Ready for an emotional Siri? Computer voices are learning how to sound scared or excited

fujitsus new speech synthesis technology will tell alarmed siri

Fujitsu has announced the development of speech synthesis technology that can tell you information in an appropriate tone, thereby doing away with the chore of determining what emotion to feel.  In case of emergencies, the technology will deliver messages in an alarming tone, which in turn will give you a clue that you should freak out. In less stressful conditions, the technology will convey messages according to your preference. According to a press release sent out by the company, messages can be conveyed through “voices that are perceived to be endearing, or distinctive voices for particular characters.” No word yet if it can do an angry James Earl Jones voice. Our fingers are crossed.

“Current speech synthesis technology, widely employed in society, is able to read-out a variety of texts, but in a monotone voice,” said Fujitsu. “For this reason, there is a need for synthesized speech to be able to convey spoken text to listeners in accordance with the given circumstances, making it easy to understand.”

Speech synthesis technology is commonly used for broadcasting traffic conditions, disaster prevention announcements, museum audio guides, and car GPS systems. Existing speech synthesis tools use large volumes of pre-recorded speech waveforms. Fujitsu uses a different method, focusing instead on voice quality, intonation, and pauses. Aside from the ability to determine the appropriate tone, the technology also conveys messages in a clear voice despite noisy environments. There’s no solid timeline for releasing the technology, but Fujitsu says practical applications could come by the end of 2014.