2018/08/16 by Filip Korzeniowski, Gerhard Widmer, Korzeniowski, Filip +1
Arts and Humanities · Computer Science · Engineering · Psychology · #Artificial intelligence #Artificial neural network #Audio and Speech Processing (eess.AS) #Chord (peer-to-peer) #Cognitive science #Computer science #Computer vision #Diverse Musicological Studies #FOS: Computer and information sciences #FOS: Electrical engineering #Language model #Machine Learning (cs.LG) #Machine learning #Music Technology and Sound Studies #Music and Audio Processing #Natural language processing #Overconfidence effect #Psychology #Smoothing #Sound (cs.SD) #Speech recognition #cs.LG #cs.SD #eess.AS #electronic engineering #information engineering
paper · pdf · doi:10.48550/arxiv.1808.05341
published in arXiv (Cornell University) (Cornell University) · First published in the Proceedings of the 26th European Signal Processing Conference (EUSIPCO-2018) in 2018, published by EURASIP
arxiv created 2018/08/16 · openalex publication_date 2018/08/16 · arxiv updated 2018/08/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/08
Common temporal models for automatic chord recognition model chord changes on\na frame-wise basis. Due to this fact, they are unable to capture musical\nknowledge about chord progressions. In this paper, we propose a temporal model\nthat enables explicit modelling of chord changes and durations. We then apply\nN-gram models and a neural-network-based acoustic model within this framework,\nand evaluate the effect of model overconfidence. Our results show that model\noverconfidence plays only a minor role (but target smoothing still improves the\nacoustic model), and that stronger chord language models do improve recognition\nresults, however their effects are small compared to other domains.\n