Interspeech'2005 - Eurospeech
In this paper we present of a statistical framework to demodulate speech resonances, which are modeled as AM-FM signals. The first approach utilizes bandpass filtering and a standard demodulation algorithm which regularizes instantaneous amplitude and frequency estimates. The second employs particle filtering techniques to allow temporal variations of the parameters that are connected with spectral characteristics of the analyzed signal. Results are presented on both synthetic and real speech signals and improved performance is demonstrated. Both approaches appear to cope quite satisfactorily with the nonstationarity of speech signals.
Bibliographic reference. Katsamanis, Athanassios / Maragos, Petros (2005): "Advances in statistical estimation and tracking of AM-FM speech components", In INTERSPEECH-2005, 1125-1128.