FIG. 18.

At MV energies, the MTFs measured with the slit and edge methods are equivalent: (a) 4 × oversampled LSFs derived from either the simulated AS1000 response to a 0.5 mm-thick Ta edge or to an the ideal slit (FLS simulation). A good match is achieved although high frequency noise, mostly beyond the Nyquist frequency of 1.25 mm⁻¹, is seen in the edge-generated LSF. Note that more than 500 000 CPU-hours of simulation time, which was prohibitively expensive, would have been required to match experimental fluence levels; (b) MTFs computed from the simulated LSFs match within the quantum noise-induced error of the edge simulation; (c) an excellent match is obtained in the MTFs of the Varian AS1000 imager measured at 6 MV using the slit (17 cGy exposure) and edge (100 cGy exposure) methods.