Parametric Time-Scale Methods in Signal Analysis
Abstract
The researchers have separated the Kaiser-Teager algorithm for separating the contributions of the amplitude modulation and frequency modulation of a signal consisting of multiple components. To achieve this separation, they use an instantaneous non-linear operator, which turns out to be the determinant of a Toeplitz matrix formed with the signal samples. Because of its instantaneously adaptive nature, they can use this algorithm to track parameter variations are not too rapid. The researchers demonstrate this using a synthetic signal containing two AM-FI4 signal components and a speech signal. They also point out the methods relationship to Prony's method. The researchers proposed a demodulator-filter approach for tracking harmonic signals and applied it to speech processing.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 30, 1993
- Accession Number
- ADA274212
Entities
People
- R. Kumaresan
Organizations
- University of Rhode Island