Light-DPCVR: A dataset on the impact of light on dynamic point cloud quality and user behavior in virtual reality

Visual quality assessment of dynamic point clouds has become an important matter for emerging use cases and applications in eXtended Reality (XR). Understanding how users perceive quality in XR is essential to enable improvements in compression and rendering processes that will improve the Quality of Experience (QoE) for end users. While numerous studies have investigated the perceptual quality of static and dynamic point clouds, very few have examined influential factors exclusive to volumetric media, such as rendering parameters and lighting, and how they impact the perceived quality and user experience in the context of XR applications with six degrees of freedom (6DoF) interaction. In this paper, we present Light-DPCVR: a subjective dataset collected through a quality assessment experiment in Virtual Reality (VR) complemented by gaze and exploration tracking. This study investigates the role of virtual lighting direction, a rendering parameter exclusive to 3D content, in the perceived quality of rendered dynamic point clouds coded with Video-based Point Cloud Compression (V-PCC) in VR, and its influence on observer behavior. Our results reveal how lighting affects the perception of V-PCC artifacts and user exploration. Furthermore, we assess the efficiency of our test methodology for volumetric videos in VR, and build a benchmark of objective quality metrics performance on our dataset and give access to our data to enlighten future development.