
+    if (!model->get_internal_only()) {
+      external_wake_word_models.push_back(model);
+    }
+  }
+  return external_wake_word_models;
+}
+
+void MicroWakeWord::add_wake_word_model(WakeWordModel *model) { this->wake_word_models_.push_back(model); }
+
 #ifdef USE_MICRO_WAKE_WORD_VAD
-void MicroWakeWord::add_vad_model(const uint8_t *model_start, float probability_cutoff, size_t sliding_window_size,
+void MicroWakeWord::add_vad_model(const uint8_t *model_start, uint8_t probability_cutoff, size_t sliding_window_size,
                                   size_t tensor_arena_size) {
   this->vad_model_ = make_unique<VADModel>(model_start, probability_cutoff, sliding_window_size, tensor_arena_size);
 }
 #endif
 
+void MicroWakeWord::suspend_task_() {
+  if (this->inference_task_handle_ != nullptr) {
+    vTaskSuspend(this->inference_task_handle_);
+  }
+}
+
+void MicroWakeWord::resume_task_() {
+  if (this->inference_task_handle_ != nullptr) {
+    vTaskResume(this->inference_task_handle_);
+  }
+}
+
 void MicroWakeWord::loop() {
+  uint32_t event_group_bits = xEventGroupGetBits(this->event_group_);
+
+  if (event_group_bits & EventGroupBits::ERROR_MEMORY) {
+    xEventGroupClearBits(this->event_group_, EventGroupBits::ERROR_MEMORY);
+    ESP_LOGE(TAG, "Encountered an error allocating buffers");
+  }
+
+  if (event_group_bits & EventGroupBits::ERROR_INFERENCE) {
+    xEventGroupClearBits(this->event_group_, EventGroupBits::ERROR_INFERENCE);
+    ESP_LOGE(TAG, "Encountered an error while performing an inference");
+  }
+
+  if (event_group_bits & EventGroupBits::WARNING_FULL_RING_BUFFER) {
+    xEventGroupClearBits(this->event_group_, EventGroupBits::WARNING_FULL_RING_BUFFER);
+    ESP_LOGW(TAG, "Not enough free bytes in ring buffer to store incoming audio data. Resetting the ring buffer. Wake "
+                  "word detection accuracy will temporarily be reduced.");
+  }
+
+  if (event_group_bits & EventGroupBits::TASK_STARTING) {
+    ESP_LOGD(TAG, "Inference task has started, attempting to allocate memory for buffers");
+    xEventGroupClearBits(this->event_group_, EventGroupBits::TASK_STARTING);
+  }
+
+  if (event_group_bits & EventGroupBits::TASK_RUNNING) {
+    ESP_LOGD(TAG, "Inference task is running");
+
+    xEventGroupClearBits(this->event_group_, EventGroupBits::TASK_RUNNING);
+    this->set_state_(State::DETECTING_WAKE_WORD);
+  }
+
+  if (event_group_bits & EventGroupBits::TASK_STOPPING) {
+    ESP_LOGD(TAG, "Inference task is stopping, deallocating buffers");
+    xEventGroupClearBits(this->event_group_, EventGroupBits::TASK_STOPPING);
+  }
+
+  if ((event_group_bits & EventGroupBits::TASK_STOPPED)) {
+    ESP_LOGD(TAG, "Inference task is finished, freeing task resources");
+    vTaskDelete(this->inference_task_handle_);
+    this->inference_task_handle_ = nullptr;
+    xEventGroupClearBits(this->event_group_, ALL_BITS);
+    xQueueReset(this->detection_queue_);
+    this->set_state_(State::STOPPED);
+  }
+
+  if ((this->pending_start_) && (this->state_ == State::STOPPED)) {
+    this->set_state_(State::STARTING);
+    this->pending_start_ = false;
+  }
+
+  if ((this->pending_stop_) && (this->state_ == State::DETECTING_WAKE_WORD)) {
+    this->set_state_(State::STOPPING);
+    this->pending_stop_ = false;
+  }
+
   switch (this->state_) {
-    case State::IDLE:
-      break;
-    case State::START_MICROPHONE:
-      ESP_LOGD(TAG, "Starting Microphone");
-      this->microphone_->start();
-      this->set_state_(State::STARTING_MICROPHONE);
-      break;
-    case State::STARTING_MICROPHONE:
-      if (this->microphone_->is_running()) {
-        this->set_state_(State::DETECTING_WAKE_WORD);
-      }
-      break;
-    case State::DETECTING_WAKE_WORD:
-      while (this->has_enough_samples_()) {
-        this->update_model_probabilities_();
-        if (this->detect_wake_words_()) {
-          ESP_LOGD(TAG, "Wake Word '%s' Detected", (this->detected_wake_word_).c_str());
-          this->detected_ = true;
-          this->set_state_(State::STOP_MICROPHONE);
+    case State::STARTING:
+      if ((this->inference_task_handle_ == nullptr) && !this->status_has_error()) {
+        // Setup preprocesor feature generator. If done in the task, it would lock the task to its initial core, as it
+        // uses floating point operations.
+        if (!FrontendPopulateState(&this->frontend_config_, &this->frontend_state_, AUDIO_SAMPLE_FREQUENCY)) {
+          this->status_momentary_error(
+              "Failed to allocate buffers for spectrogram feature processor, attempting again in 1 second", 1000);
+          return;
+        }
+
+        xTaskCreate(MicroWakeWord::inference_task, "mww", INFERENCE_TASK_STACK_SIZE, (void *) this,
+                    INFERENCE_TASK_PRIORITY, &this->inference_task_handle_);
+
+        if (this->inference_task_handle_ == nullptr) {
+          FrontendFreeStateContents(&this->frontend_state_);  // Deallocate frontend state
+          this->status_momentary_error("Task failed to start, attempting again in 1 second", 1000);
         }
       }
       break;
-    case State::STOP_MICROPHONE:
-      ESP_LOGD(TAG, "Stopping Microphone");
-      this->microphone_->stop();
-      this->set_state_(State::STOPPING_MICROPHONE);
-      this->unload_models_();
-      this->deallocate_buffers_();
-      break;
-    case State::STOPPING_MICROPHONE:
-      if (this->microphone_->is_stopped()) {
-        this->set_state_(State::IDLE);
-        if (this->detected_) {
-          this->wake_word_detected_trigger_->trigger(this->detected_wake_word_);
-          this->detected_ = false;
-          this->detected_wake_word_ = "";
+    case State::DETECTING_WAKE_WORD: {
+      DetectionEvent detection_event;
+      while (xQueueReceive(this->detection_queue_, &detection_event, 0)) {
+        if (detection_event.blocked_by_vad) {
+          ESP_LOGD(TAG, "Wake word model predicts '%s', but VAD model doesn't.", detection_event.wake_word->c_str());
+        } else {
+          constexpr float uint8_to_float_divisor =
+              255.0f;  // Converting a quantized uint8 probability to floating point
+          ESP_LOGD(TAG, "Detected '%s' with sliding average probability is %.2f and max probability is %.2f",
+                   detection_event.wake_word->c_str(), (detection_event.average_probability / uint8_to_float_divisor),
+                   (detection_event.max_probability / uint8_to_float_divisor));
+          this->wake_word_detected_trigger_->trigger(*detection_event.wake_word);
+          if (this->stop_after_detection_) {
+            this->stop();
+          }
         }
       }
       break;
+    }
+    case State::STOPPING:
+      xEventGroupSetBits(this->event_group_, EventGroupBits::COMMAND_STOP);
+      break;
+    case State::STOPPED:
+      break;
   }
 }
 
@@ -177,199 +350,40 @@ void MicroWakeWord::start() {
     return;
   }
 
-  if (this->state_ != State::IDLE) {
-    ESP_LOGW(TAG, "Wake word is already running");
+  if (this->is_running()) {
+    ESP_LOGW(TAG, "Wake word detection is already running");
     return;
   }
 
-  if (!this->load_models_() || !this->allocate_buffers_()) {
-    ESP_LOGE(TAG, "Failed to load the wake word model(s) or allocate buffers");
-    this->status_set_error();
-  } else {
-    this->status_clear_error();
-  }
+  ESP_LOGD(TAG, "Starting wake word detection");
 
-  if (this->status_has_error()) {
-    ESP_LOGW(TAG, "Wake word component has an error. Please check logs");
-    return;
-  }
-
-  this->reset_states_();
-  this->set_state_(State::START_MICROPHONE);
+  this->pending_start_ = true;
+  this->pending_stop_ = false;
 }
 
 void MicroWakeWord::stop() {
-  if (this->state_ == State::IDLE) {
-    ESP_LOGW(TAG, "Wake word is already stopped");
+  if (this->state_ == STOPPED)
     return;
-  }
-  if (this->state_ == State::STOPPING_MICROPHONE) {
-    ESP_LOGW(TAG, "Wake word is already stopping");
-    return;
-  }
-  this->set_state_(State::STOP_MICROPHONE);
+
+  ESP_LOGD(TAG, "Stopping wake word detection");
+
+  this->pending_start_ = false;
+  this->pending_stop_ = true;
 }
 
 void MicroWakeWord::set_state_(State state) {
-  ESP_LOGD(TAG, "State changed from %s to %s", LOG_STR_ARG(micro_wake_word_state_to_string(this->state_)),
-           LOG_STR_ARG(micro_wake_word_state_to_string(state)));
-  this->state_ = state;
+  if (this->state_ != state) {
+    ESP_LOGD(TAG, "State changed from %s to %s", LOG_STR_ARG(micro_wake_word_state_to_string(this->state_)),
+             LOG_STR_ARG(micro_wake_word_state_to_string(state)));
+    this->state_ = state;
+  }
 }
 
-bool MicroWakeWord::allocate_buffers_() {
-  ExternalRAMAllocator<int16_t> audio_samples_allocator(ExternalRAMAllocator<int16_t>::ALLOW_FAILURE);
-
-  if (this->input_buffer_ == nullptr) {
-    this->input_buffer_ = audio_samples_allocator.allocate(INPUT_BUFFER_SIZE * sizeof(int16_t));
-    if (this->input_buffer_ == nullptr) {
-      ESP_LOGE(TAG, "Could not allocate input buffer");
-      return false;
-    }
-  }
-
-  if (this->preprocessor_audio_buffer_ == nullptr) {
-    this->preprocessor_audio_buffer_ = audio_samples_allocator.allocate(this->new_samples_to_get_());
-    if (this->preprocessor_audio_buffer_ == nullptr) {
-      ESP_LOGE(TAG, "Could not allocate the audio preprocessor's buffer.");
-      return false;
-    }
-  }
-
-  if (this->ring_buffer_.use_count() == 0) {
-    this->ring_buffer_ = RingBuffer::create(BUFFER_SIZE * sizeof(int16_t));
-    if (this->ring_buffer_.use_count() == 0) {
-      ESP_LOGE(TAG, "Could not allocate ring buffer");
-      return false;
-    }
-  }
-
-  return true;
-}
-
-void MicroWakeWord::deallocate_buffers_() {
-  ExternalRAMAllocator<int16_t> audio_samples_allocator(ExternalRAMAllocator<int16_t>::ALLOW_FAILURE);
-  if (this->input_buffer_ != nullptr) {
-    audio_samples_allocator.deallocate(this->input_buffer_, INPUT_BUFFER_SIZE * sizeof(int16_t));
-    this->input_buffer_ = nullptr;
-  }
-
-  if (this->preprocessor_audio_buffer_ != nullptr) {
-    audio_samples_allocator.deallocate(this->preprocessor_audio_buffer_, this->new_samples_to_get_());
-    this->preprocessor_audio_buffer_ = nullptr;
-  }
-
-  this->ring_buffer_.reset();
-}
-
-bool MicroWakeWord::load_models_() {
-  // Setup preprocesor feature generator
-  if (!FrontendPopulateState(&this->frontend_config_, &this->frontend_state_, AUDIO_SAMPLE_FREQUENCY)) {
-    ESP_LOGD(TAG, "Failed to populate frontend state");
-    FrontendFreeStateContents(&this->frontend_state_);
-    return false;
-  }
-
-  // Setup streaming models
-  for (auto &model : this->wake_word_models_) {
-    if (!model.load_model(this->streaming_op_resolver_)) {
-      ESP_LOGE(TAG, "Failed to initialize a wake word model.");
-      return false;
-    }
-  }
-#ifdef USE_MICRO_WAKE_WORD_VAD
-  if (!this->vad_model_->load_model(this->streaming_op_resolver_)) {
-    ESP_LOGE(TAG, "Failed to initialize VAD model.");
-    return false;
-  }
-#endif
-
-  return true;
-}
-
-void MicroWakeWord::unload_models_() {
-  FrontendFreeStateContents(&this->frontend_state_);
-
-  for (auto &model : this->wake_word_models_) {
-    model.unload_model();
-  }
-#ifdef USE_MICRO_WAKE_WORD_VAD
-  this->vad_model_->unload_model();
-#endif
-}
-
-void MicroWakeWord::update_model_probabilities_() {
-  int8_t audio_features[PREPROCESSOR_FEATURE_SIZE];
-
-  if (!this->generate_features_for_window_(audio_features)) {
-    return;
-  }
-
-  // Increase the counter since the last positive detection
-  this->ignore_windows_ = std::min(this->ignore_windows_ + 1, 0);
-
-  for (auto &model : this->wake_word_models_) {
-    // Perform inference
-    model.perform_streaming_inference(audio_features);
-  }
-#ifdef USE_MICRO_WAKE_WORD_VAD
-  this->vad_model_->perform_streaming_inference(audio_features);
-#endif
-}
-
-bool MicroWakeWord::detect_wake_words_() {
-  // Verify we have processed samples since the last positive detection
-  if (this->ignore_windows_ < 0) {
-    return false;
-  }
-
-#ifdef USE_MICRO_WAKE_WORD_VAD
-  bool vad_state = this->vad_model_->determine_detected();
-#endif
-
-  for (auto &model : this->wake_word_models_) {
-    if (model.determine_detected()) {
-#ifdef USE_MICRO_WAKE_WORD_VAD
-      if (vad_state) {
-#endif
-        this->detected_wake_word_ = model.get_wake_word();
-        return true;
-#ifdef USE_MICRO_WAKE_WORD_VAD
-      } else {
-        ESP_LOGD(TAG, "Wake word model predicts %s, but VAD model doesn't.", model.get_wake_word().c_str());
-      }
-#endif
-    }
-  }
-
-  return false;
-}
-
-bool MicroWakeWord::has_enough_samples_() {
-  return this->ring_buffer_->available() >=
-         (this->features_step_size_ * (AUDIO_SAMPLE_FREQUENCY / 1000)) * sizeof(int16_t);
-}
-
-bool MicroWakeWord::generate_features_for_window_(int8_t features[PREPROCESSOR_FEATURE_SIZE]) {
-  // Ensure we have enough new audio samples in the ring buffer for a full window
-  if (!this->has_enough_samples_()) {
-    return false;
-  }
-
-  size_t bytes_read = this->ring_buffer_->read((void *) (this->preprocessor_audio_buffer_),
-                                               this->new_samples_to_get_() * sizeof(int16_t), pdMS_TO_TICKS(200));
-
-  if (bytes_read == 0) {
-    ESP_LOGE(TAG, "Could not read data from Ring Buffer");
-  } else if (bytes_read < this->new_samples_to_get_() * sizeof(int16_t)) {
-    ESP_LOGD(TAG, "Partial Read of Data by Model");
-    ESP_LOGD(TAG, "Could only read %d bytes when required %d bytes ", bytes_read,
-             (int) (this->new_samples_to_get_() * sizeof(int16_t)));
-    return false;
-  }
-
-  size_t num_samples_read;
-  struct FrontendOutput frontend_output = FrontendProcessSamples(
-      &this->frontend_state_, this->preprocessor_audio_buffer_, this->new_samples_to_get_(), &num_samples_read);
+size_t MicroWakeWord::generate_features_(int16_t *audio_buffer, size_t samples_available,
+                                         int8_t features_buffer[PREPROCESSOR_FEATURE_SIZE]) {
+  size_t processed_samples = 0;
+  struct FrontendOutput frontend_output =
+      FrontendProcessSamples(&this->frontend_state_, audio_buffer, samples_available, &processed_samples);
 
   for (size_t i = 0; i < frontend_output.size; ++i) {
     // These scaling values are set to match the TFLite audio frontend int8 output.
@@ -379,8 +393,8 @@ bool MicroWakeWord::generate_features_for_window_(int8_t features[PREPROCESSOR_F
     // for historical reasons, to match up with the output of other feature
     // generators.
     // The process is then further complicated when we quantize the model. This
-    // means we have to scale the 0.0 to 26.0 real values to the -128 to 127
-    // signed integer numbers.
+    // means we have to scale the 0.0 to 26.0 real values to the -128 (INT8_MIN)
+    // to 127 (INT8_MAX) signed integer numbers.
     // All this means that to get matching values from our integer feature
     // output into the tensor input, we have to perform:
     // input = (((feature / 25.6) / 26.0) * 256) - 128
@@ -389,74 +403,63 @@ bool MicroWakeWord::generate_features_for_window_(int8_t features[PREPROCESSOR_F
     constexpr int32_t value_scale = 256;
     constexpr int32_t value_div = 666;  // 666 = 25.6 * 26.0 after rounding
     int32_t value = ((frontend_output.values[i] * value_scale) + (value_div / 2)) / value_div;
-    value -= 128;
-    if (value < -128) {
-      value = -128;
-    }
-    if (value > 127) {
-      value = 127;
-    }
-    features[i] = value;
+
+    value += INT8_MIN;  // Adds a -128; i.e., subtracts 128
+    features_buffer[i] = static_cast<int8_t>(clamp<int32_t>(value, INT8_MIN, INT8_MAX));
   }
 
-  return true;
+  return processed_samples;
 }
 
-void MicroWakeWord::reset_states_() {
-  ESP_LOGD(TAG, "Resetting buffers and probabilities");
-  this->ring_buffer_->reset();
-  this->ignore_windows_ = -MIN_SLICES_BEFORE_DETECTION;
+void MicroWakeWord::process_probabilities_() {
+#ifdef USE_MICRO_WAKE_WORD_VAD
+  DetectionEvent vad_state = this->vad_model_->determine_detected();
+
+  this->vad_state_ = vad_state.detected;  // atomic write, so thread safe
+#endif
+
   for (auto &model : this->wake_word_models_) {
-    model.reset_probabilities();
+    if (model->get_unprocessed_probability_status()) {
+      // Only detect wake words if there is a new probability since the last check
+      DetectionEvent wake_word_state = model->determine_detected();
+      if (wake_word_state.detected) {
+#ifdef USE_MICRO_WAKE_WORD_VAD
+        if (vad_state.detected) {
+#endif
+          xQueueSend(this->detection_queue_, &wake_word_state, portMAX_DELAY);
+          model->reset_probabilities();
+#ifdef USE_MICRO_WAKE_WORD_VAD
+        } else {
+          wake_word_state.blocked_by_vad = true;
+          xQueueSend(this->detection_queue_, &wake_word_state, portMAX_DELAY);
+        }
+#endif
+      }
+    }
+  }
+}
+
+void MicroWakeWord::unload_models_() {
+  for (auto &model : this->wake_word_models_) {
+    model->unload_model();
   }
 #ifdef USE_MICRO_WAKE_WORD_VAD
-  this->vad_model_->reset_probabilities();
+  this->vad_model_->unload_model();
 #endif
 }
 
-bool MicroWakeWord::register_streaming_ops_(tflite::MicroMutableOpResolver<20> &op_resolver) {
-  if (op_resolver.AddCallOnce() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddVarHandle() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddReshape() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddReadVariable() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddStridedSlice() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddConcatenation() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddAssignVariable() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddConv2D() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddMul() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddAdd() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddMean() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddFullyConnected() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddLogistic() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddQuantize() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddDepthwiseConv2D() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddAveragePool2D() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddMaxPool2D() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddPad() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddPack() != kTfLiteOk)
-    return false;
-  if (op_resolver.AddSplitV() != kTfLiteOk)
-    return false;
+bool MicroWakeWord::update_model_probabilities_(const int8_t audio_features[PREPROCESSOR_FEATURE_SIZE]) {
+  bool success = true;
 
-  return true;
+  for (auto &model : this->wake_word_models_) {
+    // Perform inference
+    success = success & model->perform_streaming_inference(audio_features);
+  }
+#ifdef USE_MICRO_WAKE_WORD_VAD
+  success = success & this->vad_model_->perform_streaming_inference(audio_features);
+#endif
+
+  return success;
 }
 
 }  // namespace micro_wake_word

esphome/components/micro_wake_word/micro_wake_word.h

@@ -5,33 +5,27 @@
 #include "preprocessor_settings.h"
 #include "streaming_model.h"
 
+#include "esphome/components/microphone/microphone_source.h"
+
 #include "esphome/core/automation.h"
 #include "esphome/core/component.h"
 #include "esphome/core/ring_buffer.h"
 
-#include "esphome/components/microphone/microphone.h"
+#include <freertos/event_groups.h>
 
+#include <frontend.h>
 #include <frontend_util.h>
 
-#include <tensorflow/lite/core/c/common.h>
-#include <tensorflow/lite/micro/micro_interpreter.h>
-#include <tensorflow/lite/micro/micro_mutable_op_resolver.h>
-
 namespace esphome {
 namespace micro_wake_word {
 
 enum State {
-  IDLE,
-  START_MICROPHONE,
-  STARTING_MICROPHONE,
+  STARTING,
   DETECTING_WAKE_WORD,
-  STOP_MICROPHONE,
-  STOPPING_MICROPHONE,
+  STOPPING,
+  STOPPED,
 };
 
-// The number of audio slices to process before accepting a positive detection
-static const uint8_t MIN_SLICES_BEFORE_DETECTION = 74;
-
 class MicroWakeWord : public Component {
  public:
   void setup() override;
@@ -42,124 +36,95 @@ class MicroWakeWord : public Component {
   void start();
   void stop();
 
-  bool is_running() const { return this->state_ != State::IDLE; }
+  bool is_running() const { return this->state_ != State::STOPPED; }
 
   void set_features_step_size(uint8_t step_size) { this->features_step_size_ = step_size; }
 
-  void set_microphone(microphone::Microphone *microphone) { this->microphone_ = microphone; }
+  void set_microphone_source(microphone::MicrophoneSource *microphone_source) {
+    this->microphone_source_ = microphone_source;
+  }
+
+  void set_stop_after_detection(bool stop_after_detection) { this->stop_after_detection_ = stop_after_detection; }
 
   Trigger<std::string> *get_wake_word_detected_trigger() const { return this->wake_word_detected_trigger_; }
 
-  void add_wake_word_model(const uint8_t *model_start, float probability_cutoff, size_t sliding_window_average_size,
-                           const std::string &wake_word, size_t tensor_arena_size);
+  void add_wake_word_model(WakeWordModel *model);
 
 #ifdef USE_MICRO_WAKE_WORD_VAD
-  void add_vad_model(const uint8_t *model_start, float probability_cutoff, size_t sliding_window_size,
+  void add_vad_model(const uint8_t *model_start, uint8_t probability_cutoff, size_t sliding_window_size,
                      size_t tensor_arena_size);
+
+  // Intended for the voice assistant component to fetch VAD status
+  bool get_vad_state() { return this->vad_state_; }
 #endif
 
+  // Intended for the voice assistant component to access which wake words are available
+  // Since these are pointers to the WakeWordModel objects, the voice assistant component can enable or disable them
+  std::vector<WakeWordModel *> get_wake_words();
+
  protected:
-  microphone::Microphone *microphone_{nullptr};
+  microphone::MicrophoneSource *microphone_source_{nullptr};
   Trigger<std::string> *wake_word_detected_trigger_ = new Trigger<std::string>();
-  State state_{State::IDLE};
+  State state_{State::STOPPED};
 
-  std::shared_ptr<RingBuffer> ring_buffer_;
-
-  std::vector<WakeWordModel> wake_word_models_;
+  std::weak_ptr<RingBuffer> ring_buffer_;
+  std::vector<WakeWordModel *> wake_word_models_;
 
 #ifdef USE_MICRO_WAKE_WORD_VAD
   std::unique_ptr<VADModel> vad_model_;
+  bool vad_state_{false};
 #endif
 
-  tflite::MicroMutableOpResolver<20> streaming_op_resolver_;
+  bool pending_start_{false};
+  bool pending_stop_{false};
+
+  bool stop_after_detection_;
+
+  uint8_t features_step_size_;
 
   // Audio frontend handles generating spectrogram features
   struct FrontendConfig frontend_config_;
   struct FrontendState frontend_state_;
 
-  // When the wake word detection first starts, we ignore this many audio
-  // feature slices before accepting a positive detection
-  int16_t ignore_windows_{-MIN_SLICES_BEFORE_DETECTION};
+  // Handles managing the stop/state of the inference task
+  EventGroupHandle_t event_group_;
 
-  uint8_t features_step_size_;
+  // Used to send messages about the models' states to the main loop
+  QueueHandle_t detection_queue_;
 
-  // Stores audio read from the microphone before being added to the ring buffer.
-  int16_t *input_buffer_{nullptr};
-  // Stores audio to be fed into the audio frontend for generating features.
-  int16_t *preprocessor_audio_buffer_{nullptr};
+  static void inference_task(void *params);
+  TaskHandle_t inference_task_handle_{nullptr};
 
-  bool detected_{false};
-  std::string detected_wake_word_{""};
+  /// @brief Suspends the inference task
+  void suspend_task_();
+  /// @brief Resumes the inference task
+  void resume_task_();
 
   void set_state_(State state);
 
-  /// @brief Tests if there are enough samples in the ring buffer to generate new features.
-  /// @return True if enough samples, false otherwise.
-  bool has_enough_samples_();
+  /// @brief Generates spectrogram features from an input buffer of audio samples
+  /// @param audio_buffer (int16_t *) Buffer containing input audio samples
+  /// @param samples_available (size_t) Number of samples avaiable in the input buffer
+  /// @param features_buffer (int8_t *) Buffer to store generated features
+  /// @return (size_t) Number of samples processed from the input buffer
+  size_t generate_features_(int16_t *audio_buffer, size_t samples_available,
+                            int8_t features_buffer[PREPROCESSOR_FEATURE_SIZE]);
 
-  /// @brief Allocates memory for input_buffer_, preprocessor_audio_buffer_, and ring_buffer_
-  /// @return True if successful, false otherwise
-  bool allocate_buffers_();
+  /// @brief Processes any new probabilities for each model. If any wake word is detected, it will send a DetectionEvent
+  /// to the detection_queue_.
+  void process_probabilities_();
 
-  /// @brief Frees memory allocated for input_buffer_ and preprocessor_audio_buffer_
-  void deallocate_buffers_();
-
-  /// @brief Loads streaming models and prepares the feature generation frontend
-  /// @return True if successful, false otherwise
-  bool load_models_();
-
-  /// @brief Deletes each model's TFLite interpreters and frees tensor arena memory. Frees memory used by the feature
-  /// generation frontend.
+  /// @brief Deletes each model's TFLite interpreters and frees tensor arena memory.
   void unload_models_();
 
-  /** Performs inference with each configured model
-   *
-   * If enough audio samples are available, it will generate one slice of new features.
-   * It then loops through and performs inference with each of the loaded models.
-   */
-  void update_model_probabilities_();
-
-  /** Checks every model's recent probabilities to determine if the wake word has been predicted
-   *
-   * Verifies the models have processed enough new samples for accurate predictions.
-   * Sets detected_wake_word_ to the wake word, if one is detected.
-   * @return True if a wake word is predicted, false otherwise
-   */
-  bool detect_wake_words_();
-
-  /** Generates features for a window of audio samples
-   *
-   * Reads samples from the ring buffer and feeds them into the preprocessor frontend.
-   * Adapted from TFLite microspeech frontend.
-   * @param features int8_t array to store the audio features
-   * @return True if successful, false otherwise.
-   */
-  bool generate_features_for_window_(int8_t features[PREPROCESSOR_FEATURE_SIZE]);
-
-  /// @brief Resets the ring buffer, ignore_windows_, and sliding window probabilities
-  void reset_states_();
-
-  /// @brief Returns true if successfully registered the streaming model's TensorFlow operations
-  bool register_streaming_ops_(tflite::MicroMutableOpResolver<20> &op_resolver);
+  /// @brief Runs an inference with each model using the new spectrogram features
+  /// @param audio_features (int8_t *) Buffer containing new spectrogram features
+  /// @return True if successful, false if any errors were encountered
+  bool update_model_probabilities_(const int8_t audio_features[PREPROCESSOR_FEATURE_SIZE]);
 
   inline uint16_t new_samples_to_get_() { return (this->features_step_size_ * (AUDIO_SAMPLE_FREQUENCY / 1000)); }
 };
 
-template<typename... Ts> class StartAction : public Action<Ts...>, public Parented<MicroWakeWord> {
- public:
-  void play(Ts... x) override { this->parent_->start(); }
-};
-
-template<typename... Ts> class StopAction : public Action<Ts...>, public Parented<MicroWakeWord> {
- public:
-  void play(Ts... x) override { this->parent_->stop(); }
-};
-
-template<typename... Ts> class IsRunningCondition : public Condition<Ts...>, public Parented<MicroWakeWord> {
- public:
-  bool check(Ts... x) override { return this->parent_->is_running(); }
-};
-
 }  // namespace micro_wake_word
 }  // namespace esphome
 

esphome/components/micro_wake_word/preprocessor_settings.h

@@ -7,6 +7,10 @@
 namespace esphome {
 namespace micro_wake_word {
 
+// Settings for controlling the spectrogram feature generation by the preprocessor.
+// These must match the settings used when training a particular model.
+// All microWakeWord models have been trained with these specific paramters.
+
 // The number of features the audio preprocessor generates per slice
 static const uint8_t PREPROCESSOR_FEATURE_SIZE = 40;
 // Duration of each slice used as input into the preprocessor
@@ -14,6 +18,21 @@ static const uint8_t FEATURE_DURATION_MS = 30;
 // Audio sample frequency in hertz
 static const uint16_t AUDIO_SAMPLE_FREQUENCY = 16000;
 
+static const float FILTERBANK_LOWER_BAND_LIMIT = 125.0;
+static const float FILTERBANK_UPPER_BAND_LIMIT = 7500.0;
+
+static const uint8_t NOISE_REDUCTION_SMOOTHING_BITS = 10;
+static const float NOISE_REDUCTION_EVEN_SMOOTHING = 0.025;
+static const float NOISE_REDUCTION_ODD_SMOOTHING = 0.06;
+static const float NOISE_REDUCTION_MIN_SIGNAL_REMAINING = 0.05;
+
+static const bool PCAN_GAIN_CONTROL_ENABLE_PCAN = true;
+static const float PCAN_GAIN_CONTROL_STRENGTH = 0.95;
+static const float PCAN_GAIN_CONTROL_OFFSET = 80.0;
+static const uint8_t PCAN_GAIN_CONTROL_GAIN_BITS = 21;
+
+static const bool LOG_SCALE_ENABLE_LOG = true;
+static const uint8_t LOG_SCALE_SCALE_SHIFT = 6;
 }  // namespace micro_wake_word
 }  // namespace esphome
 

esphome/components/micro_wake_word/streaming_model.cpp

@@ -1,8 +1,7 @@
-#ifdef USE_ESP_IDF
-
 #include "streaming_model.h"
 
-#include "esphome/core/hal.h"
+#ifdef USE_ESP_IDF
+
 #include "esphome/core/helpers.h"
 #include "esphome/core/log.h"
 
@@ -13,18 +12,18 @@ namespace micro_wake_word {
 
 void WakeWordModel::log_model_config() {
   ESP_LOGCONFIG(TAG, "    - Wake Word: %s", this->wake_word_.c_str());
-  ESP_LOGCONFIG(TAG, "      Probability cutoff: %.3f", this->probability_cutoff_);
+  ESP_LOGCONFIG(TAG, "      Probability cutoff: %.2f", this->probability_cutoff_ / 255.0f);
   ESP_LOGCONFIG(TAG, "      Sliding window size: %d", this->sliding_window_size_);
 }
 
 void VADModel::log_model_config() {
   ESP_LOGCONFIG(TAG, "    - VAD Model");
-  ESP_LOGCONFIG(TAG, "      Probability cutoff: %.3f", this->probability_cutoff_);
+  ESP_LOGCONFIG(TAG, "      Probability cutoff: %.2f", this->probability_cutoff_ / 255.0f);
   ESP_LOGCONFIG(TAG, "      Sliding window size: %d", this->sliding_window_size_);
 }
 
-bool StreamingModel::load_model(tflite::MicroMutableOpResolver<20> &op_resolver) {
-  ExternalRAMAllocator<uint8_t> arena_allocator(ExternalRAMAllocator<uint8_t>::ALLOW_FAILURE);
+bool StreamingModel::load_model_() {
+  RAMAllocator<uint8_t> arena_allocator(RAMAllocator<uint8_t>::ALLOW_FAILURE);
 
   if (this->tensor_arena_ == nullptr) {
     this->tensor_arena_ = arena_allocator.allocate(this->tensor_arena_size_);
@@ -51,8 +50,9 @@ bool StreamingModel::load_model(tflite::MicroMutableOpResolver<20> &op_resolver)
   }
 
   if (this->interpreter_ == nullptr) {
-    this->interpreter_ = make_unique<tflite::MicroInterpreter>(
-        tflite::GetModel(this->model_start_), op_resolver, this->tensor_arena_, this->tensor_arena_size_, this->mrv_);
+    this->interpreter_ =
+        make_unique<tflite::MicroInterpreter>(tflite::GetModel(this->model_start_), this->streaming_op_resolver_,
+                                              this->tensor_arena_, this->tensor_arena_size_, this->mrv_);
     if (this->interpreter_->AllocateTensors() != kTfLiteOk) {
       ESP_LOGE(TAG, "Failed to allocate tensors for the streaming model");
       return false;
@@ -84,34 +84,55 @@ bool StreamingModel::load_model(tflite::MicroMutableOpResolver<20> &op_resolver)
     }
   }
 
+  this->loaded_ = true;
+  this->reset_probabilities();
   return true;
 }
 
 void StreamingModel::unload_model() {
   this->interpreter_.reset();
 
-  ExternalRAMAllocator<uint8_t> arena_allocator(ExternalRAMAllocator<uint8_t>::ALLOW_FAILURE);
+  RAMAllocator<uint8_t> arena_allocator(RAMAllocator<uint8_t>::ALLOW_FAILURE);
 
-  arena_allocator.deallocate(this->tensor_arena_, this->tensor_arena_size_);
-  this->tensor_arena_ = nullptr;
-  arena_allocator.deallocate(this->var_arena_, STREAMING_MODEL_VARIABLE_ARENA_SIZE);
-  this->var_arena_ = nullptr;
+  if (this->tensor_arena_ != nullptr) {
+    arena_allocator.deallocate(this->tensor_arena_, this->tensor_arena_size_);
+    this->tensor_arena_ = nullptr;
+  }
+
+  if (this->var_arena_ != nullptr) {
+    arena_allocator.deallocate(this->var_arena_, STREAMING_MODEL_VARIABLE_ARENA_SIZE);
+    this->var_arena_ = nullptr;
+  }
+
+  this->loaded_ = false;
 }
 
 bool StreamingModel::perform_streaming_inference(const int8_t features[PREPROCESSOR_FEATURE_SIZE]) {
-  if (this->interpreter_ != nullptr) {
+  if (this->enabled_ && !this->loaded_) {
+    // Model is enabled but isn't loaded
+    if (!this->load_model_()) {
+      return false;
+    }
+  }
+
+  if (!this->enabled_ && this->loaded_) {
+    // Model is disabled but still loaded
+    this->unload_model();
+    return true;
+  }
+
+  if (this->loaded_) {
     TfLiteTensor *input = this->interpreter_->input(0);
 
+    uint8_t stride = this->interpreter_->input(0)->dims->data[1];
+    this->current_stride_step_ = this->current_stride_step_ % stride;
+
     std::memmove(
         (int8_t *) (tflite::GetTensorData<int8_t>(input)) + PREPROCESSOR_FEATURE_SIZE * this->current_stride_step_,
         features, PREPROCESSOR_FEATURE_SIZE);
     ++this->current_stride_step_;
 
-    uint8_t stride = this->interpreter_->input(0)->dims->data[1];
-
     if (this->current_stride_step_ >= stride) {
-      this->current_stride_step_ = 0;
-
       TfLiteStatus invoke_status = this->interpreter_->Invoke();
       if (invoke_status != kTfLiteOk) {
         ESP_LOGW(TAG, "Streaming interpreter invoke failed");
@@ -124,65 +145,159 @@ bool StreamingModel::perform_streaming_inference(const int8_t features[PREPROCES
       if (this->last_n_index_ == this->sliding_window_size_)
         this->last_n_index_ = 0;
       this->recent_streaming_probabilities_[this->last_n_index_] = output->data.uint8[0];  // probability;
+      this->unprocessed_probability_status_ = true;
     }
-    return true;
+    this->ignore_windows_ = std::min(this->ignore_windows_ + 1, 0);
   }
-  ESP_LOGE(TAG, "Streaming interpreter is not initialized.");
-  return false;
+  return true;
 }
 
 void StreamingModel::reset_probabilities() {
   for (auto &prob : this->recent_streaming_probabilities_) {
     prob = 0;
   }
+  this->ignore_windows_ = -MIN_SLICES_BEFORE_DETECTION;
 }
 
-WakeWordModel::WakeWordModel(const uint8_t *model_start, float probability_cutoff, size_t sliding_window_average_size,
-                             const std::string &wake_word, size_t tensor_arena_size) {
+WakeWordModel::WakeWordModel(const std::string &id, const uint8_t *model_start, uint8_t probability_cutoff,
+                             size_t sliding_window_average_size, const std::string &wake_word, size_t tensor_arena_size,
+                             bool default_enabled, bool internal_only) {
+  this->id_ = id;
   this->model_start_ = model_start;
   this->probability_cutoff_ = probability_cutoff;
   this->sliding_window_size_ = sliding_window_average_size;
   this->recent_streaming_probabilities_.resize(sliding_window_average_size, 0);
   this->wake_word_ = wake_word;
   this->tensor_arena_size_ = tensor_arena_size;
+  this->register_streaming_ops_(this->streaming_op_resolver_);
+  this->current_stride_step_ = 0;
+  this->internal_only_ = internal_only;
+
+  this->pref_ = global_preferences->make_preference<bool>(fnv1_hash(id));
+  bool enabled;
+  if (this->pref_.load(&enabled)) {
+    // Use the enabled state loaded from flash
+    this->enabled_ = enabled;
+  } else {
+    // If no state saved, then use the default
+    this->enabled_ = default_enabled;
+  }
 };
 
-bool WakeWordModel::determine_detected() {
+void WakeWordModel::enable() {
+  this->enabled_ = true;
+  if (!this->internal_only_) {
+    this->pref_.save(&this->enabled_);
+  }
+}
+
+void WakeWordModel::disable() {
+  this->enabled_ = false;
+  if (!this->internal_only_) {
+    this->pref_.save(&this->enabled_);
+  }
+}
+
+DetectionEvent WakeWordModel::determine_detected() {
+  DetectionEvent detection_event;
+  detection_event.wake_word = &this->wake_word_;
+  detection_event.max_probability = 0;
+  detection_event.average_probability = 0;
+
+  if ((this->ignore_windows_ < 0) || !this->enabled_) {
+    detection_event.detected = false;
+    return detection_event;
+  }
+
   uint32_t sum = 0;
   for (auto &prob : this->recent_streaming_probabilities_) {
+    detection_event.max_probability = std::max(detection_event.max_probability, prob);
     sum += prob;
   }
 
-  float sliding_window_average = static_cast<float>(sum) / static_cast<float>(255 * this->sliding_window_size_);
+  detection_event.average_probability = sum / this->sliding_window_size_;
+  detection_event.detected = sum > this->probability_cutoff_ * this->sliding_window_size_;
 
-  // Detect the wake word if the sliding window average is above the cutoff
-  if (sliding_window_average > this->probability_cutoff_) {
-    ESP_LOGD(TAG, "The '%s' model sliding average probability is %.3f and most recent probability is %.3f",
-             this->wake_word_.c_str(), sliding_window_average,
-             this->recent_streaming_probabilities_[this->last_n_index_] / (255.0));
-    return true;
-  }
-  return false;
+  this->unprocessed_probability_status_ = false;
+  return detection_event;
 }
 
-VADModel::VADModel(const uint8_t *model_start, float probability_cutoff, size_t sliding_window_size,
+VADModel::VADModel(const uint8_t *model_start, uint8_t probability_cutoff, size_t sliding_window_size,
                    size_t tensor_arena_size) {
   this->model_start_ = model_start;
   this->probability_cutoff_ = probability_cutoff;
   this->sliding_window_size_ = sliding_window_size;
   this->recent_streaming_probabilities_.resize(sliding_window_size, 0);
   this->tensor_arena_size_ = tensor_arena_size;
-};
+  this->register_streaming_ops_(this->streaming_op_resolver_);
+}
+
+DetectionEvent VADModel::determine_detected() {
+  DetectionEvent detection_event;
+  detection_event.max_probability = 0;
+  detection_event.average_probability = 0;
+
+  if (!this->enabled_) {
+    // We disabled the VAD model for some reason... so we shouldn't block wake words from being detected
+    detection_event.detected = true;
+    return detection_event;
+  }
 
-bool VADModel::determine_detected() {
   uint32_t sum = 0;
   for (auto &prob : this->recent_streaming_probabilities_) {
+    detection_event.max_probability = std::max(detection_event.max_probability, prob);
     sum += prob;
   }
 
-  float sliding_window_average = static_cast<float>(sum) / static_cast<float>(255 * this->sliding_window_size_);
+  detection_event.average_probability = sum / this->sliding_window_size_;
+  detection_event.detected = sum > (this->probability_cutoff_ * this->sliding_window_size_);
 
-  return sliding_window_average > this->probability_cutoff_;
+  return detection_event;
+}
+
+bool StreamingModel::register_streaming_ops_(tflite::MicroMutableOpResolver<20> &op_resolver) {
+  if (op_resolver.AddCallOnce() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddVarHandle() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddReshape() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddReadVariable() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddStridedSlice() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddConcatenation() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddAssignVariable() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddConv2D() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddMul() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddAdd() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddMean() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddFullyConnected() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddLogistic() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddQuantize() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddDepthwiseConv2D() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddAveragePool2D() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddMaxPool2D() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddPad() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddPack() != kTfLiteOk)
+    return false;
+  if (op_resolver.AddSplitV() != kTfLiteOk)
+    return false;
+
+  return true;
 }
 
 }  // namespace micro_wake_word

esphome/components/micro_wake_word/streaming_model.h

@@ -4,6 +4,8 @@
 
 #include "preprocessor_settings.h"
 
+#include "esphome/core/preferences.h"
+
 #include <tensorflow/lite/core/c/common.h>
 #include <tensorflow/lite/micro/micro_interpreter.h>
 #include <tensorflow/lite/micro/micro_mutable_op_resolver.h>
@@ -11,30 +13,63 @@
 namespace esphome {
 namespace micro_wake_word {
 
+static const uint8_t MIN_SLICES_BEFORE_DETECTION = 100;
 static const uint32_t STREAMING_MODEL_VARIABLE_ARENA_SIZE = 1024;
 
+struct DetectionEvent {
+  std::string *wake_word;
+  bool detected;
+  bool partially_detection;  // Set if the most recent probability exceed the threshold, but the sliding window average
+                             // hasn't yet
+  uint8_t max_probability;
+  uint8_t average_probability;
+  bool blocked_by_vad = false;
+};
+
 class StreamingModel {
  public:
   virtual void log_model_config() = 0;
-  virtual bool determine_detected() = 0;
+  virtual DetectionEvent determine_detected() = 0;
 
+  // Performs inference on the given features.
+  //  - If the model is enabled but not loaded, it will load it
+  //  - If the model is disabled but loaded, it will unload it
+  // Returns true if sucessful or false if there is an error
   bool perform_streaming_inference(const int8_t features[PREPROCESSOR_FEATURE_SIZE]);
 
-  /// @brief Sets all recent_streaming_probabilities to 0
+  /// @brief Sets all recent_streaming_probabilities to 0 and resets the ignore window count
   void reset_probabilities();
 
-  /// @brief Allocates tensor and variable arenas and sets up the model interpreter
-  /// @param op_resolver MicroMutableOpResolver object that must exist until the model is unloaded
-  /// @return True if successful, false otherwise
-  bool load_model(tflite::MicroMutableOpResolver<20> &op_resolver);
-
   /// @brief Destroys the TFLite interpreter and frees the tensor and variable arenas' memory
   void unload_model();
 
- protected:
-  uint8_t current_stride_step_{0};
+  /// @brief Enable the model. The next performing_streaming_inference call will load it.
+  virtual void enable() { this->enabled_ = true; }
 
-  float probability_cutoff_;
+  /// @brief Disable the model. The next performing_streaming_inference call will unload it.
+  virtual void disable() { this->enabled_ = false; }
+
+  /// @brief Return true if the model is enabled.
+  bool is_enabled() { return this->enabled_; }
+
+  bool get_unprocessed_probability_status() { return this->unprocessed_probability_status_; }
+
+ protected:
+  /// @brief Allocates tensor and variable arenas and sets up the model interpreter
+  /// @return True if successful, false otherwise
+  bool load_model_();
+  /// @brief Returns true if successfully registered the streaming model's TensorFlow operations
+  bool register_streaming_ops_(tflite::MicroMutableOpResolver<20> &op_resolver);
+
+  tflite::MicroMutableOpResolver<20> streaming_op_resolver_;
+
+  bool loaded_{false};
+  bool enabled_{true};
+  bool unprocessed_probability_status_{false};
+  uint8_t current_stride_step_{0};
+  int16_t ignore_windows_{-MIN_SLICES_BEFORE_DETECTION};
+
+  uint8_t probability_cutoff_;  // Quantized probability cutoff mapping 0.0 - 1.0 to 0 - 255
   size_t sliding_window_size_;
   size_t last_n_index_{0};
   size_t tensor_arena_size_;
@@ -50,32 +85,62 @@ class StreamingModel {
 
 class WakeWordModel final : public StreamingModel {
  public:
-  WakeWordModel(const uint8_t *model_start, float probability_cutoff, size_t sliding_window_average_size,
-                const std::string &wake_word, size_t tensor_arena_size);
+  /// @brief Constructs a wake word model object
+  /// @param id (std::string) identifier for this model
+  /// @param model_start (const uint8_t *) pointer to the start of the model's TFLite FlatBuffer
+  /// @param probability_cutoff (uint8_t) probability cutoff for acceping the wake word has been said
+  /// @param sliding_window_average_size (size_t) the length of the sliding window computing the mean rolling
+  ///                                    probability
+  /// @param wake_word (std::string) Friendly name of the wake word
+  /// @param tensor_arena_size (size_t) Size in bytes for allocating the tensor arena
+  /// @param default_enabled (bool) If true, it will be enabled by default on first boot
+  /// @param internal_only (bool) If true, the model will not be exposed to HomeAssistant as an available model
+  WakeWordModel(const std::string &id, const uint8_t *model_start, uint8_t probability_cutoff,
+                size_t sliding_window_average_size, const std::string &wake_word, size_t tensor_arena_size,
+                bool default_enabled, bool internal_only);
 
   void log_model_config() override;
 
   /// @brief Checks for the wake word by comparing the mean probability in the sliding window with the probability
   /// cutoff
   /// @return True if wake word is detected, false otherwise
-  bool determine_detected() override;
+  DetectionEvent determine_detected() override;
 
+  const std::string &get_id() const { return this->id_; }
   const std::string &get_wake_word() const { return this->wake_word_; }
 
+  void add_trained_language(const std::string &language) { this->trained_languages_.push_back(language); }
+  const std::vector<std::string> &get_trained_languages() const { return this->trained_languages_; }
+
+  /// @brief Enable the model and save to flash. The next performing_streaming_inference call will load it.
+  void enable() override;
+
+  /// @brief Disable the model and save to flash. The next performing_streaming_inference call will unload it.
+  void disable() override;
+
+  bool get_internal_only() { return this->internal_only_; }
+
  protected:
+  std::string id_;
   std::string wake_word_;
+  std::vector<std::string> trained_languages_;
+
+  bool internal_only_;
+
+  ESPPreferenceObject pref_;
 };
 
 class VADModel final : public StreamingModel {
  public:
-  VADModel(const uint8_t *model_start, float probability_cutoff, size_t sliding_window_size, size_t tensor_arena_size);
+  VADModel(const uint8_t *model_start, uint8_t probability_cutoff, size_t sliding_window_size,
+           size_t tensor_arena_size);
 
   void log_model_config() override;
 
   /// @brief Checks for voice activity by comparing the max probability in the sliding window with the probability
   /// cutoff
   /// @return True if voice activity is detected, false otherwise
-  bool determine_detected() override;
+  DetectionEvent determine_detected() override;
 };
 
 }  // namespace micro_wake_word

esphome/components/microphone/__init__.py

@@ -1,12 +1,21 @@
 from esphome import automation
 from esphome.automation import maybe_simple_id
 import esphome.codegen as cg
+from esphome.components import audio
 import esphome.config_validation as cv
-from esphome.const import CONF_ID, CONF_TRIGGER_ID
+from esphome.const import (
+    CONF_BITS_PER_SAMPLE,
+    CONF_CHANNELS,
+    CONF_GAIN_FACTOR,
+    CONF_ID,
+    CONF_MICROPHONE,
+    CONF_TRIGGER_ID,
+)
 from esphome.core import CORE
 from esphome.coroutine import coroutine_with_priority
 
-CODEOWNERS = ["@jesserockz"]
+AUTO_LOAD = ["audio"]
+CODEOWNERS = ["@jesserockz", "@kahrendt"]
 
 IS_PLATFORM_COMPONENT = True
 
@@ -15,6 +24,7 @@ CONF_ON_DATA = "on_data"
 microphone_ns = cg.esphome_ns.namespace("microphone")
 
 Microphone = microphone_ns.class_("Microphone")
+MicrophoneSource = microphone_ns.class_("MicrophoneSource")
 
 CaptureAction = microphone_ns.class_(
     "CaptureAction", automation.Action, cg.Parented.template(Microphone)
@@ -22,16 +32,23 @@ CaptureAction = microphone_ns.class_(
 StopCaptureAction = microphone_ns.class_(
     "StopCaptureAction", automation.Action, cg.Parented.template(Microphone)
 )
+MuteAction = microphone_ns.class_(
+    "MuteAction", automation.Action, cg.Parented.template(Microphone)
+)
+UnmuteAction = microphone_ns.class_(
+    "UnmuteAction", automation.Action, cg.Parented.template(Microphone)
+)
 
 
 DataTrigger = microphone_ns.class_(
     "DataTrigger",
-    automation.Trigger.template(cg.std_vector.template(cg.int16).operator("ref")),
+    automation.Trigger.template(cg.std_vector.template(cg.uint8).operator("ref")),
 )
 
 IsCapturingCondition = microphone_ns.class_(
     "IsCapturingCondition", automation.Condition
 )
+IsMutedCondition = microphone_ns.class_("IsMutedCondition", automation.Condition)
 
 
 async def setup_microphone_core_(var, config):
@@ -39,7 +56,7 @@ async def setup_microphone_core_(var, config):
         trigger = cg.new_Pvariable(conf[CONF_TRIGGER_ID], var)
         await automation.build_automation(
             trigger,
-            [(cg.std_vector.template(cg.int16).operator("ref").operator("const"), "x")],
+            [(cg.std_vector.template(cg.uint8).operator("ref").operator("const"), "x")],
             conf,
         )
 
@@ -50,7 +67,7 @@ async def register_microphone(var, config):
     await setup_microphone_core_(var, config)
 
 
-MICROPHONE_SCHEMA = cv.Schema(
+MICROPHONE_SCHEMA = cv.Schema.extend(audio.AUDIO_COMPONENT_SCHEMA).extend(
     {
         cv.Optional(CONF_ON_DATA): automation.validate_automation(
             {
@@ -64,7 +81,104 @@ MICROPHONE_SCHEMA = cv.Schema(
 MICROPHONE_ACTION_SCHEMA = maybe_simple_id({cv.GenerateID(): cv.use_id(Microphone)})
 
 
-async def media_player_action(config, action_id, template_arg, args):
+def microphone_source_schema(
+    min_bits_per_sample: int = 16,
+    max_bits_per_sample: int = 16,
+    min_channels: int = 1,
+    max_channels: int = 1,
+):
+    """Schema for a microphone source
+
+    Components requesting microphone data should use this schema instead of accessing a microphone directly.
+
+    Args:
+      min_bits_per_sample (int, optional): Minimum number of bits per sample the requesting component supports. Defaults to 16.
+      max_bits_per_sample (int, optional): Maximum number of bits per sample the requesting component supports. Defaults to 16.
+      min_channels (int, optional): Minimum number of channels the requesting component supports. Defaults to 1.
+      max_channels (int, optional): Maximum number of channels the requesting component supports. Defaults to 1.
+    """
+
+    def _validate_unique_channels(config):
+        if len(config) != len(set(config)):
+            raise cv.Invalid("Channels must be unique")
+        return config
+
+    return cv.All(
+        automation.maybe_conf(
+            CONF_MICROPHONE,
+            {
+                cv.GenerateID(CONF_ID): cv.declare_id(MicrophoneSource),
+                cv.GenerateID(CONF_MICROPHONE): cv.use_id(Microphone),
+                cv.Optional(CONF_BITS_PER_SAMPLE, default=16): cv.int_range(
+                    min_bits_per_sample, max_bits_per_sample
+                ),
+                cv.Optional(CONF_CHANNELS, default="0"): cv.All(
+                    cv.ensure_list(cv.int_range(0, 7)),
+                    cv.Length(min=min_channels, max=max_channels),
+                    _validate_unique_channels,
+                ),
+                cv.Optional(CONF_GAIN_FACTOR, default="1"): cv.int_range(1, 64),
+            },
+        ),
+    )
+
+
+_UNDEF = object()
+
+
+def final_validate_microphone_source_schema(
+    component_name: str, sample_rate: int = _UNDEF
+):
+    """Validates that the microphone source can provide audio in the correct format. In particular it validates the sample rate and the enabled channels.
+
+    Note that:
+      - MicrophoneSource class automatically handles converting bits per sample, so no need to validate
+      - microphone_source_schema already validates that channels are unique and specifies the max number of channels the component supports
+
+    Args:
+        component_name (str): The name of the component requesting mic audio
+        sample_rate (int, optional): The sample rate the component requesting mic audio requires
+    """
+
+    def _validate_audio_compatability(config):
+        if sample_rate is not _UNDEF:
+            # Issues require changing the microphone configuration
+            #  - Verifies sample rates match
+            audio.final_validate_audio_schema(
+                component_name,
+                audio_device=CONF_MICROPHONE,
+                sample_rate=sample_rate,
+                audio_device_issue=True,
+            )(config)
+
+        # Issues require changing the MicrophoneSource configuration
+        # - Verifies that each of the enabled channels are available
+        audio.final_validate_audio_schema(
+            component_name,
+            audio_device=CONF_MICROPHONE,
+            enabled_channels=config[CONF_CHANNELS],
+            audio_device_issue=False,
+        )(config)
+
+        return config
+
+    return _validate_audio_compatability
+
+
+async def microphone_source_to_code(config):
+    mic = await cg.get_variable(config[CONF_MICROPHONE])
+    mic_source = cg.new_Pvariable(
+        config[CONF_ID],
+        mic,
+        config[CONF_BITS_PER_SAMPLE],
+        config[CONF_GAIN_FACTOR],
+    )
+    for channel in config[CONF_CHANNELS]:
+        cg.add(mic_source.add_channel(channel))
+    return mic_source
+
+
+async def microphone_action(config, action_id, template_arg, args):
     var = cg.new_Pvariable(action_id, template_arg)
     await cg.register_parented(var, config[CONF_ID])
     return var
@@ -72,15 +186,25 @@ async def media_player_action(config, action_id, template_arg, args):
 
 automation.register_action(
     "microphone.capture", CaptureAction, MICROPHONE_ACTION_SCHEMA
-)(media_player_action)
+)(microphone_action)
 
 automation.register_action(
     "microphone.stop_capture", StopCaptureAction, MICROPHONE_ACTION_SCHEMA
-)(media_player_action)
+)(microphone_action)
+
+automation.register_action("microphone.mute", MuteAction, MICROPHONE_ACTION_SCHEMA)(
+    microphone_action
+)
+automation.register_action("microphone.unmute", UnmuteAction, MICROPHONE_ACTION_SCHEMA)(
+    microphone_action
+)
 
 automation.register_condition(
     "microphone.is_capturing", IsCapturingCondition, MICROPHONE_ACTION_SCHEMA
-)(media_player_action)
+)(microphone_action)
+automation.register_condition(
+    "microphone.is_muted", IsMutedCondition, MICROPHONE_ACTION_SCHEMA
+)(microphone_action)
 
 
 @coroutine_with_priority(100.0)

esphome/components/microphone/automation.h

@@ -16,10 +16,17 @@ template<typename... Ts> class StopCaptureAction : public Action<Ts...>, public
   void play(Ts... x) override { this->parent_->stop(); }
 };
 
-class DataTrigger : public Trigger<const std::vector<int16_t> &> {
+template<typename... Ts> class MuteAction : public Action<Ts...>, public Parented<Microphone> {
+  void play(Ts... x) override { this->parent_->set_mute_state(true); }
+};
+template<typename... Ts> class UnmuteAction : public Action<Ts...>, public Parented<Microphone> {
+  void play(Ts... x) override { this->parent_->set_mute_state(false); }
+};
+
+class DataTrigger : public Trigger<const std::vector<uint8_t> &> {
  public:
   explicit DataTrigger(Microphone *mic) {
-    mic->add_data_callback([this](const std::vector<int16_t> &data) { this->trigger(data); });
+    mic->add_data_callback([this](const std::vector<uint8_t> &data) { this->trigger(data); });
   }
 };
 
@@ -28,5 +35,10 @@ template<typename... Ts> class IsCapturingCondition : public Condition<Ts...>, p
   bool check(Ts... x) override { return this->parent_->is_running(); }
 };
 
+template<typename... Ts> class IsMutedCondition : public Condition<Ts...>, public Parented<Microphone> {
+ public:
+  bool check(Ts... x) override { return this->parent_->get_mute_state(); }
+};
+
 }  // namespace microphone
 }  // namespace esphome

esphome/components/microphone/microphone.cpp

@@ -0,0 +1,21 @@
+#include "microphone.h"
+
+namespace esphome {
+namespace microphone {
+
+void Microphone::add_data_callback(std::function<void(const std::vector<uint8_t> &)> &&data_callback) {
+  std::function<void(const std::vector<uint8_t> &)> mute_handled_callback =
+      [this, data_callback](const std::vector<uint8_t> &data) { data_callback(this->silence_audio_(data)); };
+  this->data_callbacks_.add(std::move(mute_handled_callback));
+}
+
+std::vector<uint8_t> Microphone::silence_audio_(std::vector<uint8_t> data) {
+  if (this->mute_state_) {
+    std::memset((void *) data.data(), 0, data.size());
+  }
+
+  return data;
+}
+
+}  // namespace microphone
+}  // namespace esphome

esphome/components/microphone/microphone.h

@@ -1,5 +1,7 @@
 #pragma once
 
+#include "esphome/components/audio/audio.h"
+
 #include <cstddef>
 #include <cstdint>
 #include <functional>
@@ -20,18 +22,25 @@ class Microphone {
  public:
   virtual void start() = 0;
   virtual void stop() = 0;
-  void add_data_callback(std::function<void(const std::vector<int16_t> &)> &&data_callback) {
-    this->data_callbacks_.add(std::move(data_callback));
-  }
-  virtual size_t read(int16_t *buf, size_t len) = 0;
+  void add_data_callback(std::function<void(const std::vector<uint8_t> &)> &&data_callback);
 
   bool is_running() const { return this->state_ == STATE_RUNNING; }
   bool is_stopped() const { return this->state_ == STATE_STOPPED; }
 
- protected:
-  State state_{STATE_STOPPED};
+  void set_mute_state(bool is_muted) { this->mute_state_ = is_muted; }
+  bool get_mute_state() { return this->mute_state_; }
 
-  CallbackManager<void(const std::vector<int16_t> &)> data_callbacks_{};
+  audio::AudioStreamInfo get_audio_stream_info() { return this->audio_stream_info_; }
+
+ protected:
+  std::vector<uint8_t> silence_audio_(std::vector<uint8_t> data);
+
+  State state_{STATE_STOPPED};
+  bool mute_state_{false};
+
+  audio::AudioStreamInfo audio_stream_info_;
+
+  CallbackManager<void(const std::vector<uint8_t> &)> data_callbacks_{};
 };
 
 }  // namespace microphone

esphome/components/microphone/microphone_source.cpp

@@ -0,0 +1,98 @@
+#include "microphone_source.h"
+
+namespace esphome {
+namespace microphone {
+
+void MicrophoneSource::add_data_callback(std::function<void(const std::vector<uint8_t> &)> &&data_callback) {
+  std::function<void(const std::vector<uint8_t> &)> filtered_callback =
+      [this, data_callback](const std::vector<uint8_t> &data) {
+        if (this->enabled_) {
+          data_callback(this->process_audio_(data));
+        }
+      };
+  this->mic_->add_data_callback(std::move(filtered_callback));
+}
+
+void MicrophoneSource::start() {
+  if (!this->enabled_) {
+    this->enabled_ = true;
+    this->mic_->start();
+  }
+}
+void MicrophoneSource::stop() {
+  if (this->enabled_) {
+    this->enabled_ = false;
+    this->mic_->stop();
+  }
+}
+
+std::vector<uint8_t> MicrophoneSource::process_audio_(const std::vector<uint8_t> &data) {
+  // Bit depth conversions are obtained by truncating bits or padding with zeros - no dithering is applied.
+
+  const size_t source_bytes_per_sample = this->mic_->get_audio_stream_info().samples_to_bytes(1);
+  const size_t source_channels = this->mic_->get_audio_stream_info().get_channels();
+
+  const size_t source_bytes_per_frame = this->mic_->get_audio_stream_info().frames_to_bytes(1);
+
+  const uint32_t total_frames = this->mic_->get_audio_stream_info().bytes_to_frames(data.size());
+  const size_t target_bytes_per_sample = (this->bits_per_sample_ + 7) / 8;
+  const size_t target_bytes_per_frame = target_bytes_per_sample * this->channels_.count();
+
+  std::vector<uint8_t> filtered_data;
+  filtered_data.reserve(target_bytes_per_frame * total_frames);
+
+  const int32_t target_min_value = -(1 << (8 * target_bytes_per_sample - 1));
+  const int32_t target_max_value = (1 << (8 * target_bytes_per_sample - 1)) - 1;
+
+  for (size_t frame_index = 0; frame_index < total_frames; ++frame_index) {
+    for (size_t channel_index = 0; channel_index < source_channels; ++channel_index) {
+      if (this->channels_.test(channel_index)) {
+        // Channel's current sample is included in the target mask. Convert bits per sample, if necessary.
+
+        size_t sample_index = frame_index * source_bytes_per_frame + channel_index * source_bytes_per_sample;
+
+        int32_t sample = 0;
+
+        // Copy the data into the most significant bits of the sample variable to ensure the sign bit is correct
+        uint8_t bit_offset = (4 - source_bytes_per_sample) * 8;
+        for (int i = 0; i < source_bytes_per_sample; ++i) {
+          sample |= data[sample_index + i] << bit_offset;
+          bit_offset += 8;
+        }
+
+        // Shift data back to the least significant bits
+        if (source_bytes_per_sample >= target_bytes_per_sample) {
+          // Keep source bytes per sample of data so that the gain multiplication uses all significant bits instead of
+          // shifting to the target bytes per sample immediately, potentially losing information.
+          sample >>= (4 - source_bytes_per_sample) * 8;  // ``source_bytes_per_sample`` bytes of valid data
+        } else {
+          // Keep padded zeros to match the target bytes per sample
+          sample >>= (4 - target_bytes_per_sample) * 8;  // ``target_bytes_per_sample`` bytes of valid data
+        }
+
+        // Apply gain using multiplication
+        sample *= this->gain_factor_;
+
+        // Match target output bytes by shifting out the least significant bits
+        if (source_bytes_per_sample > target_bytes_per_sample) {
+          sample >>= 8 * (source_bytes_per_sample -
+                          target_bytes_per_sample);  //  ``target_bytes_per_sample`` bytes of valid data
+        }
+
+        // Clamp ``sample`` to the target bytes per sample range in case gain multiplication overflows
+        sample = clamp<int32_t>(sample, target_min_value, target_max_value);
+
+        // Copy ``target_bytes_per_sample`` bytes to the output buffer.
+        for (int i = 0; i < target_bytes_per_sample; ++i) {
+          filtered_data.push_back(static_cast<uint8_t>(sample));
+          sample >>= 8;
+        }
+      }
+    }
+  }
+
+  return filtered_data;
+}
+
+}  // namespace microphone
+}  // namespace esphome