Subspace and envelope subtraction algorithms for noise reduction in cochlear implants
Franklina M. B. Toledo, P.C. Loizou, Arthur P. Lobo · 2004
The performance of two noise reduction algorithms is evaluated using 14 subjects fitted with the Clarion S-Series and Clarion II implant devices. The first algorithm, based on signal subspace principles, is used for pre-processing sentences embedded in +5 dB noise. The second algorithm is based on the subtraction of the noisy speech envelopes from an estimate of the noise envelopes. The noise envelopes are estimated continuously using a variation of the minimum statistics algorithm. Results showed that the subspace algorithm produced significant improvements in sentence recognition scores compared to the subjects' daily strategy. Small improvements were also obtained for a few subjects with the envelope subtraction algorithm.