The issue is the following and has stumped me for a good half day now: In documentation, it is clearly stated that there should be 52 entries for VoiceStyleData[], ranging a bunch of languages. However 3 different machines I tried, with different licences, I am consistently getting only 8 and only in English.
This is exactly the expression I evaluated on a newly rebooted machine, the only difference is I told it to display as a table using the //Dataset function. No previous lines are executed. My version is 14.1.0.0, Windows version. Nothing else was modified from a “normal” install.
To me (14.2 on Windows) it looks like VoiceStyleData is just a simple collection of seven named English speakers. Try evaluating the examples in the documentation, without calling ElevenLabs or GoogleSpeech.
Thanks Hans! This worked and I got Hedda to speak! However, I cannot imagine that all 52 languages from the documentation need to be installed this way, including other voices for German itself. Knowing Mathematica, there has to be a better way.
There are ~50 locales/languages one can load (additional) on Windows (at least Windows Pro). When adding, you have a checkbox choice. You can select all the locales/languages and see how many are then available to Mathematica after loading. And How many are useful. I indicated a solution where you only chose German.
One question was
“Any ideas how to particularly install the German voices would be
greatly appreciated.”
So you don’t have to do this each time. You could have chosen more items to install more in one go.
On a Windows system I believe Mathematica may be utilizing Speech Application Programming Interface (SAPI) 4 and 5. I don’t know what the equivalent is on Mac OS or Nix variants. But there are some Text-To-Speech TTS standards and file formats. The files that create the rules of how sound is produced from text etc.
There are free Text to Speech Natural Voices (SAPI 4 and SAPI 5). I have not tried them nor can vouch. But that would be another way to add Voices to your system.
But I believe when the VoiceStyleData[] examples was created and on whose ever system, that is what they had ~52. As Hans Milton posted if you re-run the examples you get something different based on your system.
