Previous research comparing musically trained and untrained individuals has yielded valuable insights into music cognition and behaviour. Here, we explore two aspects of musical engagement previously studied separately, auditory-visual correspondences and sensorimotor skills, in a novel real-time drawing paradigm. To that end, musically trained and untrained participants were presented with 18 short sequences of pure tones varying in pitch, loudness and tempo, as well as two short musical excerpts. Using an electronic graphics tablet, participants were asked to represent the sound stimuli visually by drawing along with them while they were played. Results revealed that the majority of participants represented pitch with height (higher on the tablet referring to higher pitches), and loudness with the thickness of the line (thicker line for louder sounds). However, musically untrained participants showed a greater diversity of representation strategies and tended to neglect pitch information if unchanged over time. Investigating the performance accuracy in a subgroup of participants revealed that, while pitch-height correspondences were generally represented more accurately than loudness–thickness correspondences, musically trained participants’ representations of pitch and loudness were more accurate. Results are discussed in terms of cross-modal correspondences, the perception of time, and sensorimotor skills.