Digital Signal Processing Reference
In-Depth Information
power
.
.
.
.
.
.
.
.
.
.
128 time-sample signal
power
.
.
.
.
.
.
.
.
.
.
64 time-sample signal
power
.
.
.
.
.
.
.
.
.
.
16 time-sample signal
Figure 7.1.
Comparison of the simple periodogram weighted by a Hanning window
(left column) and the minimum variance method (right column) for the estimation of the
power spectral density of a sinusoid (0.25 Hz, sampling frequency 1 Hz), not very noisy
(20 dB), on a time support of 128 samples, 64 samples and 16 samples. Horizontal axes:
frequency. Vertical axes: spectrum in dB
Experimental comparisons with other algorithms are developed on non-
stationary signals in [BON 90, LAT 87]. Synthesis sheets were drawn up in order to
briefly present a set of time-frequency approaches for the MV estimator [BAS 92].
See also [FER 86] for an adaptive MV estimator. As regards the applications, the
reader will be able to consult [PAD 95, PAD 96] for an application in mechanics,
[LEP 98] in seismics on the characterization of avalanche signals and [MAR 95] in
room acoustics for an original adaptation in the time-octave domain.
Search WWH ::
Custom Search