jakezhaojb / medleydb

medleydb

Python tools for using MedleyDB.

Created by Rachel Bittner rachel (dot) bittner (at) nyu (dot) edu and Justin Salamon justin (dot) salamon (at) nyu (dot) edu.

This code is released along with MedleyDB:

http://medleydb.weebly.com (or http://marl.smusic.nyu.edu/medleydb)

This code is a component of the work presented in the following publication:

R. Bittner, J. Salamon, M. Tierney, M. Mauch, C. Cannam and J. P. Bello, "MedleyDB: A Multitrack Dataset for Annotation-Intensive MIR Research", in 15th International Society for Music Information Retrieval Conference, Taipei, Taiwan, Oct. 2014.

Setup

First clone this repository:

git clone https://github.com/rabitt/medleydb.git

Next, set the environment variable MEDLEYDB_PATH to the local path where the MedleyDB directory (or MedleyDB_sample) lives:

export MEDLEYDB_PATH="path/to/your/copy/of/MedleyDB"

To avoid doing this step every time, copy the line above to .bash_profile or .bash_rc.

Dependencies

• pyyaml
• sox (not strictly necessary)

If you use pip and homebrew, you can install them by doing:

[sudo] pip install pyyaml
brew install sox

Usage and Examples

To load the module:

import medleydb as mdb

Loading and accessing the full dataset

Load the dataset to a list of MultiTrack objects:

mtrack_list = mdb.load_all_multitracks()

Some attributes of a multitrack:

multitrack_1 = next(mtrack_list)
multitrack_2 = next(mtrack_list)

multitrack_1.has_bleed
multitrack_2.has_bleed

multitrack_1.artist
multitrack_2.artist

multitrack_1.is_instrumental
multitrack_2.is_instrumental

multitrack_1.melody1_annotation

Some attributes of a particular stem:

example_stem = multitrack_1.stems[0]
example_stem.instrument
example_stem.pitch_annotation

Loading a subset of the dataset

A subset of the dataset can be loaded by passing a list of track folder paths:

import glob
import os
import medleydb as mdb

example_list = glob.glob(os.path.join(mdb.AUDIO_DIR, 'ClaraBerry_*'))
dataset_subset = mdb.load_multitracks(example_list)

Getting filepaths for audio of a chosen instrument

The following code will get a list of paths to stems that are labeled as a clarinet:

import medleydb as mdb

dataset = mdb.load_all_multitracks()
clarinet_files = mdb.get_files_for_instrument('clarinet')

To see a full list of possible instrument labels:

import medleydb as mdb

instruments = mdb.get_valid_instrument_labels()
print instruments