Salience of Frequency Micro-modulations in Popular Music

Abstract

Singing voices attract auditory attention in music unlike other sounds. In a previous study, we investigated the salience of instruments and vocals using a detection task in which cued target sounds were to be detected in musical mixtures. The presentation order of cue and mixture signals influenced the detection of all targets except the lead vocals, indicating that listeners focus on voices regardless of whether these are cued or not, highlighting a unique vocal salience in music mixtures. The aim of the present online study was to investigate the extent to which phonological cues, musical features of the main melody, or frequency micro-modulation (FMM) inherent in singing voices contribute to this vocal salience. FMM was either eliminated by using an autotune effect (Experiment 1) or transferred to other instruments (Experiment 2). Detection accuracy was influenced by presentation order for all instrumental targets and the autotuned vocals, but not for the unmodified vocals, suggesting that neither the phonological cues that could provide a facilitated processing of speech-like sounds nor the musical features of the main melody are sufficient to drive vocal salience. Transferring FMM from vocals to instruments or autotuned vocals reduced the magnitude of the order effect considerably. These findings suggest that FMM is an important acoustical feature contributing to vocal salience in musical mixtures.