A number of studies have investigated whether cross-modal correspondence effect occurs in purely automatic manner or whether top-down processes can be involved in the processing. The current study addresses the disparity in the research conducting two experiments, using a classical audiovisual cross-modal correspondence paradigm and testing possible involvement of the endogenous component in the effect. Experiment 1 replicated previous findings and showed presence of cross-modal correspondence between pitch and spatial position. However, the effect was significant only in upper spatial position. Experiment 2 showed that task-related pitch probability manipulation made the cross-modal correspondence effect to disappear, however revealing an asymmetrical pattern that was highly dependent on pitch probability and spatial position. Overall, the results suggested a non-automaticity of the cross-modal correspondence effect and a possible involvement of endogenous component in the effect.