Technical discussions related to Analog Devices DSPs (including Blackfin, TigerSHARC, SHARC and ADSP-21xx DSPs).
|
Dear Dr. Mike, > Copy a program from the ADI web site and feed it fixed point data. > Not integer data. Scale your data so the fft will work correctly > - that's the main trick. Would you write me the link on ADI web site where i can find programs for adsp2188? I'am lookin for an FFT function for 16 bit signed integer real and imag time signals. What is the difference between the fixed point and the integer data? So, i have to scale the input data or the result? The fftN functions returns with the value of the bolck exponent. That means, i have to scale only the result of the fft? My main problem whit the fftN is that the values and the shape of the frquency spectra doesn't match the requested. May i ask you to help me to find an FFT function in assembly code for adsp2188 processor? Thanks in advance! Sandor |