The basic goal of muting is to remove horizontally moving energy. Such energy is unrelated to a deeper image. Typically muting is a weighting function zeroes data generally beyond some value of (g-s)/t. There is no question that muting removes much horizontally moving energy, but more can be done. Because of back-scattering, horizontally moving energy can often be found inside the mute zone. The way to get rid of it is to use a dip filter instead of a weighting function. Before modern high-density recording, slow moving noises were often aliased on the geophone cable, so dip filtering wasn't feasible. If the emergent angle isn't close enough to vertical, that is, if dt/dg isn't small enough, then the waves can't have come from the exploration target. On explosion data, filtering is not so easily applied in shot space as it is in geophone space because data is not very densely recorded in shot space. Don't fall into the trap of thinking that this dip filtering can be done on a common-midpoint gather. Back-scattered ground roll has no moveout on a common-midpoint gather.
Marine water-bottom scatter is frequently so strong that it is poorly suppressed by conventional processing. Earlier we saw the reason: point scatterers imply hyperbolic arrivals, which have steep dip, hence they have the stacking velocities of sediment rather than water. What is needed are two dip filters--one to reject waves leaving the shots at nonpenetrating angles, and the other to reject waves arriving at the geophones at nonpenetrating angles.
Present-day field arrays filter on the basis of spatial frequency kx. More high-frequency energy would be left in the data if the recording equipment used dip filters instead of spatial-frequency k filters. The causal dip filters described earlier might work nicely.