You can also find all these publications on my Zenodo directory.
All the code that I can publish is available in my Framasoft Gitlab page.
ASMD: an automatic framework for compiling multimodal datasets with audio and scores
This paper describes an open-source Python framework for handling datasets for music processing tasks, built with the aim of improving the reproducibility of research projects in music computing and assessing the generalization abilities of machine learning models. The framework enables the automatic download and installation of several commonly used datasets for multimodal music processing. Specifically, we provide a Python API to access the datasets through Boolean set operations based on particular attributes, such as intersections and unions of composers, instruments, and so on. The framework is designed to ease the inclusion of new datasets and the respective ground-truth annotations so that one can build, convert, and extend one’s own collection as well as distribute it by means of a compliant format to take advantage of the API. All code and ground-truth are released under suitable open licenses.
A Convolutional Approach to Melody Line Identification in Symbolic Scores
Have you ever tried to understand what is the melody in a music score? Even if it seems easy for us, its definition is difficutl to formalize under strict mathemathical terms, and, thus, it is a complex task for a computer. We show a new state-of-art method adopting convolutional neural network to tackle this challenge: can a computer automatically identify the melody in a music score? See code and demos at: github website
On the Adoption of Standard Encoding Formats to Ensure Interoperability of Music Digital Archives: The IEEE 1599 Format
Have you ever wanted to merge music information from different websites? In this paper we propose a standard music file format, the IEEE 1599, which was born with the exact purpose of representing multimodal music information. We explain how this format can be useful for digital libraries, archives and scientific dataset, allowing to overcome updating and copyright issues.
Multimodal Music Information Processing and Retrieval: Survey and Future Challenges
This is an exhaustive review of the literature exploiting multiple sources of information to create amazing applications. Here, you can discover how heterogeneous aspects of music information have been used for typical music information retrieval tasks. This is the very first definition of the field “Multimodal Music Information Retrieval”. We show that many works (more than 80 papers) have already approached the field and that the multimodal approach can really improve upon standard approaches. We also list the future challenges and the methods that must still be tested in this promising field.
Symbolic Music Similarity Through a Graph-Based Representation
We describe a new method to represent music information at the symbolic level (such as music scores). We test it in a few retrieval tasks, showing that it has some benefits for a generic stylistic representation. It is able to represent music both in its contrapuntal and harmonic texture. It paves new ways in automatic music composition, music information retrieval and any task which deals with music scores (music transcription, optical music recognition, audio-to-score alignment, etc).
Enhanced Wikifonia Leadsheet Dataset
EWLD (Enhanced Wikifonia Leadsheet Dataset) is a music leadsheet dataset with more than 5.000 scores that comes with a lot of metadata about composers, works, lyrics and features. It is designed for musicological and research purposes.
A Public Domain version, named OpenEWLD, is available at https://framagit.org/sapo/OpenEWLD.
You can find an in-deep discussion in my Master Thesis.
Please, use the following paper as reference:
Simonetta, Federico, Carnovalini, Filippo, Orio, Nicola, & Rodà, Antonio. (2018). Symbolic Music Similarity Through a Graph-Based Representation. In Proceedings of the Audio Mostly 2018 on Sound in Immersion and Emotion (pp. 26:1–26:7). New York, NY, USA: ACM. http://doi.org/10.1145/3243274.3243301 Zenodo link: https://zenodo.org/record/2537059
My Master Thesis:
F. Simonetta, “Graph based representation of the music symbolic level. A music information retrieval application”, Università di Padova, 2018. Zenodo link: https://zenodo.org/record/1476564