Although the electrolarynx (EL) is commonly used in the rehabilitation of laryngectomy patients, the resulting speech quality is decidedly poor and unnatural. To better understand the causes of this poor speech quality, I investigated how several aberrant properties of EL speech contribute to this unnatural quality. Using a combination of perceptual experiments and acoustic modeling, I was able to quantify the effect these aberrant properties have on EL speech quality. The results of this work lay the groundwork to direct future efforts to improve the state of EL speech.
I have worked on research projects in and out of the speech-processing field. These include incorporating space-time adaptive radar processing algorithms into ground moving target tracking, developing new space-based radar processing algorithms, and creating a simplified version of the mixed-excitation linear predictive (MELP) speech coder. My current research activities involve using super-resolution techniques to detect airport runway anomalies using real-beam radar and working on incorporating the reduced MELP speech coder into communication systems used by the military.
The broad scope of the SHBT program gives a student a wide variety of research areas to pursue, and in my case, a completely unexpected one. The interdisciplinary and inter-institutional nature of the SHBT program also provides a unique opportunity for students to interact with both the clinical and the scientific/engineering worlds. This gives students the option of applying the scientific concepts learned in the coursework towards research that has clinical applications or towards research that is more purely scientific in nature. I chose the former path and have found that the quantitative training I received is not only applicable toward clinical problems, but towards problems in fields completely unrelated to speech and hearing.
