DLfM Ottoman-Turkish Makam Recognition and Tonic Identification Experiments This repository hosts the experiments conducted, which demonstrates MORTY (our mode and tonic estmation toolbox) in the paper: Karakurt, A., Şentürk S., & Serra X. (2016). MORTY: A Toolbox for Mode Recognition and Tonic Identification. 3rd International Digital Libraries for Musicology Workshop. New York, USA Please cite the paper above, if you are using the contents of this repository for your works. The companion page of the paper is accesible via the CompMusic website. Structure of the Repository The scripts are located in the base folder along with several miscallenaeous files (the license, readme, setup and requirement files). The folder ./data links to the relevant commit in our makam recognition dataset, the folds and the summary of the evaluation obtained from all experiments. Due to file size constraints features, training models, results and evaluation files are not included in this folder and stored in Zenodo (link) instead. The folder ./dlfm_code has the relevant Python and MATLAB modules for the training, testing and evaluation. Installation If you want to install the Python package, it is recommended to install the package and dependencies into a virtualenv. In the terminal, do the following: virtualenv env source env/bin/activate python setup.py install The package and some of its dependencies use several modules in Essentia. Follow the instructions to install the library. Now you can install the rest of the dependencies: pip install -r requirements Experimentation Scripts We use Jupyter notebooks for the general experimentation and MATLAB for statistical significance tests. To reproduce the experiments you can run the scripts in the order given below: setup_feature_training.ipynb: Create the folds in the stratified 10 fold scheme, extract distribution features, train models. testing_evaluation.ipynb: Find optimal minimum peak ratio (Figure 3 in the paper), testing and evaluation. plot_tonicdist_confusionmat.ipynb: Display the tonic identification errors for all parameter sets with 7.5 cent bin size and the confusion matrix in mode recognition for the best parameter set summarize_evaluation.m: Read and store a summarization of the evaluation obtained for all parameter sets stat_significance.m: Conduct statistical significance tests. The last block in the MATLAB code is where the tests are carried semi-automatically. The parameters that are checked for significance are commented. License The source code hosted in this repository is licenced under Affero GPL version 3. The data (the features, models, figures, results etc.) are licenced under Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.