* fixes for unlocking mutex and minimizing time under lock
* call .get() on future returned from speechSynthesizer->SpeakTextAsync
* mod_azure_tts: various fixes, including dangling session lock and reducing latency
* mod_deepgram_transcribe: add support for transcribing filler words
* explicit flags for boost
* remove pragma statements
* remove some logging in mod_azure_tts
* pragma back in
* compile error fixes
* mod_azure_tts: set Playback-File-Type to tts_stream