Authors: Danilo Pani Alessandro Pani Luigi Raffo
Publish Date: 2014/07/03
Volume: 70, Issue: 3, Pages: 1555-1576
Abstract
Many online blind audio source separation BASS algorithms have been presented so far to the scientific community but only a few of them have been evaluated in terms of their realtime performance In this paper we consider a wellestablished BASS method oriented to voices separation evaluating its performance in terms of separation quality allowed by a realtime embedded computing implementation also considering novel and stateoftheart improvements to the it To this aim the algorithm has been implemented and ported for realtime execution onto an advanced lowpower digital signal processor targeted for complexdomain applications The optimized embedded implementation is able to perform up to five iterations of the gradient for any input frame of data achieving good separation levels up to 118 dB of signal to interference ratio on custom recording in real environments The proposed solution doubles the performance of a Coptimized version running on a traditional PC processor achieving a better separation result with lower power requirements
Keywords: