AnsweredAssumed Answered

IFFT in BF561 using DSP library functions.

Question asked by Anooj on Feb 4, 2010
Latest reply on Feb 24, 2010 by AndreasL



I am trying to implement IFFT algorithm. I have been using the DSP library funstions to implement.


I have used

twidfftrad2_fr16(twiddle, N_FFT); to find the twiddle factors and,



void ifft_fr16(const complex_fract16  input[],

               complex_fract16        output[],

               const complex_fract16  twiddle_table[],

               int                    twiddle_size,

               int                    fft_size,

               int                    *block_exponent

               int                    scale_method);   


to find the IFFT.


My input is complex data with range of real and imaginary values being -1, 0 or 1 (only these three values)....


The final output that I got doesn't match the output obtained in MATLAB for same set of inputs...


There also seems to be a problem while converting the values from 32 bit to 16bit...


What is the best way to convert the value from 32bit to 16 bit format....