Automatic segmentation of acoustic musical signals using hidden Markov models

摘要: In this paper, we address an important step toward our goal of automatic musical accompaniment-the segmentation problem. Given a score to piece monophonic music and sampled recording performance that score, attempt segment the data into sequence contiguous regions corresponding notes rests in score. Within framework hidden Markov model, model prior knowledge, perform unsupervised learning parameters, compute globally minimizes posterior expected number errors. We also show how produce "online" estimates position. present examples experimental results, readers are encouraged access actual sound have made available from these experiments.