It's hard not to see it as researchers doing typical ivory tower gatekeeping. If it were only a few of them that neglected to provide simple tools and installation instructions for common people, I would chalk it up to laziness. But this is the standard way with TTS projects, every time it takes me ten minutes of head scratching until I get the thing to work. It's been this way for years.

To make it twice as frustrating, these new systems never seem to trickle down to FOSS software users; they get incorporated into commercial products but the average linux desktop user with vision problems is still left suffering with Festival. Any such user who wants to use these new models is left to figure out how to integrate these systems themselves.

Anyway, I do appreciate them publishing this. I got it working now and it will suit my needs well; it's the best CPU-based TTS that I've managed to get running thus-far, and I think the quality will be good enough for narrating ebooks. Good enough for now, I'll figure out how to get Firefox using this later.