<p>Sex differences in responses to social signals are traditionally attributed to downstream processes of receiver motivation and signal valuation. However, an alternative hypothesis suggests that such dimorphism originates earlier at the level of perceptual organization. Multimodal communication provides an ideal model to test these possibilities, as sensory streams must be bound before valuation. Here I examine sex-specific sensitivity to the temporal structure of audio-visual displays in the starling <i>Sturnus vulgaris</i>. Using a robotic taxidermic model in field conditions, I examined early reflexive reactions to distinguish how each sex binds song and wing movements into perceptual events. I manipulated the temporal relationship between these components while keeping the signals identical. The results reveal systematic sex differences in sensitivity to temporal shifts. While adding a visual component to an acoustic baseline triggered renewed responses in both sexes, only females showed dishabituation when the timing between familiar modalities was altered. Males failed to react to those shifts even at a low-cost reflexive level. These findings reveal systematic sex differences in sensitivity to audio–visual temporal structure that are evident even in early reflexive responses. I conclude that sex differences in social behaviour can arise before downstream motivational and decision-making processes, identifying early signal processing as an additional source of receiver specialization.</p>