We've come a long way since the first text-to-speech began reading out aloud. Today, voice synthesis is capable of reading text with a surprisingly natural feel... but things are about to be pushed to the next level. AT&T is researching a system that will be able to read eBooks using diferent voices for each character and even making "funny voices" is kid's books - just like a normal person would do when reading to their kids.
In the future, AT&T hopes to be able to automatically assign specific voice types to characters automatically (funny mellow voices, more scary ones, etc.) and even use parent's voices so kids can hear their parents reading to them... even when they're not around. (That should make things interesting... as the potential for misuse and abuse is... well... staggering!)
Just like anyone can easily tweak their photos these days, using software like photoshop; it seems that soon we won't be able to trust anything we hear. (Well... in reality we already can't, but with more powerful tools becoming more accessible to "regular people", I suspect we'll see a boom in "abuse" scenarios.)
No comments:
Post a Comment