The Nuts and Bolts of Music Source Separation - Machine Learning Techniques for Audio Modeling

Explore machine learning techniques for audio data modeling in this 14-minute conference talk from Data Science Conference Europe 2022, focusing on music source separation. Gain insights into multidisciplinary approaches, audio data modeling methods, masking techniques, and waveform models through practical examples. Learn about state-of-the-art methods including MobileNet implementation and discover essential digital signal processing concepts. Understand the best practices and practical tips for tackling audio source separation challenges while exploring the intersection of data science and computational art.