Is it support Float Data Types in OpenMax DL?

Hello,

For the FFT, I think that there is no single precision floating point version – is this correct? In this case how to convert the input to S32 integers on the input and back to float on the output.

Thanks,

Regards,
Gaurang

Hi,

Yes you are correct that there is no floating point versions of OpenMAX DL.

You will have to scale your float values such that they don’t overflow the maximum integer values:

i = (int) f * scale

Regards,

Hedley