Tech Trends in Music Education: Final Project Overview

Name That Bird: An Android Application for Crowdsourcing the Recording and Identification of Birdsongs

Currently, there exists no Android application that enables users to simply record birdsongs and contribute their recordings to a database. Such a resource would be beneficial to many types of users, such as bird enthusiasts, ornithologists, data scientists, and bird-influenced musicians such as myself. Birdsong apps currently in the marketplace mostly come in the form of field guide apps such as iBird [1], Audubon [2], Sibley [3]; and call playback apps such as blueBill [4]. Additionally, there are two apps, BirdSongId [5] and WeBIRD [6], that are soon entering the marketplace that can perform automatic call recognition. However, while there are some web services, such as AmiBio [7] and ARBIMON [8], that are aimed at collecting recordings in order to improve automatic biodiversity monitoring algorithms, no such smart phone app exists that allows the user to record, identify, upload, and explore birdsong recordings to a database manually. Such a resource would be a very valuable asset that could provide larger data sets for data scientists that would increase the accuracy of automatic birdsong recognition algorithms employed in apps such as [5, 6]. While sound data sets such as Million Song Dataset [9], Freesound.org [10], Macaulay Library [11], and XenoCanto.org [12] provide access to many songs, sounds, and birdsongs, they are not set up in such a way to faciliate the automatic upload of sounds using means such as an Android application. Thus, I am writing a smartphone application that will facilitate the crowdsourcing of birdsong recordings into a central database that will be freely accessible to the public.

Name That Bird is an application that will facilitate the crowdsourcing of a large database of birdsong recordings. The database will then be used as the primary data set for training and testing automatic birdsong detection and species classification algorithms.

The process for building this application will include building several versions of the app, integrating user feedback into the design process as the app progresses. First, I will create a basic Android app with recording and playback functionality. This component of the app will function similar to the Heard [13] and Shoudio [14] smart phone applications. Then, I will set up a server to store the audio files and create a MySql database for storing metadata, such as the path to the audio file and any useful metadata. Once the database is functioning, I will integrate the Android app with the database, which involves making it possible for a user to upload a sound recording and other data to the database by using the app. Next, I will perform the first round of user testing with a pre-selected group of ornithologists and bird enthusiasts. I will then incorporate their feedback into a second iteration of the app, especially their comments on the user interface and experience. Finally, I will perform another round of user testing and release the app to the general public.

This is a huge project that involves writing large amounts of code to perform many different tasks. Thus, the timeline for creating this app extends beyond the duration of the class. However, I plan to at least create an app that can allow users to record, identify, and store the recordings on their mobile devices. I will only incorporate the database and server functionality to the extent that time permits.

References

[1] iBird: http://ibird.com/

[2] Audubon: http://www.audubonguides.com/field-guides/bird-identification-app.html

[3] Sibley: http://www.sibleyguides.com/about/the-sibley-eguide-to-birds-app/

[4] blueBill: http://bluebill.tidalwave.it/mobile/

[5] BirdSongId: http://www.isoperla.co.uk/BirdSongIdiPhone.html

[6] WeBIRD: http://grow.cals.wisc.edu/environment/smart-birding

[7] AmiBio: http://www.amibio-project.eu/

[8] ARBIMON: http://arbimon.com/

[9] Million Song Dataset: http://labrosa.ee.columbia.edu/millionsong/

[10] Freesound.org: http://freesound.org/

[11] Macaulay Library: http://macaulaylibrary.org/