Sorry @testman, I can’t edit pool post after 5 min. I think only moderators can do this.
For now, I don’t think so that Mycroft will have some personnality. It’s just an assistant who respond to request. It will be a very poor AI at first based on “if then else” action, if I well understand (maybe @ryanleesipes could explain better than me the two approach in AI)
@Winael@testman I’ve played around with how to make it so that we can have some personality added in. We’ve also talked about using chatbots to make for decent conversational capacity (cuz some of them are really good and have performed well on the Turing tests).
Not any kind of computer. I’d pick Julie London or Sarah Vaughan, myself, but they’re not available. But a real voice, as much as possible. Interesting article about tts at http://lwn.net/Articles/656376/, I thought.
@ryanleesipes about personality I loved the way of how TARS from Interstellar is set up :
TARS’ personality can be characterized as witty, sarcastic, and humorous, traits programmed into him to make him a better suited companion.
But I think that kind of traits are difficult to program
A chat box would most definitely be interesting. I was thinking of something like:
“Mycroft, start a conversation”
"Hello, how are you?"
And saying a command with Mycroft in the beginning of a command or a pause for a considerable amount of time would end the seamless conversation.
Unfortunately the choices that are presented in this thread won’t ever happen. Most entities like this use TTS voices so, we are limited to only those that are available. Still, it’s fun to pretend. =)
If we’re looking at voices, how about Peter Jones (the voice of the Book in Hitchhiker’s Guide To The Galaxy), Stephen Fry (Wooster and Jeeves), Hugo Weaving (Agent Smith in The Matrix), Patrick Stewart (Jean-Luc Picard in ST:NG), Ian McKellan (Gandalf in LOTR), or Eric Idle (of Monty Python fame.)
Yes I have heard of phonemes @Winael however I don’t think phonemes are going to be up to par anytime soon. I could be wrong. I usually tinker with TTS since it is widely used and easy to implement. Do you know of anyone besides Ircam who is working with them for voice synthesis?