google text-to-speech supports custom voices, as described here, here, and here.
Note that a custom voice is described not by a voice name, but by an object containing a model name and a reported usage.
I think in order to support this, it might be best to extend the Speech Credentials panel so that when google is selected, in addition to the JSON service key file there is a section below guarded by a checkbox labeled "Use custom voices". When checked, that opens a section that allows the user to add/edit/delete a list of custom voices. For each, they will provide a voice name that they choose, a language (from dropdown), the associated model, and the associated usage.
Then, in the applications view when they select google for TTS and a language, we will augment the list of voices with the custom voices, showing the name they gave, e.g. "Sally (custom)"
google text-to-speech supports custom voices, as described here, here, and here.
Note that a custom voice is described not by a voice name, but by an object containing a model name and a reported usage.
I think in order to support this, it might be best to extend the Speech Credentials panel so that when google is selected, in addition to the JSON service key file there is a section below guarded by a checkbox labeled "Use custom voices". When checked, that opens a section that allows the user to add/edit/delete a list of custom voices. For each, they will provide a voice name that they choose, a language (from dropdown), the associated model, and the associated usage.
Then, in the applications view when they select google for TTS and a language, we will augment the list of voices with the custom voices, showing the name they gave, e.g. "Sally (custom)"