Hi,
I don’t know if here is the right place to ask these questions. If I want to use the fixed-point arithmetic (for faster executions), what kind of numeric representation will be the best choice for implementing and IIR or FIR filter in DSP? Should I really take care of this matter in my programming? As far as I know multiplications for integer numbers are dangerous, so is it better to convert all of my numbers to fractional representations (assuming that the computations in this conversion are not important and they can be done before and then the result would be saved in the chip memory) ?
I would appreciate your help.
Regards,
Vala