Social Virtual Reality (SVR) systems are growing in both popularity and breadth. New systems are attempting to support a wider range of increasingly large social situations. A key characteristic of SVR systems is user capacity. Many systems are designed for small group discussion or team games, and support up to 20-30 users. Some support larger counts, but via special features coupled to a specific social situation. As SVR broaches medium to large gatherings (10s-1000s of people), capacity becomes a major challenge. Naïve implementations where each user communicates directly with every other will quickly overwhelm the resources of the system. The constraints on scalability have been known since the first SVR systems were created. However, while there are many systems out there, their scalability mechanisms are often application specific, and rarely openly discussed. In this paper we explore scalability in the open-source SVR Ubiq. We consider two partitioning schemes to increase relay-server capacity. We examine the current failure modes of the relay-server system, and demonstrate how these schemes improve capacity by a factor of 2-3x. We look at the interactions between the schemes and simulated user behaviour, to see what lessons can be gleaned for research into scalability for SVR.

Image-warping, a per-pixel deformation of one image into another, is an essential component in immersive visual experiences such as virtual reality or augmented reality. The primary issue with image warping is disocclusions, where occluded (and hence unknown) parts of the input image would be required to compose the output image. We introduce a new image warping method, Metameric image inpainting - an approach for hole-filling in real-time with foundations in human visual perception. Our method estimates image feature statistics of disoccluded regions from their neighbours. These statistics are inpainted and used to synthesise visuals in real-time that are less noticeable to study participants, particularly in peripheral vision. Our method offers speed improvements over the standard structured image inpainting methods while improving realism over colour-based inpainting such as push-pull. Hence, our work paves the way towards future applications such as depth image-based rendering, 6-DoF 360 rendering, and remote render-streaming.

Distributed virtual environments (DVEs) are challenging to create as the goals of consistency and responsiveness become contradictory under increasing latency. DVEs have been considered as both distributed transactional databases and force-reflection systems. Both are good approaches, but they do have drawbacks. Transactional systems do not support Level 3 (L3) collaboration: manipulating the same degree-of-freedom at the same time. Force-reflection requires a client-server architecture and stabilisation techniques. With Consensus Based Networking (CBN), we suggest DVEs be considered as a distributed data-fusion problem. Many simulations run in parallel and exchange their states, with remote states integrated with continous authority. Over time the exchanges average out local differences, performing a distribued-average of a consistent, shared state. CBN aims to build simulations that are highly responsive, but consistent enough for use cases such as the piano-movers problem. CBN's support for heterogeneous nodes can transparently couple different input methods, avoid the requirement of determinism, and provide more options for personal control over the shared experience. Our work is early, however we demonstrate many successes, including L3 collaboration in room-scale VR, 1000's of interacting objects, complex configurations such as stacking, and transparent coupling of haptic devices. These have been shown before, but each with a different technique; CBN supports them all within a single, unified system.

While they have long been a subject of academic study, social virtual reality (SVR) systems are now attracting increasingly large audiences on current consumer virtual reality systems. The design space of SVR systems is very large, and relatively little is known about how these systems should be constructed in order to be usable and efficient. In this paper we present Ubiq, a toolkit that focuses on facilitating the construction of SVR systems. We argue for the design strategy of Ubiq and its scope. Ubiq is built on the Unity platform. It provides core functionality of many SVR systems such as connection management, voice, avatars, etc. However, its design remains easy to extend. We demonstrate examples built on Ubiq and how it has been successfully used in classroom teaching. Ubiq is open source (Apache License) and thus enables several use cases that commercial systems cannot.

To peripheral vision, a pair of physically different images can look the same. Such pairs are metamers relative to each other, just as physically-different spectra of light are perceived as the same color. We propose a real-time method to compute such ventral metamers for foveated rendering where, in particular for near-eye displays, the largest part of the framebuffer maps to the periphery. This improves in quality over state-of-the-art foveation methods which blur the periphery. Work in Vision Science has established how peripheral stimuli are ventral metamers if their statistics are similar. Existing methods, however, require a costly optimization process to find such metamers. To this end, we propose a novel type of statistics particularly well-suited for practical real-time rendering: smooth moments of steerable filter responses. These can be extracted from images in time constant in the number of pixels and in parallel over all pixels using a GPU. Further, we show that they can be compressed effectively and transmitted at low bandwidth. Finally, computing realizations of those statistics can again be performed in constant time and in parallel. This enables a new level of quality for foveated applications such as such as remote rendering, level-of-detail and Monte-Carlo denoising. In a user study, we finally show how human task performance increases and foveation artifacts are less suspicious, when using our method compared to common blurring.

Mobile HMDs must sacrifice compute performance to achieve ergonomic and power requirements for extended use. Consequently, applications must either reduce rendering and simulation complexity - along with the richness of the experience - or offload complexity to a server. Within the context of edge-computing, a popular way to do this is through render streaming. Render streaming has been demonstrated for desktops and consoles. It has also been explored for HMDs. However, the latency requirements of head tracking make this application much more challenging. While mobile GPUs are not yet as capable as their desktop counterparts, we note that they are becoming more powerful and efficient. With the hard requirements of VR, it is worth continuing to investigate what schemes could optimally balance load, latency and quality. We propose an alternative we call edge-physics: streaming at the scene-graph level from a simulation running on edge-resources, analogous to cluster rendering. Scene streaming is not only straightforward, but compute and bandwidth efficient. The most demanding loops run locally. Jobs that hit the power-wall of mobile CPUs are off-loaded, while improving GPUs are leveraged, maximising compute utilisation. In this paper we create a prototypical implementation and evaluate its potential in terms of fidelity, bandwidth and performance. We show that an effective system which maintains high consistencies on typical edge-links can be easily built, but that some traditional concepts are not applicable, and a better understanding of the perception of motion is required to evaluate such a system comprehensively.

Grounded haptic devices can provide a variety of forces but have limited working volumes. Wearable haptic devices operate over a large volume but are relatively restricted in the types of stimuli they can generate. We propose the concept of docking haptics, in which different types of haptic devices are dynamically docked at run time. This creates a hybrid system, where the potential feedback depends on the user’s location. We show a prototype docking haptic workspace, combining a grounded six degree-of-freedom force feedback arm with a hand exoskeleton. We are able to create the sensation of weight on the hand when it is within reach of the grounded device, but away from the grounded device, hand-referenced force feedback is still available. A user study demonstrates that users can successfully discriminate weight when using docking haptics, but not with the exoskeleton alone. Such hybrid systems would be able to change configuration further, for example docking two grounded devices to a hand in order to deliver twice the force, or extend the working volume. We suggest that the docking haptics concept can thus extend the practical utility of haptics in user interfaces.

We suggest a rasterization pipeline tailored towards the needs of HMDs, where latency and field-of-view requirements pose new challenges beyond those of traditional desktop displays. Instead of image warping for low latency, or using multiple passes for foveation, we show how both can be produced directly in a single perceptual rasterization pass. We do this with per-fragment ray-casting. This is enabled by derivations of tight space-time-fovea pixel bounds, introducing just enough flexibility for the requisite geometric tests, but retaining most of the simplicity and efficiency of the traditional rasterizaton pipeline. To produce foveated images, we rasterize to an image with spatially varying pixel density. To compensate for latency, we extend the image formation model to directly produce "rolling" images where the time at each pixel depends on its display location. Our approach overcomes limitations of warping with respect to disocclusions, object motion and view-dependent shading, as well as geometric aliasing artifacts in other foveated rendering techniques. A set of perceptual user studies demonstrates the efficacy of our approach.

Many techniques facilitate real-time collision detection against complex models. These typically work by pre-computing information about the spatial distribution of geometry into a form that can be quickly queried. When models deform though, expensive pre-computations are impractical. We present radial fields: a variant of distance fields parameterised in cylindrical space, rather than Cartesian space. This 2D parameterisation significantly reduces the memory and computation requirements of the field, while introducing minimal overhead in collision detection tests. The interior of the mesh is defined implicitly for the entire domain. Importantly, it maps well to the hardware rasteriser of the GPU. Radial fields are much more application-specific than traditional distance fields. For these applications - such as collision detection with articulated characters-however, the benefits are substantial.

Real-time interactive systems such as virtual environments have high performance requirements, and profiling is a key part of the optimisation process to meet them. Traditional techniques based on metadata and static analysis have difficulty following causality in asynchronous systems. In this paper we explore a new technique for such systems. Timestamped samples of the system state are recorded at instrumentation points at runtime. These are assembled into a graph, and edges between dependent samples recovered. This approach minimises the invasiveness of the instrumentation, while retaining high accuracy. We describe how our instrumentation can be implemented natively in common environments, how its output can be processed into a graph describing causality, and how heterogeneous data sources can be incorporated into this to maximise the scope of the profiling. Across three case studies, we demonstrate the efficacy of this approach, and how it supports a variety of metrics for comprehensively bench-marking distributed virtual environments.

Video see-through augmented reality (VSAR) is an effective way of combing real and virtual scenes for head-mounted human computer interfaces. In this paper we present the AR-Rift 2 system, a cost-effective prototype VSAR system based around the Oculus Rift CV1 head-mounted display (HMD). Current consumer camera systems however typically have latencies far higher than the rendering pipeline of current consumer HMDs. They also have lower update rate than the display. We thus measure the latency of the video and implement a simple image-warping method to ensure smooth movement of the video.

Latency is detrimental to interactive systems, especially pseudo-physical systems that emulate real-world behaviour. It prevents users from making quick corrections to their movement, and causes their experience to deviate from their expectations. Latency is a result of the processing and transport delays inherent in current computer systems. As such, while a number of studies have hypothesized that any latency will have a degrading effect, few have been able to test this for latencies less than ∼ 50 ms. In this study we investigate the effects of latency on pointing and steering tasks. We design an apparatus with a latency lower than typical interactive systems, using it to perform interaction tasks based on Fitts's law and the Steering law. We find evidence that latency begins to affect performance at ∼ 16 ms, and that the effect is non-linear. Further, we find latency does not affect the various components of an aiming motion equally. We propose a three stage characterisation of pointing movements with each stage affected independently by latency. We suggest that understanding how users execute movement is essential for studying latency at low levels, as high level metrics such as total movement time may be misleading.

Latency - the delay between a user's action and the response to this action - is known to be detrimental to virtual reality. Latency is typically considered to be a discrete value characterising a delay, constant in time and space - but this characterisation is incomplete. Latency changes across the display during scan-out, and how it does so is dependent on the rendering approach used. In this study, we present an ultra-low latency real-time ray-casting renderer for virtual reality, implemented on an FPGA. Our renderer has a latency of ~1 ms from 'tracker to pixel'. Its frameless nature means that the region of the display with the lowest latency immediately follows the scan-beam. This is in contrast to frame-based systems such as those using typical GPUs, for which the latency increases as scan-out proceeds. Using a series of high and low speed videos of our system in use, we confirm its latency of ~1 ms. We examine how the renderer performs when driving a traditional sequential scan-out display on a readily available HMO, the Oculus Rift OK2. We contrast this with an equivalent apparatus built using a GPU. Using captured human head motion and a set of image quality measures, we assess the ability of these systems to faithfully recreate the stimuli of an ideal virtual reality system - one with a zero latency tracker, renderer and display running at 1 kHz. Finally, we examine the results of these quality measures, and how each rendering approach is affected by velocity of movement and display persistence. We find that our system, with a lower average latency, can more faithfully draw what the ideal virtual reality system would. Further, we find that with low display persistence, the sensitivity to velocity of both systems is lowered, but that it is much lower for ours.

Consumer virtual reality systems are now becoming widely available. We report on a study on presence and embodiment within virtual reality that was conducted 'in the wild', in that data was collected from devices owned by consumers in uncontrolled settings, not in a traditional laboratory setting. Users of Samsung Gear VR and Google Cardboard devices were invited by web pages and email invitation to download and run an app that presented a scenario where the participant would sit in a bar watching a singer. Each participant saw one of eight variations of the scenario: with or without a self-avatar; singer inviting the participant to tap along or not; singer looking at the participant or not. Despite the uncontrolled situation of the experiment, results from an in-app questionnaire showed tentative evidence that a self-avatar had a positive effect on self-report of presence and embodiment, and that the singer inviting the participant to tap along had a negative effect on self-report of embodiment. We discuss the limitations of the study and the platforms, and the potential for future open virtual reality experiments.

Reconfigurable hardware has been used before for low latency image synthesis. These are typically low level implementations with tight vertical integration. For example the apparatus of both Regan et al and Ng et al had the tracker driven by the same device performing the rendering. Reconfigurable hardware combined with the dataflow programming model can make application specific rendering hardware cost effective. Our sprite renderer has comparable scope to both prior examples, but our dataflow graph can be adapted to other use cases with an effort comparable to GPU shader programming.

Latency of interactive computer systems is a product of the processing, transport and synchronisation delays inherent to the components that create them. In a virtual environment (VE) system, latency is known to be detrimental to a user's sense of immersion, physical performance and comfort level. Accurately measuring the latency of a VE system for study or optimisation, is not straightforward. A number of authors have developed techniques for characterising latency, which have become progressively more accessible and easier to use. In this paper, we characterise these techniques. We describe a simple mechanical simulator designed to simulate a VE with various amounts of latency that can be finely controlled (to within 3ms). We develop a new latency measurement technique called Automated Frame Counting to assist in assessing latency using high speed video (to within 1ms). We use the mechanical simulator to measure the accuracy of Steed's and Di Luca's measurement techniques, proposing improvements where they may be made. We use the methods to measure latency of a number of interactive systems that may be of interest to the VE engineer, with a significant level of confidence. All techniques were found to be highly capable however Steed's Method is both accurate and easy to use without requiring specialised hardware.