It is well known that a sound coming from a particular location can capture visual attention, as previously shown in crossmodal cueing studies. The current study shows that when performing a task involving the discrimination of a visual target, an irrelevant auditory cue near the target location results in performance benefits relative to a neutral condition. In contrast, when the irrelevant auditory cue comes from a non-target location there are performance costs. We have determined whether it is possible to suppress such auditory capture by endogenously focusing visual attention to a restricted area in space. Previous studies have shown that endogenously focusing attention can eliminate capture by visual stimuli. Unlike these studies, the current study shows that even when attention is highly visually focused, auditory stimuli still capture visual attention. However, in this case there are only performance costs from auditory stimuli coming from the non-target location; cueing benefits of a sound presented near the visual target are no longer present.