All assume far field behavior as sources are assumed to be 'point' sources - unless you want to try to model sources as driver arrays - and the exact FER modeling of the acoustical impedance of surfaces is still evolving and have limitations for small spaces, especially the near field.
I prefer EASE, but then that is what I am most familiar, while friends also use CATT-A.
Both CATT-A and EASE are fine programs that provide quite a bit of functionality - albeit for a non-trivial price.
