I wonder if the little peak in the beginning comes from the crossover - mixed second and third order. Since I will be ordering parts for my Unitized waveguide crossover, I wil try to measure the individual responses again and set up a LeCleach crossover to see, if this could be improved even more.