I had been breaking my head over convolution for the past one week, and last night there was the revelation I had been waiting for. After that moment of truth, I no longer know why it took so much time-what was there not to understand?
Anyway, once I had nailed convolution, I figured out a neat way to implement circular convolution in MATLAB. For convolution of 2 sequences of length n, this will give the desired result:
out=out + input1(i)*input2;
end %end of for building convolved sequence.
Figuring that out kind of made my day today. This is what Richard Feynman called “The Pleasure of Finding Things Out”.