A human voice provides meaning and identity simultaneously. For over fifty years, companies such as AT&T and IBM have sought to have a computer synthesize speech that matches the expressive meaning and individualized identity capabilities of a human voice. Bell Labs first demonstrated the Voder at the 1939 New York World's Fair. Haskins Laboratories completed the Pattern Playback in 1950, and MIT developed MITtalk, later DECTalk, as the first computer based synthesizer in the 1980's.
Despite their efforts, and billions of dollars of investment, humans have been unwilling to accept the output of TTS engines as equivalent to human speech. For most applications, few humans will tolerate even a single sentence of voice synthesis.
Expressive human speech is prosodic — it is musical, with melodies and rhythms that clarify words and phrases, as well as create the auditory context to enhance a listener's comprehension. The prosodic musical overlay communicates emotion and lowers the cognitive load. Natural sounding speech, whether spoken or synthesized, enables the human mind to handle listening and comprehension as a background task, not requiring full concentration.
Lessac has developed a new automated method for producing near human quality expressive speech from plain text; synthesized speech produced by the Lessac Automated Narrator sounds quite similar to speech from a skilled newscaster. Lessac was among the highest-ranked in the Blizzard Challenge in 2010, 2011 and 2012, an annual global competition among 10 to 20 mainly academic research institutions to develop the best voice synthesizer from a common set of voice data. The entrant that ranked first relied on the output of Lessac's front-end text parser to drive their synthesizer. Most of the Blizzard Challenge entries each year are far superior to available commercial text-to-speech engines.
The foundation for Lessac's unique technology is the 50 years of voice research by the late renowned speech professor and practitioner Arthur Lessac. Since the early 1930s, Arthur Lessac and his team of practitioners have been investigating how the human body and voice function naturally and instinctively. Arthur Lessac established the Lessac Training and Research Institute to extend his legacy.
Lessac Technologies recently launched LessacAudioBooks.com , where we offer nearly 400 audiobooks synthesized using our automated narration technology. This represents over 6,000 hours of automated narration, all of which was produced in the internet cloud in less than 30 elapsed hours. Many of these audiobooks can be downloaded for free, and most of the titles that are not free are only $0.99. Sample the first chapter of each of these audiobooks for free.
Please click here to play a synthesized version of this page.
256 Highland Street
West Newton, MA 02465