Rhythmic auditory stimulation (RAS) is a gait intervention in which gait-disordered patients synchronise footsteps to music or metronome cues. Musical 'groove', the tendency of music to induce movement, has previously been shown to be associated with faster gait, however, why groove affects gait remains unclear. One mechanism by which groove may affect gait is that of beat salience: music that is higher in groove has more salient musical beats, and higher beat salience might reduce the cognitive demands of perceiving the beat and synchronizing footsteps to it. If groove's effects on gait are driven primarily by the impact of beat salience on cognitive demands, then groove's effects might only be present in contexts in which it is relevant to reduce cognitive demands. Such contexts could include task parameters that increase cognitive demands (such as the requirement to synchronise to the beat), or individual differences that may make synchronisation more cognitively demanding. Here, we examined whether high beat salience can account for the effects of high-groove music on gait. First, we increased the beat salience of low-groove music to be similar to that of high-groove music by embedding metronome beats in low and high-groove music. We examined whether low-groove music with high beat salience elicited similar effects on gait as high-groove music. Second, we examined the effect of removing the requirement to synchronise footsteps to the beat (i.e., allowing participants to walk freely with the music), which is thought to remove the cognitive demand of synchronizing movements to the beat. We tested two populations thought to be sensitive to the cognitive demands of synchronisation, weak beat-perceivers and older adults. We found that increasing the beat salience of low-groove music increased stride velocity, but strides were still slower than with high-groove music. Similarly, removing the requirement to synchronise elicited faster, less variable gait, and reduced bias for stability, but high-groove music still elicited faster strides than low-groove music. These findings suggest that beat salience contributes to groove's effect on gait, but it does not fully account for it. Despite reducing task difficulty by equalizing beat salience and removing the requirement to synchronise, high-groove music still elicited faster, less variable gait. Therefore, other properties of groove also appear to play a role in groove's effect on gait.