Project employee and Ph.D. student Amir Poorjam presented a paper entitled Dominant Distortion Classification for Pre-Processing of Vowels in Remote Biomedical Voice Analysis at Interspeech 2017 in Stockholm on August 21 2017. The paper was presented as part of a session on Pathological Speech and Language and was co-authored by Max Little, Jesper Rindom Jensen, and Mads Græsbøll Christensen, all of whom are involved in the Parkinsons project.
Abstract: Advances in speech signal analysis facilitates the development of techniques for remote biomedical voice assessment. However, the performance of these techniques are affected by noise
and distortion in signals. In this paper, we focus on the vowel /a/ as the widely-used voice signal for pathological voice assessments and investigate the impact of four major types of distortion that are commonly present during recording or transmission in voice analysis, namely background noise, reverberation, peak clipping and compression on Mel-frequency cepstral coefficients (MFCCs) as the most popular features in many voice assessments. Then, we propose a new distortion classification
approach to detect the most dominant distortion in such voice signals. The proposed method involves MFCCs as frame-level features and a support vector machine as the classifier to detect the presence and type of distortion in frames of a given voice signal. Experimental results over the healthy and Parkinson’s voices show the effectiveness of the proposed approach in distortion detection.