Musicians’ body behaviour has a preponderant role in audience perception. We investigated how performers’ motion is perceived depending on the musical style and musical expertise. To further explore the effect of visual input, stimuli were presented in audio-only, audio–visual and visual-only conditions. We used motion and audio recordings of expert saxophone players playing two contrasting excerpts (positively and negatively valenced). For each excerpt, stimuli represented five motion degrees with increasing quantity of motion (QoM) and distinct predominant gestures. In the experiment (online and in-person), 384 participants rated performance recordings for expressiveness, professionalism and overall quality. Results revealed that, for the positively valenced excerpt, ratings increased as a function of QoM, whilst for the negatively valenced, the recording with predominant flap motion was favoured. Musicianship did not have a significant effect in motion perception. Concerning multisensory integration, both musicians and non-musicians presented visual dominance in the positively valenced excerpt, whereas in the negatively valenced, musicians shifted to auditory dominance. Our findings demonstrate that musical style not only determines the way observers perceive musicians’ movement as adequate, but also that it can promote changes in multisensory integration.