On the Importance of Temporal Context in Proximity Kernels: A Vocal Separation Case Study
This work addresses a specific bottleneck in source separation for audio processing, representing an incremental improvement.
The paper tackled the problem of unreliable single-frame similarity metrics in musical source separation by incorporating temporal context into kernels, which led to a considerable improvement in separation quality for vocal separation.
Musical source separation methods exploit source-specific spectral characteristics to facilitate the decomposition process. Kernel Additive Modelling (KAM) models a source applying robust statistics to time-frequency bins as specified by a source-specific kernel, a function defining similarity between bins. Kernels in existing approaches are typically defined using metrics between single time frames. In the presence of noise and other sound sources information from a single-frame, however, turns out to be unreliable and often incorrect frames are selected as similar. In this paper, we incorporate a temporal context into the kernel to provide additional information stabilizing the similarity search. Evaluated in the context of vocal separation, our simple extension led to a considerable improvement in separation quality compared to previous kernels.