I would like to use openmax dl library(optimized cortex-a instead of fftw library.I am using beagle board(cortex-a and I am developing my application in Linux.

For the FFT, I want to use single precision floating point and complex data types. In this case my input and output is float data.

I check Openmax DL library datatypes and FFT API,but I don't get any float datatypes in this library.

How to convert float to integer value without missing floating value?