Configure Text-to-Speech by endpoint

Currently Speech-to-Text is configured by endpoint, and Text-to-Speech is configured by region.  Configuration by region does not work in air gapped clouds and prevents TTS from functions.  

Please add ability to configure TTS by endpoint. 