The experience of groove is associated with the urge to move to a musical rhythm. Here we focus on the relevance of audio features, obtained using music information retrieval (MIR) tools, for explaining the perception of groove and music-related movement. In Study 1 we extracted audio features from clips of real music previously rated on perceived groove. Measures of variability, such as the variance of the audio signal’s RMS curve and spectral flux (particularly in low frequencies), predicted groove ratings. Additionally, we dissociated two forms of event density, showing that an algorithm that emphasizes variability between beats predicted groove ratings better. In Study 2 we manipulated RMS levels and groove category (low, mid, and high groove) to confirm that perceived groove is not a function of loudness. In Study 3 we utilized novel music clips that manipulated the frequency of bass and bass drum (low vs. high) and attack time (short vs. long). Groove ratings and tapping velocities tended to be higher and tapping variability tended to be lower when the bass instruments had lower frequencies. The present findings emphasize the multifaceted nature of groove by linking audio and musical qualities to subjective experience and motor behavior.