Hi,
I'm new to Hexagon development and have made a small program that computes the fft of a real signal but the output is not correct. The input is sinusoidal but the fft output has three spikes and I can't figure out the cause of this? Any advice appreciated!
Please find test code below
Thanks
// --------------- Simplified Cpp code ---------------
#define FFTSZ 256
CWord2x32 Input[FFTSZ] __attribute__((aligned(8*FFTSZ)));
CWord2x32 Output[FFTSZ] __attribute__((aligned(8)));
Word32 buffer[FFTSZ];
pFile_r = fopen("audio/input.bin", "rb");
pFile_w = fopen("audio/output.bin", "wb");
int Nfft = FFTSZ;
Word32 Z = 0;
result = fread(buffer, sizeof(typeof(buffer[0])), Nfft, pFile_r);
for(int i=0;i<Nfft;i++)
Input[i] = Q6_P_combine_RI(buffer[i], Z); // complex transform with real-input only
fft32x16(Input, Nfft, Wtwiddles, Output); // do transform
fwrite(Output, sizeof(typeof(Output[0])), Nfft, pFile_w);
// --------------- Matlab code ---------------
x= sin(2*pi*16*[0:255]/256); % no discontinuities
fid = fopen('../audio/input.bin', 'wb');
fwrite(fid, round(x*2^31), 'int32');
fclose(fid);
fid = fopen('../audio/output.bin', 'rb');
y = fread(fid, 'int32');
fclose(fid);
Y = (y(1:2:end)+1i*y(2:2:end))*2^-31;
subplot(313),plot(abs(Y));grid;
RESOLVED!!!!
Hi,
Is there some documentation on how to use the fft32x16? I have run the test (using make clean->make sim) which compares two fft realizations but that doesn't help me much with my problem. When I read the input (C:\Qualcomm\HEXAGON_Tools\5.0.14\examples\libcore\SigProc\cxFFT32x16\test\data) from Matlab and computes the fft the output does not resemble the output file.
This forum in general seems quite dead and I wonder if there is a better forum for this kind of question?