Include ssrc parameter when sending a SPEAKING payload

As it is a required parameter.

Don't unnecessarily send a second SPEAKING payload after
connecting to voice

We do need to send a SPEAKING payload in order to set our SSRC value
after connecting to voice, yet that can be with a `state` of 0
(SpeakingState.none).

There is no reason to send 2 packets; one marking ourselves as
speaking, and then not speaking.

discord/gateway.py

@@ -915,6 +915,7 @@ class DiscordVoiceWebSocket:
             'd': {
                 'speaking': int(state),
                 'delay': 0,
+                'ssrc': self._connection.ssrc,
             },
         }
 
@@ -990,7 +991,11 @@ class DiscordVoiceWebSocket:
     async def load_secret_key(self, data: Dict[str, Any]) -> None:
         _log.debug('received secret key for voice connection')
         self.secret_key = self._connection.secret_key = data['secret_key']
-        await self.speak()
+
+        # Send a speak command with the "not speaking" state.
+        # This also tells Discord our SSRC value, which Discord requires
+        # before sending any voice data (and is the real reason why we
+        # call this here).
         await self.speak(SpeakingState.none)
 
     async def poll_event(self) -> None: