Have you tried the Google Services app? That uses the latest and greatest TTS from Google. I think their new voices are really good.
If you have Sonos speakers you could have a look at the HTTP-API from them, if I remember correctly it is possible to enable Microsoft TTS instead of the default implementation of Google (it uses the old API / voices of Google btw).
Experimental support for TTS. Today the following providers are available:
- Microsoft Cognitive Services (Bing Text to Speech API)
- AWS Polly
- Google (default)
- macOS say command
source: GitHub - jishi/node-sonos-http-api: An HTTP API bridge for Sonos easing automation. Hostable on any node.js capable device, like a raspberry pi or similar.