Short description

Electrolaryngeal (EL) and other alaryngeal speech presents unique acoustic challenges due to its artificial voicing and reduced prosodic variation. At our lab we are working on providing more natural voice production for alaryngeal speech and alternatively on voice conversion to provide those speakers with a more natural sounding voice. Expressive speech production is difficult since the direct control of the excitation signal by the brain is broken. We therefore try to make use of biosignals such as heart rate, or skin conductance that have shown to correlate with emotions in the voice.

Your Tasks

• Review current research in correlations of biosignal and voicing properies
• Use machine learning systems to determine a certain emotional state of a speaker given biosignals
• Analyze latency, real-time factor, and robustness to EL-specific distortions
• Document the methodologies, experimental setups, and results

Your Profile/Prerequisites

• Strong interest in biomedical applications and speech technology
• Experience with Python is beneficial
• Familiarity with speech signal processing and speech communication

Contact:

• Martin Hagmüller (hagmueller@tugraz.at or 0316/873 4377)
• Benedikt Mayrhofer (benedikt.mayrhofer@tugraz.at)