CITATION — REFERENCE ENTRY
Approximate Note Transcription for the Improved Identification of Difficult Chords — Proceedings of the 11th International Society for Music Information Retrieval Conference (ISMIR 2010)
- Key
- mauch2010-nnls-chroma
- Authors
- Mauch, Matthias; Dixon, Simon
- Issued
- 2010
- Type
- paper-conference
- Container
- Proceedings of the 11th International Society for Music Information Retrieval Conference (ISMIR 2010)
- Pages
- 135-140
Raw CSL JSON
{
"DOI": "10.5281/zenodo.1416598",
"URL": "https://archives.ismir.net/ismir2010/paper/000025.pdf",
"page": "135-140",
"type": "paper-conference",
"title": "Approximate Note Transcription for the Improved Identification of Difficult Chords",
"author": [
{
"given": "Matthias",
"family": "Mauch"
},
{
"given": "Simon",
"family": "Dixon"
}
],
"issued": {
"date-parts": [
[
2010
]
]
},
"event-place": "Utrecht, Netherlands",
"container-title": "Proceedings of the 11th International Society for Music Information Retrieval Conference (ISMIR 2010)"
}
Claims
-
In Mauch and Dixon's baseline, inverted chords were recognized around 40 percentage points worse than the same chord type in root position.
"performance for these chords is around 40 percentage points worse than for the same chord type in root position."
-
Mauch and Dixon performed an approximate note transcription with non-negative least squares (NNLS) before computing chroma features.
"We do so by performing a prior approximate transcription using an existing technique to solve non-negative least squares problems (NNLS)."
-
The partials of notes played in chords compromise correct recognition; the bass note's overtones in particular fall at the fundamental frequencies of other notes.
"The motivation for this is the observation that the partials of the notes played in chords compromise the correct recognition of chords. The bass note in particular usually has overtones at frequencies where other notes have their fundamental frequencies."
-
The approach achieved 80% accuracy on the MIREX 2009 song collection and metric, against the 2009 top result of 74%.
"We achieve very good results of 80% accuracy using the song collection and metric of the 2009 MIREX Chord Detection tasks. This is a significant increase over the top result (74%) in MIREX 2009."
Available in