weird behaviour of fft(real(S))
Reported by adhefe
BUG DESCRIPTION:
--------------
Consider the following code example:
s=zeros(100,1);
s(4)=1;
S=fft(s,-1);
C=cos(2*%pi*[0:99]'*3/100);
xset('window',0)
plot(real(S)-C) //It gives values smaller 1.e-14. It is just rounding errors which is fine.
//However,
rs=fft(real(S),-1)
xset('window',1)
plot(rs) // A wrong result is shown!
xset('window',2)
plot(C) // This shows what would be expected in the last plot!
//More,
is=fft(imag(S),-1) // It goes just fine.
What is wrong?
ERROR LOG:
----------
.....
HOW TO REPRODUCE THE BUG:
-------------------------
Just run the code above under scilab 5.5.2 final
OTHER INFORMATION:
------------------