how to do scaling to use float values for Inverse SP functio


I am trying to use FFTInv_CToC_SC32_Sfs Signal Processing function in my application .

I want to use float values as input values to this function like 0.000012, 0.001953,0.002049,0.003508.
I am trying to use conversion factor as 32768(i.e. 2^15) to convert input values like
inp = (OMX_S32) (0.000012 * (1 << 15)); here inp is OMX_S32 variable.
Like this I am also converting output value of FFTInv using conversion factor like…

out = (float)out_FFTInv / 32768 ;

By this way I am not getting proper output values.

how to do scaling of float values so I can use that as OMX integer and get proper output from FFTInv function ?

Rachit Shah