1 paper
Jacob Kealey, John Hershey, François Grondin
Neural networks have recently become the dominant approach to sound separation. Their good performance relies on large datasets of isolated recordings. For speech and music, isolat…