Beside using floating point over integers is there any difference between the two fixedresampling functions if your imput samples are integers ?
(ippsResamplePolyphaseFixed_16s / ippsResamplePolyphaseFixed_32f)
Is the floating point more accurate ?
Isone of them faster than the other without optimisation ?
Doesone of themtake more advantages from SSE2 optimization than the other ?
For more complete information about compiler optimizations, see our Optimization Notice.