Audio-Tokens: A toolbox for rating, sorting and comparing audio samples in the browser

Area of Science:

• Psychology
• Computer Science
• Machine Learning

Background:

• Online rating studies are crucial for collecting perceptual data.
• Existing tools may lack flexibility for diverse auditory research needs.
• Efficient annotation of audio datasets is essential for machine learning.

Purpose of the Study:

• To present a novel JavaScript toolbox for conducting online auditory rating studies.
• To facilitate the collection of various types of audio ratings, including feature and similarity judgments.
• To offer a versatile tool for both psychological research and machine learning applications.

Main Methods:

• Development of a JavaScript toolbox with visual tokens controlling audio playback.
• Integration with jsPsych plugin and availability as plain JavaScript.
• Support for single- and multidimensional feature ratings, categorical, and similarity ratings.

Main Results:

• The toolbox allows for interactive and flexible online collection of auditory data.
• It accommodates diverse rating paradigms, enhancing data collection possibilities.
• The tool is adaptable for custom applications beyond jsPsych.

Conclusions:

• The developed JavaScript toolbox provides a powerful and flexible solution for online auditory rating studies.
• It is expected to advance research in speech and music perception.
• The toolbox will aid in the curation and annotation of audio datasets for machine learning.