作者: Hamish Allan
DOI:
关键词:
摘要: When human listeners tap their feet in time to music, they are performing beat tracking. Automatic tracking has been the subject much interest research community, but higher-level metre such as measure received less attention. This dissertation presents Metanome, which uses an autocorrelation technique try determine boundaries of a musical audio signal for (tactus) locations known. It achieves success rate 81% music with signatures corresponding metrical groupings two, further experimentation is required see if it can work effectively other signatures.