Edward, I think you have an interesting idea here. It reminded me of the scene in _2001 A Space Odyssey_ where Hal 9000 is having the higher functions of his brain disconnected and he tries to sing "Daisy" while he still has enough connected to do it. From my experience working with espeak I'd say that getting a computer to sing, even badly, may be beyond what we can reasonably do today. Espeak can speak at different pitches and rates, but I don't think you can change pitch and rate while speech is going on and if you could that's still a long way off from actual singing.
If there is value in making a Sugar Activity that acts as a regular Karaoke machine that may be doable. There is Karaoke software written in Python that can do Karaoke and it can play OGG files. It may be possible to make a Sugar Activity version of this software. It's called cdgtools and the link is: http://www.kibosh.org/cdgtools/. James Simmons > We have a starting point for achieving fluency naturally in the > karaoke-colored TTS speech engine under development. We need to > provide it as a resource to all Activities that deal in text. We > definitely want it to be able to read stories to the children. > > We also should use it to create a singalong program stocked with folk > songs of the culture, because Same-Language-Subtitled Bollywood > musicals and TV singalongs have proven to be the best literacy program > in India, and indeed anywhere in the world. Singing engages parts of > the brain that are not involved in reading text, and makes singing > both more enjoyable and more memorable than reading text. Rhyme and > meter are also important aids to memory and to enjoyment. > _______________________________________________ IAEP -- It's An Education Project (not a laptop project!) IAEP@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/iaep