I saw that example, my approach is not that different from it.
I'm trying to read some values from mics and transfer them to frequency domain then later on do processing on the values and then go back to time domain and play them back to the output speakers.