This
model illustrates fixed-point computation
using the data-polymorphic FIR filter:
This
model exploits the fact that the
FIR filter actor is data polymorphic
to feed it fixed-point numbers and
fixed-point coefficients. The magnitude
transfer function is calculated
for three scenarios. The "exact"
scenario uses double-precision floating
point computations throughout. The
"fixed point" scenario
uses fixed-point computations, with
6 bits total, of which 2 bits are
integer bits. The "quantized
coefficients" scenario quantizes
the coefficients to 8 bits total,
of which 2 are integer bits, but
then performs all calculations using
double-precision floating point.
See the Block
Diagram definition.
