Pupil plane images zoomed in around the shape of the probe applied on the DM. The optical path difference is normalized to the maximum of the probe. Top row, left: expected probe, namely the influence function of the actuator. Middle: shape of the probe in closed loop without considering optical gains. Right: same as middle but including an optical gain calibration. Bottom row: differences between the obtain probe in closed loop and the expected probe. Science case: bright-1, seeing = 1.1″, τ₀ = 3 ms.