The present study addresses the impact of multimodal complexity in the transmission of a dance exercise. This study examined what kind of temporal correspondence favors the clarity of a teaching instruction in the context of a dance class. From an autoethnographic perspective, we describe a real learning situation in which a teacher marks (bodily demonstration and verbal counting) a movement exercise for a dancer. We analyze (microanalysis) the components of this marking, considering the model of the visual spatial indicator and the zero velocity and the prosodic components of the voice. It was hypothesized that when the emphasis of the movement and the emphasis of the voice are synchronized in phase this favors the clarity of the instruction. Seventy-five participants - dancers; musicians; non-musicians-non-dancers - evaluated dance exercises (audiovisual clips) with different modes of synchronization between voice and movement following two testing strategies. We found that in-phase exercise markings are significantly easier to follow and understand than when they are not. The work aims to show that a neglect in the multimodal organization of the instruction can generate distortion (noise) between the conceptualization and the perception of the message, putting at risk the result of the communication.