Skip to main content
Join the official 2019 Python Developers SurveyStart the survey!

Small example datasets of annotated vocalizations

Project description


Small example datasets of annotated vocalizations.

Useful if you have:

  • need some example vocalizations that are quick to download
  • build a tool that works with different annotation formats and you want to test that tool


There are two components to pollyglot:

  1. command-line tool that creates the small example datasets from larger publicly-available datasets
  2. package that fetches the small example datasets, which can be a dependency for your library

To use (1), you invoke the command-line tool pollymake

Cloning this repository, installing it for development (see below), and then calling

$ pollymake all

will re-make the dataset within the repository.

pollymake creates an archive from each repository. These are then uploaded to a Figshare dataset repository:

The goal of this package to share code that automates the process of creating a data repository on FigShare, and make this source open for collaboration. The formats in this repository can be parsed by the Crowsetta package. Development and tutorials for crowsetta make use of the small, quick-to-download archives of each format on Figshare that are generated from the source in this repository.

crowsetta provides tools for anyone that wants to write clean code when working with these annotation formats (or their own format) To learn more, please visit

formats + references

Below are the formats included and references for the sources.

Praat textgrid

Textgrids output by the Praat program.

Songs with Praat textgrid format are from the Birdsong Database provided by the Taylor lab at UCLA:
as presented in this paper: The .xls file containing links to songs from the Taylor lab birdsong database was created by Tim Sainburg to train generative networks for animal vocalizations:; adapted under MIT license.


.not.mat files are output by the evsonganaly GUI created by Evren Tumer in the Brainard lab. The audio file format .cbin is output by the Labview program EvTAF.

Another repository of Bengalese finch song annotated in this format is here:


A specific .xml format for a repository of labeled Bengalese Finch song. The repository is here: The repository provides data for testing a convolutional neural network to segment and label vocalizations, as shared in the repository and discussed in the paper "Automatic recognition of element classes and boundaries in the birdsong with variable sequences" by Takuya Koumura and Kazuo Okanoya (

Project information


Pollyglot (c) by David Nicholson, 2018-2019.

Code is shared under BSD-3 License.

Where applicable, data in the vocal-annotations-formats-dataset is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. (The figshare repositories are shared under CC-BY-4.0) Where the authors have not made their intentions clear with a license, citations to papers and links to the original source are included. Please raise an issue on this repository if there are any concerns about this.

You should have received a copy of the license along with this work. If not, see

Project details

Release history Release notifications

This version


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for pollyglot, version 0.1.1
Filename, size File type Python version Upload date Hashes
Filename, size pollyglot-0.1.1-py3-none-any.whl (21.2 kB) File type Wheel Python version py3 Upload date Hashes View hashes
Filename, size pollyglot-0.1.1.tar.gz (13.2 kB) File type Source Python version None Upload date Hashes View hashes

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page