Separate this, and all of these Things Around It: Music Source Separation Via Hyperellipsoidal Queries

Music source separation (MSS) is a task of extracting one or more constituent components, or composites thereof, from a musical audio mixture. Historically, music source separation has been dominated by a stem-based paradigm, leading to most systems being either a collection of single-stem extraction models, or a tightly coupled system with a fixed set of supported stems. Combined with the limited data availability, advances in music source separation have thus been mostly limited to the "VDBO" set of stems: vocals, drum, bass, and the catch-all others. Recent MSS works have begun to challenge the fixed-stem paradigm, moving towards models able to extract any musical sound as long as this target type of sound could be specified to the model as an additional query input. We generalize this idea to a query-by-region system, specifying the target based on the query, regardless of how many sound sources or which sound classes are contained within it. To do so, we use hyperellipsoidal regions as queries to allow for an intuitive yet easily parametrizable approach to specifying both the target (location) and its spread. Evaluation of the proposed system on the MoisesDB dataset demonstrated near state-of-the-art performance of the proposed system both in terms of signal-to-noise ratios and retrieval metrics.