I think the FFT precision is low, because float variables are used. When compared with Matlab, you will often get different results.. For my own solution I changed all float variables to double and got satisfying results. Hope this helps everyone who also encountered this problem !
↧