Researchers at Harvard, MIT and Massachusetts General Hospital are working together to improve the sound quality of EL speech by creating an approach that can automatically add missing pitch variation to the speech. This is a significant advance because listeners rate EL speech that includes pitch variation to be significantly more natural/human sounding than regular (monotone) EL speech. The research team is now utilizing this new understanding to begin correcting other acoustic factors that contribute to the poor quality and reduced intelligibility of EL speech. Ultimately, they intend to collaborate with companies that make EL devices to implement their improvements in systems that will make real differences in the lives of laryngectomy patients. The capability for more natural/human-sounding (and ultimately more intelligible) EL speech will allow laryngectomy patients to communicate more effectively, giving them a new voice after cancer surgery and enhancing their quality of life.
This study has also furthered the basic understanding concerning which aspects of normal speech production contribute effectively to creating natural speech sounds -- an understanding that can be useful to researchers in speech science and in addressing other speech disorders.
The talk, "F0 control in electrolarynx speech" (3aSC8) by Yoko Saikachi, Kenneth Stevens, and Robert Hillman is at 11:15 a.m. on Wednesday, November 12.
3) UNDERWATER Wi-Fi
|Contact: Jason Bardi|
American Institute of Physics