Resolve NoAudioReceived issue

Fixes: #416
Fixes: #421
Fixes: #436
Fixes: #437
Fixes: #440
Fixes: #442
Fixes: #443
Fixes: #444
Signed-off-by: rany <rany2@riseup.net>

Author: rany2

src/edge_tts/communicate.py

@@ -392,6 +392,7 @@ def __parse_metadata(self, data: bytes) -> TTSChunk:
         raise UnexpectedResponse("No WordBoundary metadata found")
 
     async def __stream(self) -> AsyncGenerator[TTSChunk, None]:
+        # TODO: support webm-24khz-16bit-mono-opus properly. Used by macOS.
         async def send_command_request() -> None:
             """Sends the command request to the service."""
             word_boundary = self.tts_config.boundary == "WordBoundary"
@@ -506,9 +507,13 @@ async def send_ssml_request() -> None:
 
                     # At termination of the stream, the service sends a binary message
                     # with no Content-Type; this is expected. What is not expected is for
-                    # an MPEG audio stream to be sent with no data.
+                    # an audio stream to be sent with no data.
                     content_type = parameters.get(b"Content-Type", None)
-                    if content_type not in [b"audio/mpeg", None]:
+                    if content_type not in (
+                        b"audio/mpeg",
+                        b"audio/webm; codec=opus",
+                        None,
+                    ):
                         raise UnexpectedResponse(
                             "Received binary message, but with an unexpected Content-Type."
                         )

src/edge_tts/constants.py

@@ -1,32 +1,27 @@
 """Constants for the edge_tts package."""
 
-BASE_URL = "api.msedgeservices.com/tts/cognitiveservices"
+BASE_URL = "speech.platform.bing.com/consumer/speech/synthesize/readaloud"
 TRUSTED_CLIENT_TOKEN = "6A5AA1D4EAFF4E9FB37E23D68491D6F4"
 
-WSS_URL = (
-    f"wss://{BASE_URL}/websocket/v1?Ocp-Apim-Subscription-Key={TRUSTED_CLIENT_TOKEN}"
-)
-VOICE_LIST = (
-    f"https://{BASE_URL}/voices/list?Ocp-Apim-Subscription-Key={TRUSTED_CLIENT_TOKEN}"
-)
+WSS_URL = f"wss://{BASE_URL}/edge/v1?TrustedClientToken={TRUSTED_CLIENT_TOKEN}"
+VOICE_LIST = f"https://{BASE_URL}/voices/list?trustedclienttoken={TRUSTED_CLIENT_TOKEN}"
 
 DEFAULT_VOICE = "en-US-EmmaMultilingualNeural"
 
-CHROMIUM_FULL_VERSION = "140.0.3485.14"
+CHROMIUM_FULL_VERSION = "143.0.3650.75"
 CHROMIUM_MAJOR_VERSION = CHROMIUM_FULL_VERSION.split(".", maxsplit=1)[0]
 SEC_MS_GEC_VERSION = f"1-{CHROMIUM_FULL_VERSION}"
 BASE_HEADERS = {
     "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36"
     f" (KHTML, like Gecko) Chrome/{CHROMIUM_MAJOR_VERSION}.0.0.0 Safari/537.36"
     f" Edg/{CHROMIUM_MAJOR_VERSION}.0.0.0",
-    "Accept-Encoding": "gzip, deflate, br",
+    "Accept-Encoding": "gzip, deflate, br, zstd",
     "Accept-Language": "en-US,en;q=0.9",
 }
 WSS_HEADERS = {
     "Pragma": "no-cache",
     "Cache-Control": "no-cache",
     "Origin": "chrome-extension://jdiccldimpdaibmpdkjnbmckianbfold",
-    "Sec-WebSocket-Protocol": "synthesize",
     "Sec-WebSocket-Version": "13",
 }
 WSS_HEADERS.update(BASE_HEADERS)