AT&T Labs - Research

Text-To-Speech (TTS) -- The Synthesis of Audible Speech from Text

> Home   |   Demo   |   FAQ   |   Publications   |   Contact

AT&T Natural Voices™

AT&T Natural Voices™ (NV) provides high-quality conversion of text to audio for US English and US Spanish. The demo page available on this website provides samples of each available voice for evaluation. Note that text supplied for conversion must be in-language, i.e. no translation. NV is available for use in three basic ways.

  • The AT&T Developer Program offers speech synthesis and recognition as cloud services for almost any networked device through its Speech API.

  • Wizzard Speech LLC provides sales, licensing and support for desktop and server SDK editions of Natural Voices through version 5.3 for US English and US Spanish. Visit their website for further details.

  • AT&T Labs may also license speech technology directly to strategic partners.

Please be advised that the version of AT&T Natural Voices used in the online Research demo may change without warning and may differ in quality and voice selection from commercially available product versions.

The demo site may also be unavailable from time to time. We will post warnings when outages are known in advance.

What is TTS?

TTS, or Text-To-Speech, is the creation of audible speech from computer readable text. There is more background in the FAQ.

Our research group's charter is to increase the naturalness of speech synthesis significantly while maintaining good intelligibility. Next-Generation TTS was introduced in 1998 and marked a dramatic leap in naturalness. Now called AT&T Natural Voices™, it is still broadening the range of applications in which TTS can be deployed.

In some cases, such as telephone response, a TTS voice may supply the only feedback a user gets, and the quality of an entire service may be judged by that voice. TTS is closest to the customer's ear.(R)


   > TV Commercial Highlights AT&T Voice Technology

In April 2006 AT&T launched a national television commercial highlighting voice technology, its long development history within AT&T, and its use in current services. The robot hand at the beginning is demonstrating the Voder ("VOice coDER"), shown by Homer Dudley at the 1939 World's Fair. A human operator used the contols to manually control the sound output.

Please note the following regarding some of the material appearing in this AVI video clip:

(c) 2006 AT&T Operations, Inc. All Rights Reserved.    WarGames TM & (c) 1983 Metro-Goldwyn-Mayer Studios Inc. All Rights Reserved.    TM & (c) H-B.    TM @ (c) Turner Ent. Co. (s06)

   > Red Planet

Red Planet, the Warner Bros. movie released on 11/09/2000, uses our TTS as the voices of the space ship (female) and the astronauts' space suits (male). Here's the press release.

   > I Am Sam

This movie, starring Sean Penn and Michelle Pfeiffer, used our "Crystal" as the voice of Pfeiffer's voice-activated cell phone. Please note that we take no responsiblity for the infuriatingly inaccurate speech recognition in the movie :-)

   > Discover Magazine

Discover magazine describes synthesized speech as both a triumph of technology and the fruition of a very old dream. The article follows the evolution of AT&T Natural Voices™, the productized version of our work.

Unfortunately, the article credits us with inventing just about every facet of modern TTS technology. As I understand it, AT&T was asked to verify certain facts but did not get to preview the article prior to publication. (In fact, my mother-in-law read it long before any of us.) We take some comfort knowing that people active in the field are well aware who invented what. And I'm sure our friends and colleagues will refresh our memories from time to time :-)

Let Your Application Speak For Itself!

TTS technology can be useful whenever a computerized application needs to communicate with a customer or user, and so is especially useful in telephone services. Though recorded material still provides the highest quality, recordings are often impractical due to cost or time constraints. Some typical uses for TTS technology are:

  • customer support dialog systems, e.g. help desks
  • interactive voice response (IVR), e.g. banking, class enrollment
  • unified messaging, especially email reading
  • screen reader programs for the sight-impaired