>I have 2 serially connected multirate filters. One has 18 taps and the
>other has 61 taps. The decimation factors are 2, 2. Now I want to
combine
>the 2 filters into one. The new coefficient file has 78 taps by using
>convolution. In FDAtool, the order is 137 after cascading. I think it is
>from 137=17+60*2 . So I guess I cannot just use convolution to combine 2
>multirate filters.
>
>The problem is that the coefficients from FDAtool still shows in two
>sections, which stand for stage 1 (18 taps) and stage 2 (61 taps). Does
>anyone know how FDAtool expand the coefficients from 61 taps to 122
taps?
>Then I can do convolution on 18 and 122. This should be the key to solve
>the problem.
>
>Billy
>
>

Okay, I am here again to answer my question. I put zeros between the taps
and expanded the coefficient. It works well as I expected. I may need more
theory to support it.

I have 2 serially connected multirate filters. One has 18 taps and the
other has 61 taps. The decimation factors are 2, 2. Now I want to combine
the 2 filters into one. The new coefficient file has 78 taps by using
convolution. In FDAtool, the order is 137 after cascading. I think it is
from 137=17+60*2 . So I guess I cannot just use convolution to combine 2
multirate filters.

The problem is that the coefficients from FDAtool still shows in two
sections, which stand for stage 1 (18 taps) and stage 2 (61 taps). Does
anyone know how FDAtool expand the coefficients from 61 taps to 122 taps?
Then I can do convolution on 18 and 122. This should be the key to solve
the problem.

Billy