| Paper: | SLP-L1.6 | ||
| Session: | Speech Coding for Network Applications | ||
| Time: | Tuesday, May 16, 12:10 - 12:30 | ||
| Presentation: | Lecture | ||
| Topic: | Speech and Spoken Language Processing: Speech Coding | ||
| Title: | PACKET LOSS CONCEALMENT WITH NATURAL VARIATIONS USING HMM | ||
| Authors: | Manohar N. Murthi, University of Miami, United States; Christoffer A. Rødbro, Søren Vang Andersen, Søren Holdt Jensen, Aalborg University, Denmark | ||
| Abstract: | Packet loss concealment (PLC) at a receiver has a substantial effect on the speech quality in Voice over IP. Most conventional PLC systems have largely relied upon variations of signal repetition and overlap-add interpolation which can produce speech signals that do not follow the larger overall statistical trends. In this paper, we demonstrate how Hidden Markov Models can be utilized to effect PLC based on statistical signal processing. In particular, we show how HMM based PLC yields conditional density functions that can be utilized by various statistical estimation methods that produce signal parameter estimates that produce more natural variation than conventional PLC methods, thereby providing much better speech quality. | ||
©2018 Conference Management Services, Inc. -||- email: webmaster@icassp2006.org -||- Last updated Friday, August 17, 2012