Teach DFT index k as cycles per recording

A student reading a |X[k]| plot treated k as a time sample and the bars as signal values. Saying that k counts full cycles in the whole recording fixed it at once: F = k / (recording seconds), with no formula to memorize. It also sets up the next probe: applying the rule to bins above N/2 gives frequencies the signal does not contain, which leads the student to mirror pairs and the Nyquist limit on their own.