作者: P. Chordia , Gregoire Tronel , Ajay Srinivasamurthy , Sidharth Subramanian
DOI:
关键词:
摘要: In this paper, we propose a beat tracking and similarity based approach to rhythm description in Indian Classical Music. We present an algorithm that uses matrix inter onset interval histogram automatically extract the sub-beat structure long-term periodicity of musical piece. From information, can then obtain rank ordered set candidates for tāla cycle period naḍe (sub-beat structure). e tempo, locations along with provide beer overall is tested on manually annotated Carnatic music dataset (CMDB) light classical (ILCMDB). allowed metrical levels recognition accuracy ILCMDB 79.3% 72.4% tāla, respectively. difficult CMDB was poorer 68.6% 51.1% analysis algorithm's performance motivates us explore knowledge approaches recognition.