ABSTRACT
It is generally agreed that vocal clarity is one of the most important quality parameters of popular music mixes (Ronen, 2015; Hermes et al., 2017, pp. 10–11). However, there is currently no formally proven basis of what makes recorded singing clear, nor a generally agreed-on definition of vocal clarity. Two computational predictors exist for the spectral clarity of a range of isolated sound sources, including piano, guitar, strings and one vocal take (Hermes et al., 2017). It might be possible to develop similar predictors specifically for vocal clarity in mixes if this depends on a set of clearly defined, generalizable acoustic parameters. Findings in this area would be useful as they could, for example, inform assistive, artificially intelligent mix tools for vocals or help aspiring mix engineers further their skills. In order to assess vocal clarity further, it appears beneficial to draw upon the knowledge of both scientists and creative practitioners in a cross-disciplinary approach. A broad overview of current findings on vocal clarity can be found in the first section. The aforementioned computational predictors of single sound spectral clarity (Hermes et al., 2017) are subsequently tested in the context of a vocal mix in the second section in order to assess whether they offer a useful starting point for measuring vocal clarity in mixes. This is initially done as an autoethnographic study, but feedback is also sought from other audio professionals in a short, indicative pilot listening test. Based on this assessment, suggestions for further research are presented.
