Both acoustic intensity and loudness contribute to time-series models of perceived affect in response to music.

Abstract

Research in the fields of music and emotion has investigated relationships between acoustic properties of music and listeners' perception of affect in response to music. Specifically, the dimensions of arousal and valence that comprise a two-dimensional circumplex model of experienced affect (Russell, 1980, 2003) are commonly measured in experiments investigating the multiplicity of acoustic cues associated with perceived affect (Bailes & Dean, 2012; Dean & Bailes, 2010; Dean, Bailes, & Schubert, 2011; Gingras, Marin, & Fitch, 2014; Olsen & Stevens, 2013; Ritossa & Rickard, 2004; Schubert, 1999, 2004, 2013). With additional use of continuous perceptual measurements, a clearer picture of the real-time factors that underpin affective response to music as it unfolds through time is beginning to emerge.From causal experiments, time-series models of continuous responses to music from the Western classical tradition have shown that manipulations of acoustic intensity profiles of performed music, as measured by decibels (dB) in SPL, significantly influence listeners' real-time perception of affect and loudness (Dean et al., 2011). A strong impact of acoustic intensity on perceived loudness in music perception is to be expected, as a long tradition of psychophysical research reports an intimate (yet not straightforward) relationship between loudness and acoustic intensity (Canevet & Scharf, 1990; Fletcher & Munson, 1933; Florentine, Popper, & Fay, 2011; Olsen, 2014; Olsen, Stevens, & Tardieu, 2010; Scharf, 1978; Stevens, 1956). One would also expect, therefore, that where intensity predicts perceived loudness and perceived affect, loudness might either share with intensity a predictive capacity for perceived affect, or mediate the role of intensity when affective responses are modeled.Somewhat contrary to this hypothesis, time-series modeling of continuously perceived affect in response to one piece from the Western classical tradition (an extract from Dvorak's Slavonic Dance Opus 46, No. 1) included better prediction of perceived affect from acoustic intensity than from continuous loudness (Dean et al., 2011), with loudness not required as a component of the optimized model. The relevance of the surprising exclusion of loudness in this particular model is assessed in the present study. Is this phenomenon specific to the particular Dvorak piece, or is it more general across music with varied complexity and familiarity? As argued above, we expect that loudness will either share a predictive capacity with intensity or mediate the role of intensity when affective responses to more diverse musical stimuli are modeled. To investigate this hypothesis further, acoustic intensity and perceived loudness must be manipulated in a way that subtly distorts the inherently strong psychophysical relationship, so their potential influences on affect can be distinguished empirically. This is accomplished here with the manipulation of synthesized reverberation, an acoustic cue active for listeners' impressions of auditory space and perceived auditory distance, which in turn has been shown to affect perceived loudness of a sound source (Butler, Levy, & Neff, 1980; Lee, Cabrera, & Martens, 2012; Mershon, Desaulniers, Kiefer, Amerson, & Mills, 1981; Stecker & Hafter, 2000; Zahorik, Brungart, & Bronkhorst, 2005; but see, Zahorik & Wightman, 2001).Thus, the aim of the present article is to further address the predictive power of continuous intensity and loudness change in time-series models of perceived affect. This is achieved by manipulating continuous loudness perception without concurrent changes in the intensity profile of the stimulus; intensity changes that would normally be deemed necessary to manipulate loudness. Here, one group of participants continuously rated the affective elements of perceived arousal and valence (Experiment 1), while another participant group continuously rated loudness (Experiment 2). …