Predicted data with a virtual source located at the middle of the model with a single-layered overburden after 20 iterations. (a) Conventional double-focusing and (b) Marchenko double-focusing. Weak multiple reflections that are predicted by our method can be seen inside the red box. The maximum and the minimum value of the grey-level scale of both figures are the same and a linear time-varying (⁠|$\frac{t}{{\rm d}t}$|⁠) gain is applied to both records to amplify weaker events.