Filters








805 Hits in 5.1 sec

Deep rank-based transposition-invariant distances on musical sequences [article]

Gaëtan Hadjeres, Frank Nielsen
2017 arXiv   pre-print
This paper introduces a method to build transposition-invariant distances on symbolic musical sequences which are learned from data.  ...  A distance could also be invariant under some transformations, such as transpositions, so that it can be used as a distance between musical motives rather than musical sequences.  ...  The idea is to define a distance based solely on the ranking of the high-level activations of a Deep Neural Network (DNN) [13] .  ... 
arXiv:1709.00740v1 fatcat:6g3rsgdcmvgjlcijvi64ugjh3i

Audio-based Musical Version Identification: Elements and Challenges [article]

Furkan Yesiler, Guillaume Doras, Rachel M. Bittner, Christopher J. Tralie, Joan Serrà
2021 arXiv   pre-print
Although this trend positively influences the number of researchers and institutions working on VI, it may also result in obscuring the literature before the deep learning era.  ...  Recent years, however, have witnessed the rise of deep learning-based approaches that take a step toward bridging the accuracy-scalability gap, yielding systems that can realistically be deployed in industrial  ...  error Data projection Classification- based training Conventional similarity measures Similarity-based training Similarity estimation For structure invariance For transposition invariance  ... 
arXiv:2109.02472v1 fatcat:tbbd66yq2vcz3ahc4z5ymethgi

Pairwise Similarity Normalization Based on a Hubness Score for Improving Cover Song Retrieval Accuracy

Jin S. SEO
2022 IEICE transactions on information and systems  
A hubness-score based normalization of the pairwise similarity is proposed for the sequence-alignment based cover song retrieval.  ...  Experiments on two cover song datasets confirm that the proposed similarity normalization improves the cover song retrieval accuracy.  ...  In this paper, Fig. 1 The music-similarity computation for the cover song identification based on the optimal transposition index and sequence alignment.  ... 
doi:10.1587/transinf.2021edl8075 fatcat:np4m3hvqwbespkxjj5bcwps4du

Adaption of String Matching Algorithms for Identification of Near-Duplicate Music Documents

Matthias Robine, Pierre Hanna, Pascal Ferraro, Julien Allali
2007 Annual International ACM SIGIR Conference on Research and Development in Information Retrieval  
The main part of the existing systems for the comparison of symbolic music are based on string matching algorithms and represent music as sequences of notes.  ...  The number of copyright registrations for music documents is increasing each year. Computer-based systems may help to detect near-duplicate music documents and plagiarisms.  ...  transposition invariant.  ... 
dblp:conf/sigir/RobineHFA07 fatcat:cffqgg5lozblpppo2yirusrjom

Temporal Pyramid Pooling Convolutional Neural Network for Cover Song Identification

Zhesong Yu, Xiaoshuo Xu, Xiaoou Chen, Deshun Yang
2019 Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence  
Most existing methods rely on hand-crafted features and sequence alignment methods, and further breakthrough is hard to achieve.  ...  We show that they could be naturally adapted to deal with key transposition in cover songs.  ...  In some sense, our model could be used for content-based music recommendation.  ... 
doi:10.24963/ijcai.2019/673 dblp:conf/ijcai/YuXCY19 fatcat:25yzxoepj5fapc7ory3h6ajo7y

Using weakly aligned score–audio pairs to train deep chroma models for cross-modal music retrieval

Frank Zalkow, Meinard Müller
2020 Zenodo  
A typical strategy is to compare score–audio pairs based on a common mid-level representation, such as chroma features.  ...  We present systematic experiments that show the effectiveness of the CTC-based model for this theme-based retrieval task.  ...  We thank Daniel Stoller for fruitful discussions on the CTC loss, and Michael Krause for proof-reading the manuscript.  ... 
doi:10.5281/zenodo.4245399 fatcat:rtla4xknznb7fbgl4hgsf4itqa

Accurate and Scalable Version Identification Using Musically-Motivated Embeddings [article]

Furkan Yesiler, Joan Serrà, Emilia Gómez
2020 arXiv   pre-print
MOVE achieves state-of-the-art performance on two publicly-available benchmark sets by learning scalable embeddings in an Euclidean distance space, using a triplet loss and a hard triplet mining strategy  ...  In this paper, we present MOVE, a musically-motivated method for accurate and scalable version identification.  ...  Based on the modifiable musical characteristics specified in Section 1 and elsewhere, such function sequentially and independently applies transposition in the pitch dimension, time stretching, and time  ... 
arXiv:1910.12551v2 fatcat:mvz3bg4anjfipdse7a6p5ywflm

Isomorphic Controllers and Dynamic Tuning: Invariant Fingering over a Tuning Continuum

Andrew Milne, William Sethares, James Plamondon
2007 Computer Music Journal  
This means that transpositional invariance on a button-row is only possible for a rank-one (i.e., equal) temperament.  ...  Musical context is also important; the tuning range of invariant rational identity may also change based on the spectrum and/or timbre of the sounds (Sethares 2004) .  ...  A layout is transpositionally invariant if every fixed interval I is fingered in the same manner, i.e., if Thus, transpositional invariance requires that the difference in locations between notes on the  ... 
doi:10.1162/comj.2007.31.4.15 fatcat:pyurawmkunfybi5atdwpptyn4q

Learning a Representation for Cover Song Identification Using Convolutional Neural Network [article]

Zhesong Yu, Xiaoshuo Xu, Xiaoou Chen, Deshun Yang
2019 arXiv   pre-print
In this paper, we propose a novel Convolutional Neural Network (CNN) architecture based on the characteristics of the cover song task.  ...  Cover song identification represents a challenging task in the field of Music Information Retrieval (MIR) due to complex musical variations between query tracks and cover versions.  ...  In this sense, our approach may also be used to retrieve similar music of the query and extended to content-based music recommendation.  ... 
arXiv:1911.00334v1 fatcat:jfago6nx4fbihakq66746edrxq

Summarizing and Comparing Music Data and Its Application on Cover Song Identification

Diego Furtado Silva, Felipe Falcão, Nazareno Andrade
2018 Zenodo  
In this paper, we propose a novel method to summarize a recording in small snippets based on its self-similarity information.  ...  While there is a multitude of music information retrieval algorithms that have distance functions as their core procedure, comparing the similarity between recordings is a costly procedure.  ...  To deal with this variation, several methods provide invariance to these issues to CSI algorithms. One example is the Optimal Transposition Index (OTI) [20] , which provides key invariance.  ... 
doi:10.5281/zenodo.1492521 fatcat:cdpd6xd6wfhsllmi4kmkoztzge

CTC-Based Learning of Chroma Features for Score–Audio Music Retrieval

Frank Zalkow, Meinard Muller
2021 IEEE/ACM Transactions on Audio Speech and Language Processing  
Strategies for comparing score and audio data are often based on a common mid-level representation, such as chroma features, which capture melodic and harmonic properties.  ...  on the features.  ...  In Section II, we review related work on score-audio retrieval, deep salience and deep chroma models, as well as on musical applications of the CTC loss.  ... 
doi:10.1109/taslp.2021.3110137 fatcat:ozokra6olbh7xfz6hnv3yjstx4

Known Artist Live Song Id: A Hashprint Approach

T. J. Tsai, Thomas Prätzlich, Meinard Müller
2016 Zenodo  
on deep neural networks).  ...  and width of the filters are designed to match the pitches on the Western musical scale, so the representation is especially suitable for considering key transpositions.  ... 
doi:10.5281/zenodo.1418222 fatcat:xfo6s5mewzherpeimfmw4daezu

Data Driven And Discriminative Projections For Large-Scale Cover Song Identification

Eric J. Humphrey, Oriol Nieto, Juan Pablo Bello
2013 Zenodo  
Intuitively, cover song retrieval algorithms are designed to be invariant to time and key transpositions.  ...  a deep network via backpropagation.  ... 
doi:10.5281/zenodo.1416547 fatcat:nvpf2jilwvaanf3sz5owklcuvq

Graph based representation of the music symbolic level. A music information retrieval application

Federico Simonetta, Antonio Rodà
2018 Zenodo  
In this work, a new music symbolic level representation system is described.  ...  It has been tested in two information retrieval tasks concerning similarity between segments of music and genre detection of a given segment.  ...  this measure transposition invariant.  ... 
doi:10.5281/zenodo.1476564 fatcat:mzk7lhqao5gbte5jo6hvlhbc3a

Computational geometric aspects of rhythm, melody, and voice-leading

Godfried Toussaint
2010 Computational geometry  
mathematics and computer science to music theory, music perception, and musicology.  ...  Recent results on the geometric and computational aspects of rhythm, melody, and voice-leading are reviewed, connections to established areas of computer science, mathematics, statistics, computational  ...  The transposition-distance (also called Cayley distance) between two sequences is the minimum number of transpositions needed to convert one sequence to the other.  ... 
doi:10.1016/j.comgeo.2007.01.003 fatcat:7uihjaxpgjcg7pipdcy3lre3bi
« Previous Showing results 1 — 15 out of 805 results