BeatNet+: Real-time rhythm analysis for diverse music audio — Transactions of the International Society for Music Information Retrieval

Claims

1. challenges-and-method
   Real-time rhythm analysis faces three challenges: accuracy without access to future input, low latency (especially on low-powered devices), and generalization to music without strong percussive components. BeatNet+ trains an auxiliary branch on versions of the same music without drums, so the main network learns a representation less dependent on percussion, and adds adaptation strategies for isolated singing voices and non-percussive music.

   "Developing real‑time music rhythm analysis systems involves addressing three key challenges: The first challenge is maintaining high accuracy while not accessing future input data, as offline models do. The second challenge is achieving low latency, especially on low‑powered devices. [...] We argue that the third challenge is generalization to various kinds of music audio. [...] The main branch is fed with full music mixtures, while the auxiliary branch is fed with the full music of the same pieces less drum tracks [...]. Another innovation is on the adaptation strategies that help develop real‑time rhythm analysis models for challenging music scenarios, including isolated singing voices and non‑percussive music."

   Locator: section: Abstract, 1 · Quote language: en