Music transcription with ISA and HMM

Abstract : We propose a new generative model for polyphonic music based on nonlinear Independent Subspace Analysis (ISA) and factorial Hidden Markov Models (HMM). ISA represents chord spectra as sums of note power spectra and note spectra as sums of instrument-dependent log-power spectra. HMM models note duration. Instrument-dependent parameters are learnt on solo excerpts and used to transcribe musical recordings as collections of notes with time-varying power and other descriptive parameters such as vibrato. We prove the relevance of our modeling assumptions by comparing them with true data distributions and by giving satisfying transcriptions of two duo recordings.
Liste complète des métadonnées

Cited literature [12 references]  Display  Hide  Download

https://hal.inria.fr/inria-00544697
Contributor : Emmanuel Vincent <>
Submitted on : Wednesday, December 8, 2010 - 4:46:03 PM
Last modification on : Thursday, March 21, 2019 - 2:20:03 PM
Document(s) archivé(s) le : Thursday, March 10, 2011 - 12:33:33 PM

File

vincent_ICA04bis.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : inria-00544697, version 1

Collections

Citation

Emmanuel Vincent, Xavier Rodet. Music transcription with ISA and HMM. 5th Int. Conf. on Independent Component Analysis and Blind Signal Separation (ICA), Sep 2004, Granada, Spain. pp.1197--1204. ⟨inria-00544697⟩

Share

Metrics

Record views

112

Files downloads

199