It’s not documented on the TTS page that you can not only pick a language and a gender, but also pick a specific voice with YAML configuration.
It’s currently based on Azure Speech Service.
I think not documenting it is done on purpose if Nabu Casa decides at some point to change the underlying service for TTS. So be aware that specifying a voice can break your setup if there are changes done by Azure or Nabu Casa.
service: tts.cloud_say
data:
language: fr-FR
entity_id: media_player.mini_salle_de_bain
options:
voice: JeromeNeural
message: Bonjour à toutes et à tous
All the available voices are listed here: