Background
For the fields of cognitive and computational musicology, online listening experiments can attract a much larger, more diverse, and intrinsically motivated group of participants than studies conducted in the lab (Burgoyne et al., 2013; Germine et al., 2012; Honing & Reips, 2008). Thereby, they can contribute to a more complete understanding of our capacity for music and the role of music in our society (Honing, 2021). While tools for creating online questionnaires have existed for a while (e.g., Qualtrix, LimeSurvey, among others), they often have limited capacity or flexibility when it comes to music playback, control over timing, or the design of the user interface.
Aims
With this background in mind, our aims are to develop an infrastructure in which multiple modes of audio playback and precise timing would be supported. Moreover, we wanted a framework in which styles, such as logos, fonts and colors, can be adapted easily per experiment. Lastly, the infrastructure should consist of reusable, testable components, and be easy to install both locally and on a server.
Main Contribution
We have implemented the MUSic-related Citizen Science Listening Experiments (MUSCLE) infrastructure, which consists of a Django backend, and a React frontend. The backend manages the experimental data, which is presented to the user and saved to the database; the frontend streamlines the design of an attractive user interface.
At this point, the infrastructure can display various types of audio player (auto-play, single or multiple play buttons), customizable forms (choice or open questions with multiple widgets to choose from, and which can be presented grouped, or spread over multiple screens), and define and carry out sophisticated experimental designs using rules definitions. We have integrated commonly used questions, such as the Goldsmiths Music
Sophistication Index and created templates for typical use cases for online music experiments (such as forced choice). The application’s dependencies are formalized as Dockerfiles. Using Docker Engine or Kubernetes, the application can be set up on a local system or a remote virtual machine.
Discussion
We have already used the infrastructure to run various online experiments (https://app.amsterdammusiclab.nl/). We were able to collect and analyze data from hundreds of users. We would like to collaborate with other music research labs to increase the reusability of the infrastructure even further, and ultimately, release the software as an open-source package.
For the fields of cognitive and computational musicology, online listening experiments can attract a much larger, more diverse, and intrinsically motivated group of participants than studies conducted in the lab (Burgoyne et al., 2013; Germine et al., 2012; Honing & Reips, 2008). Thereby, they can contribute to a more complete understanding of our capacity for music and the role of music in our society (Honing, 2021). While tools for creating online questionnaires have existed for a while (e.g., Qualtrix, LimeSurvey, among others), they often have limited capacity or flexibility when it comes to music playback, control over timing, or the design of the user interface.
Aims
With this background in mind, our aims are to develop an infrastructure in which multiple modes of audio playback and precise timing would be supported. Moreover, we wanted a framework in which styles, such as logos, fonts and colors, can be adapted easily per experiment. Lastly, the infrastructure should consist of reusable, testable components, and be easy to install both locally and on a server.
Main Contribution
We have implemented the MUSic-related Citizen Science Listening Experiments (MUSCLE) infrastructure, which consists of a Django backend, and a React frontend. The backend manages the experimental data, which is presented to the user and saved to the database; the frontend streamlines the design of an attractive user interface.
At this point, the infrastructure can display various types of audio player (auto-play, single or multiple play buttons), customizable forms (choice or open questions with multiple widgets to choose from, and which can be presented grouped, or spread over multiple screens), and define and carry out sophisticated experimental designs using rules definitions. We have integrated commonly used questions, such as the Goldsmiths Music
Sophistication Index and created templates for typical use cases for online music experiments (such as forced choice). The application’s dependencies are formalized as Dockerfiles. Using Docker Engine or Kubernetes, the application can be set up on a local system or a remote virtual machine.
Discussion
We have already used the infrastructure to run various online experiments (https://app.amsterdammusiclab.nl/). We were able to collect and analyze data from hundreds of users. We would like to collaborate with other music research labs to increase the reusability of the infrastructure even further, and ultimately, release the software as an open-source package.