Some questions about the arm_rfft_q15

[LLily.1]

I was trying to use arm_rfft_q15  .In the CMSIS documentation for arm_rfft_q15, it says the output format is Q7.9 and 6 bits should be upscaled when the input RFFT size is 128 . What is meant by this? Do I have to get the float values by dividing the output array by 512(2^9)? But, the results were wrong.

[Ozone]

Q7.9 means 7 bit integer, and 9 fractional bits.

The idea is to use this number format - and functions - for appropriate data (like 12 bit ADC values).

This makes sense for smaller/slower cores.

Converting from / to float (24 bit mantissa + 8 bit exponent) does not make sense in this context.