Master Thesis - Wind Noise Metric for Speech Communication
Core
Develop a multi-score evaluation framework to assess wind-noise reduction in speech recordings for two-way communication, balancing noise suppression with speech preservation and artifact detection.
Role type
Research Master's Thesis Student (Signal Processing)
Builds
A combined scoring framework for comparing wind-noise reduction methods
Domain
Audio signal processing and acoustics
Deliverable
research
Required skills
Signal processing, Acoustics, Speech quality evaluation, Data collection and analysis, Statistical analysis
Preferred skills
DSP implementation, Subjective listening tests, Audio recording
Technologies
Digital Signal Processing (DSP), Audio recording equipment
Responsibilities
Investigate existing speech quality evaluation methods, Develop complementary scores for wind noise, residual noise, intelligibility, and artifacts, Record or collect speech samples under varying wind conditions, Compare metric scores with subjective listener assessments, Evaluate framework effectiveness in distinguishing wind suppression from speech damage
