I do not have much math background but part of the project I am working on requires the FFT of a single vector. The matlab function fft(x) works accurately for what I need,
In MATLAB, it looks like you're doing an fft of 16 real values {1+0i, 2+0i, 3+0i, etc...} whereas in Accelerate you're doing an fft of 8 complex values {1+2i, 3+4i, 5+6i, etc...}