Our cognitive system tends to link auditory pitch with spatial location in a specific manner (ie high-pitched sounds are usually associated with an upper location, and low sounds are associated with a lower location). Recent studies have demonstrated that this cross-modality association biases the allocation of visual attention and affects performance despite the auditory stimuli being irrelevant to the behavioural task. There is, however, a discrepancy between studies in their interpretation of the underlying mechanisms. Whereas we have previously claimed that the pitch-location mapping is mediated by volitional shifts of attention (Chiou & Rich, 2012, Perception, 41, 339–353), other researchers suggest that this cross-modal effect reflects automatic shifts of attention (Mossbridge, Grabowecky, & Suzuki, 2011, Cognition, 121, 133–139). Here we report a series of three experiments examining the effects of perceptual and response-related pressure on the ability of nonpredictive pitch to bias visual attention. We compare it with two control cues: a predictive pitch that triggers voluntary attention shifts and a salient peripheral flash that evokes involuntary shifts. The results show that the effect of nonpredictive pitch is abolished by pressure at either perceptual or response levels. By contrast, the effects of the two control cues remain significant, demonstrating the robustness of informative and perceptually salient stimuli in directing attention. This distinction suggests that, in contexts of high perceptual demand and response pressure, cognitive resources are primarily engaged by the task-relevant stimuli, which effectively prevents uninformative pitch from orienting attention to its cross-modally associated location. These findings are consistent with the hypothesis that the link between pitch and location affects attentional deployment via volitional rather than automatic mechanisms.