Melody Extraction From Polyphonic Music Signals Using Pitch Contour Characteristics

By Justin Salamon, Emilia G'omez. IEEE Transactions on Audio, Speech, and Language Processing, 2012

486 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Essentia: An Audio Analysis Library for Music Information Retrieval (2013, methodology): “• the pitch salience function of a signal, the estimation of the fundamental frequency of the predominant melody by the MELODIA algorithm [56], and the pitch estimation by the YinFFT method [11]; • the Harmonic Pitch-Class Profile (HPCP) of a spectrum (also called chroma features) [28], the tuning f”
Mellotron: Multispeaker Expressive Voice Synthesis by Conditioning on Rhythm, Pitch and Global Style Tokens (2019, unknown): “The pitch data is obtained by using Yin [10] or Melodia [15].”
MedleyDB: A Multitrack Dataset for Annotation-Intensive MIR Research (2014, methodology): “Performance of Melodia [12] on different subsets of MedleyDB (MDB) for type 1 melody annotations, and comparison to performance on the MIREX datasets.”
The 2018 Signal Separation Evaluation Campaign (2018, methodology): “They are: MELO as described in [24], as well as all the method implemented in NUSSL [15]: 2DFT [25], RPCA [9], REP1 [22], REP2 [20], HPSS [8].”
MIR_EVAL: A Transparent Implementation of Common MIR Metrics (2014, methodology): “We performed melody extraction using the “SG2” algorithm evaluated in 2011 [18] and compared mir_eval’s reported scores to those of MIREX.”