An approach to assess loudness and dynamics with Web Audio native nodes

Abstract

Music content providers on the Internet like YouTube, Spotify or Apple Music, as well as a range of software playback systems like the media player “foobar2000” have a loudness normalization feature to match a series of diverse audio tracks in overall loudness. This is done to keep perceived volume differences between audio tracks as low as possible. Thus, it is important for music producers, especially mastering engineers, to master audio tracks with a particular amount of dynamic range, so that streaming services will not turn the playback volume of their tracks down. With their already low dynamic range, a listener would now even better be able to recognize their inferior sound compared to other tracks with higher dynamic range. To correctly assess the dynamics of audio material, this paper introduces two web applications that compute and visualize the loudness and dynamic range of audio material, using a subset of the loudness units described in the recommendation R 128 by the European Broadcasting Union, and using only native notes by the W3C Web Audio API.