FFT real2complex vs complex2complex


Currently I’m using the FFT lib for out-of-place and real2complex transformation of a data with 1024*1024.

The needed time is about 1.45ms which is almost the half of the time the convolutionFFT2D sample needs (about 3ms just for FFT or IFFT) and I’m just wondering if there is really that much difference in speed of complex2complex vs real2complex or in-place vs out-of-place transformations.

Does someone have some experience what type of transformation is the fastest?