 We need to decrease the resolution by one bit. Let's change it to 15-bits.
Save the coefficients again.
When you open the file, you may find the resolution decreased and all values
stay inside the range of signed 16-bits (-32768 to +32767).
That's it for this tutorial. The absolute value of IIR filter's coefficients
may exceed 1.0. If you implement them in some microprocessors, you may
use 16-bits variables, like "short" in C language. In that case,
you need to decrease the resolution to quantize coefficients.
|