Transcribing Debussy's Syrinx dynamics through Linguistic Description

摘要: Advances in computational power have enabled the manipulation of music audio files and emergence Music Information Retrieval field. One main research lines this area is that Transcription, which aims at transforming an file into musical notation. So far, most efforts focused on accurately transcribing pitch durations notes, thus neglecting other aspects score. The present work explores a novel line action context automatic transcription, focusing dynamics, by means Linguistic Description. process described paper (called MUDELD: MUsic Dynamics Extraction through Description) departs from data series representing file, requires segmentation piece phrases, currently done hand. Initial experiments been performed eight recordings Debussy's Syrinx with promising results.