Probabilistic Binary-Mask Cocktail-Party Source Separation in a Convolutional Deep Neural Network
This work addresses speech separation for audio processing applications, but it is incremental as it applies a known deep learning approach to a well-established benchmark.
The paper tackled the problem of separating two competing speech sources in a cocktail party scenario by training a convolutional deep neural network to predict probabilistic binary masks, achieving performance close to the ideal binary mask benchmark.
Separation of competing speech is a key challenge in signal processing and a feat routinely performed by the human auditory brain. A long standing benchmark of the spectrogram approach to source separation is known as the ideal binary mask. Here, we train a convolutional deep neural network, on a two-speaker cocktail party problem, to make probabilistic predictions about binary masks. Our results approach ideal binary mask performance, illustrating that relatively simple deep neural networks are capable of robust binary mask prediction. We also illustrate the trade-off between prediction statistics and separation quality.