Had not thought of that. But that will give me one count per bar that goes past the encoder, yet there are four transitions that could potentially be counted. Of course, the pulses are not precisely symmetrical, so the four sub-counts are potentially not the same size.