21,942 Hits in 5.7 sec

Evaluating Low-Level Features for Beat Classification and Tracking

Fabien Gouyon, Simon Dixon, Gerhard Widmer
2007 2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07  
In this paper, we address the question of which low-level acoustical features are the most adequate for identifying music beats computationally.  ...  We compare two ways of evaluating features: their accuracy in a song-specific classification task (classifying beats vs nonbeats) and their performance as a front-end to a beat tracking system.  ...  Acknowledgments This research was partly funded by the projects S2S 2 and In-terfaces2Music. Thanks to Anssi Klapuri, Stephen Hainsworth, Giorgos Emmanouil, Matthew Davies and Juan Bello.  ... 
doi:10.1109/icassp.2007.367318 dblp:conf/icassp/GouyonDW07 fatcat:ewklyj4dwjf2tk54fsxoirntue

Beat Critic: Beat Tracking Octave Error Identification By Metrical Profile Analysis

Leigh M. Smith
2010 Zenodo  
Thanks are due to Geoffroy Peeters for provision of the beat-tracker and onset detection code.  ...  EVALUATION Two evaluation strategies for octave errors are possible: 1) evaluation of beat tracking, where the phase of the beat tracking is correct, but the beat frequency is twice the true rate and 2  ...  CONCLUSIONS A method for the detection of octave errors in beat tracking has been proposed and evaluated.  ... 
doi:10.5281/zenodo.1417891 fatcat:euzbxjuc3be6dng2de6jlv36fu

Enhancing downbeat detection when facing different music styles

Simon Durand, Bertrand David, Gael Richard
2014 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)  
We estimate the time signature by examining the similarity of frames at the beat level. The features are selected through a linear SVM model or a weighted sum.  ...  The whole system is evaluated on five different datasets of various musical styles and shows improvement over the state of the art.  ...  CONCLUSION Evaluation results show that using complementary high level musically inspired features is efficient for downbeat detection when facing different music styles.  ... 
doi:10.1109/icassp.2014.6854177 dblp:conf/icassp/DurandDR14 fatcat:tym5inplcbbcffzn5kj4wwgbia

Audio Signal Representations for Indexing in the Transform Domain

Emmanuel Ravelli, Gaël Richard, Laurent Daudet
2010 IEEE Transactions on Audio, Speech, and Language Processing  
We show that this new audio codec allows efficient transform-domain audio indexing for 3 different applications, namely beat tracking, chord recognition and musical genre classification.  ...  MDCT for AAC, or hybrid PQF/MDCT for MP3) have a sufficient time resolution for some rhythmic features, but a poor frequency resolution, which prevents their use in tonality-related applications.  ...  ACKNOWLEDGMENT The authors would like to thank Matthew Davies from QMUL and Juan P. Bello from NYU for providing their Matlab code.  ... 
doi:10.1109/tasl.2009.2025099 fatcat:56n3nzdzjzaqjjfisnckmzrw3y

Capturing the Temporal Domain in Echonest Features for Improved Classification Effectiveness [chapter]

Alexander Schindler, Andreas Rauber
2014 Lecture Notes in Computer Science  
, and can be effectively used for large scale music genre classification.  ...  We evaluate the performance on four traditional music genre classification test collections and compare them to state of the art audio descriptors.  ...  From these low-level features some mid-and high-level audio descriptors are derived (e.g. tempo, key, time signature, etc.).  ... 
doi:10.1007/978-3-319-12093-5_13 fatcat:wycyqeczwvgdxm3vy22bxuh4au

Multi-Task Learning of Tempo and Beat: Learning One to Improve the Other

Sebastian Böck, Matthew Davies, Peter Knees
2019 Zenodo  
In this paper, we propose a multi-task learning approach for simultaneous tempo estimation and beat tracking of musical audio.  ...  The multi-task learning is achieved by globally aggregating the skip connections of a beat tracking system built around temporal convolutional networks, and feeding them into a tempo classification layer  ...  for beat [4] and joint beat and downbeat tracking [6] .  ... 
doi:10.5281/zenodo.3527849 fatcat:w7a3sjlvord3zacq2w3qbqumrq

Using voice suppression algorithms to improve beat tracking in the presence of highly predominant vocals

Jose R. Zapata, Emilia Gomez
2013 2013 IEEE International Conference on Acoustics, Speech and Signal Processing  
Finally, we evaluate all the pairwise combinations between beat tracking and voice suppression methods.  ...  Then, we use seven state-of-the-art audio voice suppression techniques and a simple low pass filter to improve beat tracking estimations in the later case.  ...  and allow a better mid-level representation for beat tracking.  ... 
doi:10.1109/icassp.2013.6637607 dblp:conf/icassp/ZapataG13 fatcat:5mxeqofbhndwvpwwxum6ajx3mu

Content-Based Music Information Retrieval: Current Directions and Future Challenges

M.A. Casey, R. Veltkamp, M. Goto, M. Leman, C. Rhodes, M. Slaney
2008 Proceedings of the IEEE  
Some of the music collections available are approaching the scale of ten million tracks and this has posed a major challenge for searching, retrieving, and organizing music content.  ...  retrieval) and other cues (e.g., music notation and symbolic representation), and identifies some of the major challenges for the coming years.  ...  the other low-level features.  ... 
doi:10.1109/jproc.2008.916370 fatcat:ynpw7lyf6fchfdzzl5jll22cee

Cover song detection: From high scores to general classification

Suman Ravuri, Daniel P.W. Ellis
2010 2010 IEEE International Conference on Acoustics, Speech and Signal Processing  
The input to the system is a reference track and test track, and the output from the system is a binary classification of whether the reference/test pair is either from a reference/cover or reference/non-cover  ...  The system differs from state-of-the-art detectors by calculating multiple input features, performing a novel type of test song normalization in order to combat against "impostor" tracks, and performing  ...  We also experimented with mixing tempo levels (i.e. using 240 beats/minute for the reference track and 120 beats/minutes for the test track), but including these cross-tempos resulted in no performance  ... 
doi:10.1109/icassp.2010.5496214 dblp:conf/icassp/RavuriE10 fatcat:tyfmzodxmnfrripf5hapch4h64

Improving Rhythmic Similarity Computation By Beat Histogram Transformations

Matthias Gruhne, Christian Dittmar, Daniel Gärtner
2009 Zenodo  
An important pre-requisite for these search methods is the semantic classification, which requires suitable low-and mid-level features.  ...  These techniques require specialized classifiers and the beat histogram cannot be used as feature in conjunction with other low-level features.  ... 
doi:10.5281/zenodo.1417635 fatcat:kpiuo7dzqvbxxhsefw35fdvx6e

Joint Beat And Downbeat Tracking With Recurrent Neural Networks

Sebastian Böck, Florian Krebs, Gerhard Widmer
2016 Zenodo  
We use the recently published beat and downbeat annotations for the GTZAN dataset, the Klapuri, and the SMC set (built specifically to comprise hard-to-track musical pieces) for evaluation.  ...  Results & Discussion Since our system jointly tracks beats and downbeats, we compare with both downbeat and beat tracking algorithms. First of all, we evaluate on completely unseen data.  ... 
doi:10.5281/zenodo.1415836 fatcat:5u3avicz3bdtlgnw7umg7p55dm

Ibt: A Real-Time Tempo And Beat Tracking System

João Lobato Oliveira, Fabien Gouyon, Luis Gustavo Martins, Luís Paulo Reis
2010 Zenodo  
SYSTEM DESCRIPTION Audio Feature Extraction According to recent comparative studies evaluating alternative onset detection functions [5] and the accuracy of several low-level features applied to beat  ...  Although this paper does not provide experiments with respect to the usefulness of diverse low-level features as input to tracking beats [9] [2], it should be noted that a particularity of the proposed  ... 
doi:10.5281/zenodo.1416469 fatcat:i4m3seayyjck3fiolidzvjbdku

Audio Features Dedicated to the Detection of Four Basic Emotions [chapter]

Jacek Grekow
2015 Lecture Notes in Computer Science  
We examined the effect of low-level, rhythm and tonal features on the accuracy of the constructed classifiers.  ...  We selected features and found sets of features that were the most useful for detecting individual emotions.  ...  In the selected features, we have a representative of low-level (L), rhythm (R) and tonal (T) features.  ... 
doi:10.1007/978-3-319-24369-6_49 fatcat:rp7ozrfmhbdg7a3gg22xdhs3m4

In Search of Automatic Rhythm Analysis Methods for Turkish and Indian Art Music

Ajay Srinivasamurthy, André Holzapfel, Xavier Serra
2014 Journal of New Music Research  
We define and describe three relevant rhythm annotation tasks for these cultures -beat tracking, meter estimation, and downbeat detection.  ...  We then evaluate several methodologies from the state of the art in Music Information Retrieval (MIR) for these tasks, using manually annotated datasets of Turkish and Indian music.  ...  We evaluate beat tracking on the Turkish low-level-annotated dataset and the Carnatic low-level-annotated dataset introduced in Section 3.2, using KLA and ELL beat tracking algorithms.  ... 
doi:10.1080/09298215.2013.879902 fatcat:ehr432lz7bdvdpnuywtk542pe4

Audio-Based Music Classification With A Pretrained Convolutional Network

Sander Dieleman, Philemon Brakel, Benjamin Schrauwen
2011 Zenodo  
We then trained and evaluated the network as an MLP with backpropagation, for each of the classification tasks.  ...  Network Layout The input of the network consists of beat-aligned chroma and timbre features for a given track, so there are 24 input dimensions in total.  ... 
doi:10.5281/zenodo.1415187 fatcat:txgwoju3lbfbpc6eanolqdqmmi
« Previous Showing results 1 — 15 out of 21,942 results