Maybe I’ve been watching too many episodes of “Mob Wives,” but when a press release hit my desk about Mafia Voice, a text-to-voice generator promising that I can talk like Big Ang and the girls, I was intrigued.
Well, not so intrigued that I wanted to pay for the software download—although my boss might find it amusing—but I wondered what other sorts of fun I could have with voice generators and accents. In my quest I found myself doing quite a bit of, ahem, research, and saw that there is a real need for this technology. Typing in words and listening to them played back in various accents and languages helps people who have lost their voice due to medical conditions, those trying to master foreign languages, and actors hoping to improve their skills.
One of the earliest innovators of text-to-speech was AT&T Labs, which produces AT&T Natural Voices.
“The type of text-to-speech we do is called a “concatenative” system, meaning that we record a human speaker to make a voice database,” the company said. “We re-use small chunks of the recordings to create new sentences containing words that were never recorded. Further, we do “unit selection” synthesis. This means that we use large voice databases and do clever searches on-the-fly to find chunks in the voice database that best match the requested sentences.”
Personally, I found AT&T’s Crystal and her other friends who “spoke” in U.S., U.K., and Indian-accented English, as well as Spanish, German and French, very robotic, but it was fun to type in outrageous sentences and hear them played back.
A site that I really liked is The International Dialects of English Archive, a free, online archive of primary source dialect and accent recordings founded and directed by Paul Meier, a dialect coach at the University of Kansas. All of the site’s recordings are in English, are of native speakers, and include both English language dialects and English spoken in the accents of other languages, including those spoken in Africa, Asia, Europe, and South America. Also featured are speech samples of what the site calls “desirable” non-regional style of American English.
Since I moved to Savannah, Ga. from New York last year, people tell me all the time that I talk funny. Bless their hearts. Growing up in a suburb outside of Manhattan, I prefer to think that I have that “desirable” non-regional accent. If I want to fit in, maybe I’ll go the site and click on the “Georgia” sample of accents. But when I get nostalgic, all I have to do click on my home state of New York, and hear a variety of real people speaking with Bronx, Brooklyn, Queens, Long Island and other accents that remind me that there’s no place like home.