Sliding spectra: a new perspective.
The sliding spectrum generated by using a fixed number of past data samples suffers from the tradeoff between spectral resolution and temporal resolution. To overcome this problem, an infinite impulse response filter is used in place of the finite length data window. By transforming the data into the spectrum domain and propagating the result through the filter's poles and zeros, an associate set of sliding spectra which vary in characteristics are generated. The sliding spectra with fine spectral resolution may be used to detail the power distribution over frequency during periods of stationarities while others can be examined for detection of transients in the signal statistics.
|Main Author:||Amin, Moeness G.|