--- Quote Start ---
@Kaz, I need only 16 bit at the final sum as I have a 16 bit dac so at the output of the polynomial I MUST have a 16 bit signal.
So as I have understood from your last post I do not need the leading zeros in the signals input of the sum and also at the output signal.. Am I right ?
Thank you for your time !
--- Quote End ---
you are right.
with only 16 bit resolution you are dealing with scaled down result(not absolute values).
A constant term must also be scaled down by 2^16 before entering addition and so you can just ignore it.