Variation in multitrack mixes : analysis of low-level audio signal features

To further the development of intelligent music production tools, towards generating mixesthat would realistically be created by a human mix-engineer, it is important to understand whatkind of mixes can be created, and are typically created, by human mix-engineers. This paperpresents an analysis of 1501 mixes, over 10 different songs, created by mix-engineers. Theprimary dimensions of variation in the full dataset of mixes were ‘amplitude’, ‘brightness’,‘bass’ and ‘width’, as determined by feature-extraction and subsequent principal componentanalysis. The distribution of representative features approximated a normal distribution andthis is then used to obtain general trends and tolerance bounds for these features. The resultspresented here are useful as parametric guidance for intelligent music production systems