[dingus]

There's also a terminal incantation to input a text file and output a multi-hour long sound file of the entire document processed with text-to-speech.

It's about what you'd expect it to sound like... almost unbearable. But maybe the updated Siri voices are better, I haven't tried it for a few years.

[ivansavz]

I believe the Alex voice is not synth, but actual phoneme voice files of a real human.

When I look in `/System/Library/Speech/Voices` and run `du -sh *` I see Alex.SpeechVoice is 877MB so godda be something wav-like in there.