I found this linked from synthesizers.com. Analyze a voice and press piano keys to render a somewhat understandable rendition of the original speech. There’s a brief view of the guts, but it doesn’t look like Midibox stuff. But, hey, it’s controlled with Linux!
