ARI Guest Talk: 17.07.2024, 1.30pm
The voice is an essential component of speech conveying both linguistic and paralinguistic information. In particular, it facilitates the encoding of (i) prosody, intonation, and emphasis, (ii) emotion, sentiment, mood, attitude, and intention (iii) relationship dynamics, including intimacy, confidentiality, respect, dominance, and submission, (iv) health characteristics, (v) speaker identity including age, gender, and personality traits, but also (vi) semantic differences in some languages. As a result, individuals suffering from speech pathologies often encounter difficulties in communication scenarios including face-to-face conversations and verbal telecommunications.
In the first part of this talk we will have a look at the tracking of fundamental frequency (F0) in pathological speech, which is pivotal to many processing steps that may follow. Problems in the tracking of F0 in pathological speech include high levels of additive and modulation noise, and challenges and ambiguities in the acquisition of the ground truth. In the second part, we will touch upon more advanced approaches to modelling the voice source. Here we will look at a model of glottal area waveforms capable of (de)coupling resulting in (de)synchronization effects, as well as a 3D kinematic model of the vibrating surfaces of the vocal folds.
Philipp Aichinger is affiliated with the Medical University of Vienna (Dept. ENT, Div. Phon.-Log.), where he leads the Speech and Hearing Science Lab. After graduating in Electrical Engineering and Audio Engineering (Dipl.-Ing., TU Graz, KU Graz) in 2009, he obtained the Ph.D. (Dr.techn.) degree from the TU Graz (Inst. Signal Process. Speech Comm.) in 2015. Philipp is first or senior author of currently 15 Journal papers and 23 conference papers. He is a senior member of IEEE (Signal Processing Society) and member of the Acoustical Society of America (ASA), the Audio Engineering Society (AES), the American Speech-Language-Hearing Association (ASHA), the Association for Research in Otolaryngology (ARO), and the International Speech Communication Association (ISCA). Philipp was a visiting researcher to the UK/FAU Erlangen-Nuremberg, Germany in 2022. He is associated editor of IEEE/ACM Transactions on Audio, Speech and Language Processing.