Diegetic and Object-Based Spatial Audio in Cinematic VR: A PRISMA-Guided Systematic Review with a Functional Taxonomy and Validation Framework

Vimala Perumal^1*Zeeshan Jawed Shah^1,2*

• ¹Multimedia University, Cyberjaya, Malaysia
• ²Prince Mohammad Bin Fahd University, Al Khobar, Saudi Arabia

Cinematic VR (CVR) removes the director’s frame, creating the challenge of guiding audience attention without breaking immersion. This systematic review examines two audio modalities with strong potential to act as narrative agents—diegetic audio (sounds from within the story world) and object-based spatial audio (discrete sound “objects” rendered with positional metadata)—to synthesize empirical evidence on how they guide attention, shape affect and presence, and to consolidate the methods used to validate these effects. Searches in IEEE Xplore, ACM Digital Library, Scopus, and Web of Science (June 2025) identified studies that used diegetic and/or object-based spatial audio as narrative devices in CVR with empirical user data; non-diegetic-only or purely technical papers without user measures were excluded except as baselines. Following PRISMA-2020, we qualitatively synthesized 18 studies employing behavioral (head/eye tracking), subjective (presence/engagement), and physiological (HR/EMG/EDA/PPG) measures. Across studies, world-locked, off-screen diegetic cues were repeatedly reported to redirect gaze and shorten time-to-ROI after cuts, while object-based rendering enabled precise, dynamic cue placement and was commonly associated with higher presence/immersion and affective arousal relative to non-spatial or head-locked baselines; however, methodological heterogeneity and small-to-moderate samples limit cross-study comparability and certainty. We contribute (i) a functional taxonomy of narrative audio techniques aligned to diegetic/object-based practice (Chion, 1994/2019; Rumsey, 2001/2012); (ii) a Validation Triangulation Framework integrating behavioral, subjective, and physiological evidence (Cacioppo, Tassinary, & Berntson, 2007; Laborde, Mosley, & Thayer, 2017); and (iii) a Minimum Reporting & Sharing Standard for CVR Narrative Audio (MRSS-CVR) specifying what to report, how to preregister, and how to share data/metadata in line with PRISMA 2020/PRISMA-S and FAIR principles (Page et al., 2021; Rethlefsen et al., 2021; Wilkinson et al., 2016). Limitations include small samples, inconsistent reporting, and limited direct measures of narrative comprehension. Overall, the available evidence suggests that diegetic and object-based spatial audio can support narrative guidance in CVR, but the strength of inference remains constrained by small samples, heterogeneous methods, and limited direct measures of narrative comprehension. The proposed taxonomy and validation framework are therefore positioned as evidence-informed, forward-looking tools to improve comparability and enable cumulative progress (PRISMA 2020 abstract guidance followed; no protocol registration).