Forums

Scaling of fir polyphase filters

Started by mitrobge June 26, 2008
In a FIR filter the scaling of the input in order to avoid overflow at the
output can be specified by using the 1st norm of the vector of
coefficients.

ymax=xmax*scaling*norm_1
for xmax=ymax=1
scaling=1/norm_1

How to use this relation in a multistage interpolator with polyphase
decomposition?

Thanks


On Jun 26, 7:19 am, "mitrobge" <dimitris.mitrovge...@gmail.com> wrote:
> In a FIR filter the scaling of the input in order to avoid overflow at the > output can be specified by using the 1st norm of the vector of > coefficients.
> How to use this relation in a multistage interpolator with polyphase > decomposition?
Think through the operation in the context of the conditions would create the largest values, either at the output or at some intermediate point.