Improving Piano Music Transcription by Elman Dynamic Neural Networks
Identifieur interne : 000164 ( Main/Exploration ); précédent : 000163; suivant : 000165Improving Piano Music Transcription by Elman Dynamic Neural Networks
Auteurs : G. Costantini [Italie] ; M. Todisco [Italie] ; M. Carota [Italie]Source :
- Lecture Notes in Electrical Engineering [ 1876-1100 ] ; 2010.
Abstract
Abstract: In this paper, we present two methods based on neural networks for the automatic transcription of polyphonic piano music. The input to these methods consists in live piano music acquired by a microphone, while the pitch of all the notes in the corresponding score forms the output. The aim of this work is to compare the accuracy achieved using a feed-forward neural network, such as the MLP (MultiLayer Perceptron), with that supplied by a recurrent neural network, such as the ENN (Elman Neural Network). Signal processing techniques based on the CQT (Constant-Q Transform) are used in order to create a time-frequency representation of the input signals. The processing phases involve non-negative matrix factorization (NMF) for onset detection. Since large scale tests were required, the whole process (synthesis of audio data generated starting from MIDI files, comparison of the results with the original score) has been automated. Test, validation and training sets have been generated with reference to three different musical styles respectively represented by J. S. Bach’s inventions, F. Chopin’s nocturnes and C. Debussy’s preludes.
Url:
DOI: 10.1007/978-90-481-3606-3_78
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream Istex, to step Corpus: 000468
- to stream Istex, to step Curation: 000468
- to stream Istex, to step Checkpoint: 000109
- to stream Main, to step Merge: 000162
- to stream Main, to step Curation: 000164
Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title xml:lang="en">Improving Piano Music Transcription by Elman Dynamic Neural Networks</title><author><name sortKey="Costantini, G" sort="Costantini, G" uniqKey="Costantini G" first="G." last="Costantini">G. Costantini</name></author><author><name sortKey="Todisco, M" sort="Todisco, M" uniqKey="Todisco M" first="M." last="Todisco">M. Todisco</name></author><author><name sortKey="Carota, M" sort="Carota, M" uniqKey="Carota M" first="M." last="Carota">M. Carota</name></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:A8E4A6005013ADA58B3139070BA1E364D44F6DD2</idno><date when="2010" year="2010">2010</date><idno type="doi">10.1007/978-90-481-3606-3_78</idno><idno type="url">https://api.istex.fr/document/A8E4A6005013ADA58B3139070BA1E364D44F6DD2/fulltext/pdf</idno><idno type="wicri:Area/Istex/Corpus">000468</idno><idno type="wicri:explorRef" wicri:stream="Istex" wicri:step="Corpus" wicri:corpus="ISTEX">000468</idno><idno type="wicri:Area/Istex/Curation">000468</idno><idno type="wicri:Area/Istex/Checkpoint">000109</idno><idno type="wicri:explorRef" wicri:stream="Istex" wicri:step="Checkpoint">000109</idno><idno type="wicri:doubleKey">1876-1100:2010:Costantini G:improving:piano:music</idno><idno type="wicri:Area/Main/Merge">000162</idno><idno type="wicri:Area/Main/Curation">000164</idno><idno type="wicri:Area/Main/Exploration">000164</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">Improving Piano Music Transcription by Elman Dynamic Neural Networks</title><author><name sortKey="Costantini, G" sort="Costantini, G" uniqKey="Costantini G" first="G." last="Costantini">G. Costantini</name><affiliation wicri:level="3"><country xml:lang="fr">Italie</country><wicri:regionArea>Department of Electronic Engineering, University of Rome “Tor Vergata”, via del Politecnico, 1, 00133, Rome</wicri:regionArea><placeName><settlement type="city">Rome</settlement><region nuts="2">Latium</region></placeName></affiliation></author><author><name sortKey="Todisco, M" sort="Todisco, M" uniqKey="Todisco M" first="M." last="Todisco">M. Todisco</name><affiliation wicri:level="3"><country xml:lang="fr">Italie</country><wicri:regionArea>Institute of acoustics “O. M. Corbino”, via del Fosso del Cavaliere, 100, 00133, Rome</wicri:regionArea><placeName><settlement type="city">Rome</settlement><region nuts="2">Latium</region></placeName></affiliation></author><author><name sortKey="Carota, M" sort="Carota, M" uniqKey="Carota M" first="M." last="Carota">M. Carota</name><affiliation wicri:level="3"><country xml:lang="fr">Italie</country><wicri:regionArea>Department of Electronic Engineering, University of Rome “Tor Vergata”, via del Politecnico, 1, 00133, Rome</wicri:regionArea><placeName><settlement type="city">Rome</settlement><region nuts="2">Latium</region></placeName></affiliation></author></analytic><monogr></monogr><series><title level="s">Lecture Notes in Electrical Engineering</title><imprint><date>2010</date></imprint><idno type="ISSN">1876-1100</idno><idno type="eISSN">1876-1119</idno><idno type="ISSN">1876-1100</idno></series></biblStruct></sourceDesc><seriesStmt><idno type="ISSN">1876-1100</idno></seriesStmt></fileDesc><profileDesc><textClass></textClass><langUsage><language ident="en">en</language></langUsage></profileDesc></teiHeader><front><div type="abstract" xml:lang="en">Abstract: In this paper, we present two methods based on neural networks for the automatic transcription of polyphonic piano music. The input to these methods consists in live piano music acquired by a microphone, while the pitch of all the notes in the corresponding score forms the output. The aim of this work is to compare the accuracy achieved using a feed-forward neural network, such as the MLP (MultiLayer Perceptron), with that supplied by a recurrent neural network, such as the ENN (Elman Neural Network). Signal processing techniques based on the CQT (Constant-Q Transform) are used in order to create a time-frequency representation of the input signals. The processing phases involve non-negative matrix factorization (NMF) for onset detection. Since large scale tests were required, the whole process (synthesis of audio data generated starting from MIDI files, comparison of the results with the original score) has been automated. Test, validation and training sets have been generated with reference to three different musical styles respectively represented by J. S. Bach’s inventions, F. Chopin’s nocturnes and C. Debussy’s preludes.</div></front></TEI><affiliations><list><country><li>Italie</li></country><region><li>Latium</li></region><settlement><li>Rome</li></settlement></list><tree><country name="Italie"><region name="Latium"><name sortKey="Costantini, G" sort="Costantini, G" uniqKey="Costantini G" first="G." last="Costantini">G. Costantini</name></region><name sortKey="Carota, M" sort="Carota, M" uniqKey="Carota M" first="M." last="Carota">M. Carota</name><name sortKey="Todisco, M" sort="Todisco, M" uniqKey="Todisco M" first="M." last="Todisco">M. Todisco</name></country></tree></affiliations></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Musique/explor/DebussyV1/Data/Main/Exploration
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000164 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Exploration/biblio.hfd -nk 000164 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/Musique
|area= DebussyV1
|flux= Main
|étape= Exploration
|type= RBID
|clé= ISTEX:A8E4A6005013ADA58B3139070BA1E364D44F6DD2
|texte= Improving Piano Music Transcription by Elman Dynamic Neural Networks
}}
| This area was generated with Dilib version V0.6.33. |  |