Digital Signal Processing Reference
In-Depth Information
x(k)
x(k)
x(k)
g M
g 2
g 1
v M (k)
v 2 (k)
v 1 (k)
+
+
+
M
2
1
y M (k)
y 2 (k)
y 1 (k)
STFT
STFT
STFT
Y M (ω)
Y 2 (ω)
Y 1 (ω)
H M (ω)
H 2 (ω)
H 1 (ω)
Inverse STFT
z(k)
FIG. 1.5 General block diagram of a DMA system for processing speech signals.
ferent way to gain a better understanding why DMAs can achieve maximal
directional gain.
1.2 The Paradigm for the Design and Implementation of
Differential Arrays
This topic is designed to address the challenging problem of extracting a de-
sired speech signal from interference and noise using DMA techniques. Differ-
ent from the traditional approach that constructs a general N th order DMA
in the time domain as illustrated in Fig. 1.4, we adopt a new and more general
paradigm, in most of this work, as shown in Fig. 1.5. This paradigm starts
with a clean speech signal x ( k ) (from a desired speaker) passing through some
Search WWH ::




Custom Search