Contrast iterations in simulation, using EFC to calculate DM commands. The data sets were taken sequentially with single actuators (blue), classic sinc probes (orange), and sharp sinc probes (green), all using three pairs in the PW estimation. The EFC regularization was held constant across all iterations. The loop was run at a wavelength of λ = 783 nm and with a loop gain of 0.8, and the mean contrast was measured in a circular DH spanning 3–9 λ/D. All three shapes of probes were applied with a probe amplitude that adds a mean contrast of 10⁻⁶ to the DH, in each iteration. The single-actuator probes provide a slightly better contrast than the classic sinc probes at higher iteration numbers (after iteration ~45), while the sharp sinc probes start performing better sooner (after iteration ~15). The loop does not converge in the data shown here, but it reaches a level below 10⁻⁸, which is lower than the comparison hardware data shown in Fig. 10. The simulations are noiseless.