We present a new algorithm for lossy speech compression. The new algorithm is based on a simple technique for embedding a compressed sensing mechanism within a conventional Kalman filter. As such, it is capable of constructing compressed representations using significantly less samples than what is usually considered necessary.
By: Avishy Carmi; Dimitri Kanevsky; Bhuvana Ramabhadran
Published in: RC24814 in 2009
Questions about this service can be mailed to reports@us.ibm.com .
