I am trying a simple example using frequency sampling. Therefore I have N=16 weights.
W=[w0 w1.....w15]
and I make then as follows
W=[1 1 1 1 1 0 0 0 0 0 0 0 1 1 1 1 ]
as magnitude and the phase is -k*pi*(n-1)/N where k is from 0 to N/2
giving a frequency sampling complex vector of
H=[h0 h1 h2 ....h15]
Now according to textbooks I have to have conjugate symmetry but exactly how does this work with the dc bit? The only way I could get it to work so I had real weight coefficients of the filter was to let the frequency coefficients be
h0=1 and then h15=conj(h1), h14=conj(h2), h13=conj(h3), h12=conj(h4).
Then my time-domain coefficients were real and it worked. I read that h15 should be the dc bit ie h15=h0 but this doesn't work.
I am trying a simple example using frequency sampling. Therefore I have N=16 weights.
W=[w0 w1.....w15]
and I make then as follows
W=[1 1 1 1 1 0 0 0 0 0 0 0 1 1 1 1 ]
as magnitude and the phase is -k*pi*(n-1)/N where k is from 0 to N/2
giving a frequency sampling complex vector of
H=[h0 h1 h2 ....h15]
Now according to textbooks I have to have conjugate symmetry but exactly how does this work with the dc bit? The only way I could get it to work so I had real weight coefficients of the filter was to let the frequency coefficients be
h0=1 and then h15=conj(h1), h14=conj(h2), h13=conj(h3), h12=conj(h4).
Then my time-domain coefficients were real and it worked. I read that h15 should be the dc bit ie h15=h0 but this doesn't work.
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 365 |
Nodes: | 16 (2 / 14) |
Uptime: | 09:27:31 |
Calls: | 7,763 |
Files: | 12,898 |
Messages: | 5,746,245 |