Arab-Andalusian Corpus Analysis
Description
This repository contains a docker-compose file to run a Jupyter server and the notebooks to download and analyse data and metadata from the Arab Andalusian Corpus of Dunya. The structure of the repository is the next one:
- /nawba_recognition:
- NawbaPitchAnalysis.ipynb: to visualize pitch distribution and note/class distribution of a single recording or of a group of them.
- NawbaRecognition.ipynb: to compute several experiments to evaluate the performance on nawba recognition of algorithms based on templates derived from scores.
- /demos_notebooks:
- demo_andalusian_api.ipynb: to understand the composition of the API from pycompmusic and a small application.
- download_interface_corpus.ipynb: to download data and metadata from the Corpus, to compute the pitch profile, distribution and the tonic frequency of each recording;
- metadata_visualization_interface.ipynb: to group, visualize and analyse metadata;
Installation
To run the notebooks, you need to first install docker. Here you can find the links to installation instructions for different operative systems:
Usage
In a terminal/console window, change to this directory
On MacOS or Windows, run:
docker-compose up
On Linux, run the following (this command ensures that any files you create are owned by your own user):
JUPYTER_USER_ID=$(id -u) docker-compose up
The first time you run this command it will download the required docker images (about 2GB in size).
Then accesss http://localhost:8888 with your browser and when asked for a password use the default password mir
Then, you can access the notebooks from the browser and run them. All the notebooks contain their user guides.
The use of Dunya data and metadata requires that you register with Dunya. After the registration, a personal token is provided. This token has to be added in the utilities/constants.py
file.
NB: the computation of the data and metadata can require a couple of days. For this reason, the nawba recognition experiment includes a zip file with the necessary pre-computed files.
Credits
This work is based on a collaboration between Niccolò Pretto, Miguel García Casado, Barış Bozkurt, Rafael Caro Repetto and Xavier Serra, as part of the project Musical Bridges. The notebooks use the Arab Andalusian Corpus in Dunya, created during the CompMusic project.
These notebooks are based on MIR-docker-extension (https://github.com/MTG/MIR-docker-extension). The repository includes snippets of code and algorithms from the following repositories:
- pycompmusic: https://github.com/MTG/pycompmusic
- dunya: https://github.com/MTG/dunya
- tomato: https://github.com/sertansenturk/tomato
- morty: https://github.com/altugkarakurt/morty
- dunya desktop: https://github.com/MTG/dunya-desktop
References
[1] Caro Repetto, R; Pretto, N; Chaachoo, A; Bozkurt, B; and Serra X "An open corpus for the computational research of Arab-Andalusian music", in Proceedings of Proceedings of 5th International Conference on Digital Libraries for Musicology (DLfM 2018); 2018 Sep 28; Paris, France. New York: ACM; 2018. p. 78-86. DOI: 10.1145/3273024.3273025
[2] Pretto, N; Bozkurt, B; Caro Repetto, R; and Serra, X "Nawba Recognition for Arab-Andalusian Music using Templates from Music Scores", Proceedings of the 15th Sound and Music Computing Conference 2018, SMC 2018, Limassol, Cyprus. DOI: 10.5281/zenodo.1257387