Speaker-Conditional Chain Model for Speech Separation and Extraction
This addresses speech separation and extraction for real scenarios with variable speaker numbers, though it appears incremental with hybrid methods.
The paper tackles the cocktail party problem by proposing a Speaker-Conditional Chain Model that first infers speaker identities and then uses them as conditions to extract speech sources, achieving comparable results on standard benchmarks and better adaptability for multi-round long recordings.
Speech separation has been extensively explored to tackle the cocktail party problem. However, these studies are still far from having enough generalization capabilities for real scenarios. In this work, we raise a common strategy named Speaker-Conditional Chain Model to process complex speech recordings. In the proposed method, our model first infers the identities of variable numbers of speakers from the observation based on a sequence-to-sequence model. Then, it takes the information from the inferred speakers as conditions to extract their speech sources. With the predicted speaker information from whole observation, our model is helpful to solve the problem of conventional speech separation and speaker extraction for multi-round long recordings. The experiments from standard fully-overlapped speech separation benchmarks show comparable results with prior studies, while our proposed model gets better adaptability for multi-round long recordings.