375 Hits in 3.9 sec

MIR Evaluation Practices [article]

Marius Miron, Lorenzo Porcaro
2021 Zenodo  
Slides for the lesson on "MIR Evaluation Practices", prepared for the Music Information Retrieval course of the Master in Sound and Music Computing organized by the Music Technology Group (Universitat Pompeu Fabra) (2021-03-8/9)
doi:10.5281/zenodo.4611731 fatcat:t6hvgvqc3vekbhcyz45hfcwngq

Automatic Detection Of Hindustani Talas

Marius Miron
2011 Zenodo  
The thesis aims to develop a system of Hindustani tala automatic recognition which can be trained by building a labeled corpus of Hindustani songs with tabla accompaniment. Most of the research concerning rhythm in the North Indian classical music was developed around monophonic recordings and the scope was just recognizing the tabla strokes or modeling the expressiveness of tabla solos and not the metric cycles in which these strokes usually occur, the talas. The aspects researched were
more » ... ation and stroke recognition in a polyphonic context, as recognizing the talas is a perceptually challenging task and the automatic detection proved to be even more dificult.
doi:10.5281/zenodo.1162292 fatcat:2vtev66ejbfzrdqt5o3zeulunu

Piano Fingering with Reinforcement Learning [article]

Pedro Ramoneda, Marius Miron, Xavier Serra
2021 arXiv   pre-print
Hand and finger movements are a mainstay of piano technique. Automatic Fingering from symbolic music data allows us to simulate finger and hand movements. Previous proposals achieve automatic piano fingering based on knowledge-driven or data-driven techniques. We combine both approaches with deep reinforcement learning techniques to derive piano fingering. Finally, we explore how to incorporate past experience into reinforcement learning-based piano fingering in further work.
arXiv:2111.08009v1 fatcat:crqstvfycnhajhrfvywdgnphhy

Binaural Source Separation with Convolutional Neural Networks [article]

Gerard Erruz, Marius Miron, Adan Garriga
2017 Zenodo  
This work is a study on source separation techniques for binaural music mixtures. The chosen framework uses a Convolutional Neural Network (CNN) to estimate time-frequency soft masks. This masks are used to extract the different sources from the original two-channel mixture signal. Its baseline single-channel architecture performed state-of-the-art results on monaural music mixtures under low-latency conditions. It has been extended to perform separation in two-channel signals, being the first
more » ... wo-channel CNN joint estimation architecture. This means that filters are learned for each source by taking in account both channels information. Furthermore, a specific binaural condition is included during training stage. It uses Interaural Level Difference (ILD) information to improve spatial images of extracted sources. Concurrently, we present a novel tool to create binaural scenes for testing purposes. Multiple binaural scenes are rendered from a music dataset of four instruments (voice, drums, bass and others). The CNN framework have been tested for these binaural scenes and compared with monaural and stereo results. The system showed a great amount of adaptability and good separation results in all the scenarios. These results are used to evaluate spatial information impact on separation performance.
doi:10.5281/zenodo.1095835 fatcat:mye3jo4smng7jbvvnpk57ysgdi

Audio Source Separation Using Deep Neural Networks

Pritish Chandna, Jordi Janer, Marius Miron
2016 Zenodo  
[Miron et al., 2015] However, since this thesis does not focus on informed source separation, a detailed explanation of the same will not be provided herewith.  ... 
doi:10.5281/zenodo.3755620 fatcat:girvxhgbv5cqplktyzmv22gaqu

Assessing Algorithmic Biases for Musical Version Identification [article]

Furkan Yesiler and Marius Miron and Joan Serrà and Emilia Gómez
2021 arXiv   pre-print
Version identification (VI) systems now offer accurate and scalable solutions for detecting different renditions of a musical composition, allowing the use of these systems in industrial applications and throughout the wider music ecosystem. Such use can have an important impact on various stakeholders regarding recognition and financial benefits, including how royalties are circulated for digital rights management. In this work, we take a step toward acknowledging this impact and consider VI
more » ... stems as socio-technical systems rather than isolated technologies. We propose a framework for quantifying performance disparities across 5 systems and 6 relevant side attributes: gender, popularity, country, language, year, and prevalence. We also consider 3 main stakeholders for this particular information retrieval use case: the performing artists of query tracks, those of reference (original) tracks, and the composers. By categorizing the recordings in our dataset using such attributes and stakeholders, we analyze whether the considered VI systems show any implicit biases. We find signs of disparities in identification performance for most of the groups we include in our analyses. Moreover, we also find that learning- and rule-based systems behave differently for some attributes, which suggests an additional dimension to consider along with accuracy and scalability when evaluating VI systems. Lastly, we share our dataset with attribute annotations to encourage VI researchers to take these aspects into account while building new systems.
arXiv:2109.15188v1 fatcat:4drrscwbcncszby4tuxx3vv34e

Exploring Robust Music Fingerprinting Methods With Data-driven Methodologies

Aditya Bhattacharjee, Marius Miron, Furkan Yesiler
2021 Zenodo  
Firstly, I would like to thank my co-supervisors: Marius Mirion and Furkan Yesiler, for giving me their time as well as their virtual machine's.  ...  I would also like to extend my gratitude to Marius for important assessments during the dissertation, both in preparation for it and in evaluation.  ... 
doi:10.5281/zenodo.5553864 fatcat:6x5lovub3fbqlkrq2asmig7bhm

Assessing The Tuning Of Sung Indian Classical Music

Joan Serrà, Gopala K. Koduri, Marius Miron, Xavier Serra
2011 Zenodo  
[TODO] Add abstract here.
doi:10.5281/zenodo.1415101 fatcat:lpqbbgvhircfro4dpg5rgq2o7i

PodcastMix: A dataset for separating music and speech in podcasts

Nicolás Schmidt, Marius Miron, Jordi Pons
2021 Zenodo  
Over the last few years, the popularity of podcast shows in streaming services has increased considerably. Licensed music in these shows is frequently used, but the precision of song identification services could be a˙ected by the speakers voice in the mix. This presents a major problem both for the musicians, who do not receive their respective royalty payments, and for the broadcasters, who may be exposed to legal problems for non-compliance with international copyright laws. In this Master
more » ... esis, a benchmark between two state of the art models for music source separa-tion, the ConvTasNet and the UNet, was performed against a novel Podcast-like audio dataset called PodcastMix with the objective of separating both the voice of the speakers and the background music from a podcast. In this way, the back-ground music and foreground speech source separation task was formalized. This new dataset is compound by music from the Jamendo free music streaming service, mixed with the VCTK speech dataset. The models were trained on this dataset and evaluated both in the test partition and on a dataset of real podcasts. The results show that UNet performs better than ConvTasNet in separating speakers and music from podcasts. The benchmark was performed using the Asteroid toolkit and the evaluation metrics were computed using BSSEval tool in order to measure the quality of the separations.
doi:10.5281/zenodo.5554789 fatcat:75wg7qrslnez5buxw46mublk54

Fairness, Accountability and Transparency in Music Information Research (FAT-MIR) [article]

Emilia Gomez, Andre Holzapfel, Marius Miron, Bob L. Sturm
2019 Zenodo  
This tutorial focuses on the timely issues of ethics, fairness, accountability and transparency, with particular attention paid to research in applications in music information research. These topics arise from a broader consideration of ethics in the field – related work of which was recently published in TISMIR ( These topics are also receiving attention in the broader domain of machine learning and data science, e.g., the FAT-Machine
more » ... Learning (ML) conference 2014-2018, Explainable AI workshops 2017-2018, Interpretable Machine Learning workshops, and in the context of the HUMAINT project and winter school on ethical, legal, social and economic impact of Artificial Intelligence ( This tutorial is suitable for researchers and students in MIR working in any domain, as these issues are relevant for all MIR tasks, from low- to high-level, from system to user-centered research. There are no prerequisites for taking this tutorial.
doi:10.5281/zenodo.3546227 fatcat:g7si4auptber7badmlwoz57kvm

Addressing multiple metrics of group fairness in data-driven decision making [article]

Marius Miron, Songül Tolan, Emilia Gómez, Carlos Castillo
2020 arXiv   pre-print
The Fairness, Accountability, and Transparency in Machine Learning (FAT-ML) literature proposes a varied set of group fairness metrics to measure discrimination against socio-demographic groups that are characterized by a protected feature, such as gender or race.Such a system can be deemed as either fair or unfair depending on the choice of the metric. Several metrics have been proposed, some of them incompatible with each other.We do so empirically, by observing that several of these metrics
more » ... luster together in two or three main clusters for the same groups and machine learning methods. In addition, we propose a robust way to visualize multidimensional fairness in two dimensions through a Principal Component Analysis (PCA) of the group fairness metrics. Experimental results on multiple datasets show that the PCA decomposition explains the variance between the metrics with one to three components.
arXiv:2003.04794v1 fatcat:25kkfwana5g6noexaq4apwoatq

Computational Approaches For The Understanding Of Melody In Carnatic Music

Gopala K. Koduri, Marius Miron, Joan Serrà, Xavier Serra
2011 Zenodo  
[TODO] Add abstract here.
doi:10.5281/zenodo.1415237 fatcat:3nimj3ij45hvraztaelhgwqjo4

Audio-To-Score Alignment At The Note Level For Orchestral Recordings

Marius Miron, Julio José Carabias-Orti, Jordi Janer
2014 Zenodo  
Attribution: Marius Miron, Julio José Carabias-Orti, Jordi Janer.  ...  Phenicx Project 103] is the only one addressing explicitly the topic of fine note c Marius Miron, Julio José Carabias-Orti, Jordi Janer.  ... 
doi:10.5281/zenodo.1416150 fatcat:jksc5hlmffeodcmo77wrzw5qcq

Improving Score-Informed Source Separation For Classical Music Through Note Refinement

Marius Miron, Julio José Carabias-Orti, Jordi Janer
2015 Zenodo  
[TODO] Add abstract here.
doi:10.5281/zenodo.1417688 fatcat:si3iyzbognf4dkwlxd5ifkzlqa

Data-driven Vocal Pitch Extraction for Indian Art Music Melodic Analysis

Genís Plaja I Roglans, Xavier Serra, Marius Miron
2021 Zenodo  
The melodic exploration of Indian Art Music is an emerging topic in the field of Music Information Retrieval (MIR), including several tasks to contribute to the understanding of this tradition. Ground-truth vocal melody annotations are an essential material to tackle this challenge, but the automatic extraction of these data from polyphonic audio signals is an unsolved task as of now. The state of the art on this topic is currently proposing high-performance data-driven pitch extraction models.
more » ... Nevertheless, these models are exclusively developed using Western music data (pop, rock, blues and related styles) and therefore, the performance of these algorithms on Indian Art Music signals is significantly degraded. Furthermore, given the shortage of properly annotated vocal melody ground-truth for Indian Art Music, there are no works in the literature that propose pitch extraction methods or even re-train state of the art algorithms for this music tradition. In this work, we aim at overcoming this issue by addressing two main contributions: (1) The creation of a dataset of properly annotated vocal melody for Indian Art Music, and (2) The training, evaluation and testing of a state of the art vocal pitch extraction method to obtain a trained model that outperforms the actual proposals for Indian Art Music signals. Additionally, we review the related literature on Indian Art Music melodic exploration and on pitch extraction, to provide the basis for our research. We also contribute to the mirdata Python library by integrating loaders for Indian Art Music and other World traditional music datasets and promote the research on these cultures. We hope that the outcomes of this thesis contribute to the Indian Art Music melodic understanding and to the overcoming of the actual barriers produced by the shortage of data and methodologies for melody extraction of this rich and relevant musical culture.
doi:10.5281/zenodo.5554703 fatcat:poiog55knrhfjncg7ri5lxyjiy
« Previous Showing results 1 — 15 out of 375 results