Function w(r) is referred to as the lag window. 
Fig. 2.11. Lag windows w,(/ with A(n. 
25.3 Spectral Windows 
From Eq. 2.89, 
Net , 
§(@) = Se wOiR@ ene: 
r=—(N-1) 
As Eq. 2.66 in Section 2.5.1 shows, 
aw 
1 Nel ae. Hunde ; 
Pyw) = = RO es RG) — | Pre"? do, 
1 
r=—(N-1) 
sO inserting these relations into Eq. 2.89 gives 
a 
S@) = | Py) Ww-$¢) do, (2.90) 
IT 
where 
1 N-1 
W@)=— > wr) et. (2.91) 
2m r=—-(N-1) 
$(w) is obtained as a weighted integral of the periodogram, and weighting involves 
a smoothing operation in the neighborhood of @. This operation reduces the contribution 
30 
