Hi again,
FP calculation is based on int which are 4*8 bits (aren't they ?) with the standard FP_Q = 10, it means each number is multiplicated by 1024 so the resolution do no exceed 1/1024 (not far from 0.001) but numbers can go up to 2^((4*8-1)-10)=2 097 151
if we compare, floats have a resolution of 0.0000001 (no?) wich is much better in many case
it leads to strange behaviours when we translate a code from floats to FP.
so this is my questions:
why FP_Q is set to 10 when it could be set to 16 or 20 to stick with floats resolution ?
as it is a Constant in overloaded.h we come back to the old problem: no more common overloaded.dll if we change it ;)
could we have a way to set a variable FP_Q value (could be set for exemple at 10 by default and changed by a function call if necessary) ?
and the main question: did i miss something ? (which could be more than 75% chance huhu)
