Music Rhythm Matching Based on Dynamic Step Frequency

Youyang Du, Chi Zhang, Escoffier Wang, Yunsen Tang, Yonghui Xu, Lizhen Cui

Abstract: In recent years, more and more people choose to exercise through running, but running itself is a boring process, which is difficult to stick to. Although people can choose to listen to music while running to relieve the tedium of the running process, it is observed that if they just listen to music normally, they cannot play the real role of music in helping running. The existing way is to push songs with a similar rhythm to the user’s pace. This approach lacks flexibility and the user loses the power of choice. To solve the problems mentioned above, we propose a new method, the Timely Dynamic Adjustment Matching Method, which adjusts the music according to the user’s current cadence. In addition, we invent a data augmentation model to remove the user’s temporarily incorrect step frequency or measured error step frequency for some reason and adjust it when the user changes the existing step frequency. After the experiment, users running with the app using the timely dynamic adjustment matching method have improved the stability of step frequency, and user satisfaction has increased greatly.