Parametric Assessment of Esophageal Speech in Post-Laryngectomy Patients

In this paper, we present an analysis of esophageal speech recorded for 27 patients after larynx excision. Classical time-frequency parameters were determined and compared with undeformed, natural speech. Next, we described voice source parameters using inverse filtering techniques. Also, an attempt was made to find out useful parameters for esophageal speech quality description. We discovered that some parameters responsible for speech intelligibility are similar to those in the natural human voice. Another conclusion drawn is the possibility of voice quality assessment by using glottal flow characteristics. The results are promising, showing the potential of the proposed methods in monitoring the voice rehabilitation process.