Can we estimate how much scattering might still be in the regions we have assigned to be outside the support constraint? Pierre Thibault and Veit Elser have done something along these lines.