More and more I’ve been wanting to listen to blogs rather than read them. Why? Because often I’m busy programming and whereas I don’t mind listening to the radio while I work, I do have a tough time working and reading blogs simultaneously. Similarly, sometime I want to take these text-to-podcast posts with me as I travel about or get some exercise.
I checked out a program called TextAloud with the AT&T Natural Voices (which are very nice), however, I’m crazy enough to want more.
This is what I’d like to see:
I want to hear the blog posts not as one speaker, but “in the voice of the author.” I’d like to hear Robert Scoble—-or at least the synthesized voice that he recommends–read his posts. And if Robert has an extended excerpt of Don Dodge in one of his blog posts, I want to hear the section in Don’s voice. Same goes for the posted comments. If Christopher Coulter posts a response to Robert, I want to hear it in Christopher’s voice (or the one he recommends).
I have no idea if synthesized voices can be created ad hoc. I can imagine it’s non trivial–especially if high quality is the goal. I don’t know if companies extract human-recorded word snippets creating a large dictionary or are able to do more. Likewise I don’t know if it’s possible to start with a handful of voices and morph them to approximate someone else’s voice. I’m guessing either way is not too practical today, but it would be nice.
Short of everyone creating their own Net voice, another possibility would be to have people pick their own “avatar” voices from a public pool and publish it somewhere on their “blog.” Then when a blog reader happens upon a post by Robert Scoble, let’s say, the program can do a lookup via his URL to see which voice to use. I imagine I could maintain my own list too–I could use AT&T’s Mike voice for Robert and Charles for Christopher. Sounds like too much work. Maybe we could have a voting system where people collectively decide which voice to use for each :-).