Abstract by Jared Miller
Physics and Astronomy
Evaluating acoustic simulation fidelity through machine learning
There has been a longstanding tradeoff in evaluating the quality or fidelity of sound recordings: subjective listening tests are time consuming and expensive, but objective measures often fail to capture the nuances of human perception. The research presented here seeks to address this problem by investigating the use of machine learning to evaluate the fidelity of acoustic simulations. To begin, we created a dataset of recordings representing varying levels of audio fidelity. Participants listened to each of the recordings and subjectively classified the perceived fidelity. Common sound quality metrics were calculated from each of the recordings: loudness, sharpness, roughness, and fluctuation strength. These sound quality metrics were used as inputs for various machine learning algorithms to test which best modeled the human classifications. A logistic regression model was determined to be the most advantageous, dependent on using binary classification. Introducing a reference sound, and calculating each sound quality metric relative to the reference, significantly improved accuracy.