Transforming audio from time-domain to frequency-domain and back is not trivial, because it involves:
The Superpowered FrequencyDomain feature solves all with a very simple API call: time-domain audio goes in, magnitudes and phases come out (or complex numbers if you wish). Or vice-versa. No need to dig deep into FFT if that's not your bailiwick.
It is a singular audio DSP feature, and also the fastest solution with the highest performance on mobile devices. It is, at minimum, 2.5 times faster than an Apple vDSP implementation.
The Superpowered Audio SDK has a simple frequency domain example project for both iOS and Android. You can work with magnitudes and phases within minutes, and implement your own linear phase EQ for example!
Superpowered FrequencyDomain has an interesting twist too: you can define the value of Pi. You can leave it at 3.14xxx, but some algorithms that work with phases greatly benefit if Pi is translated to something else, such as 1.0 or 0.5.