Abstract
A review of prior psychophysical and perceptual rendering literature suggests that perceptually lossless rendering, in which lossy renders are indistinguishable from their non-degraded counterparts, is a definite pos- sibility. Typically, perceptual rendering methods employ foveation, in which the region corresponding to the central field of view is rendered with higher fidelity than the region corresponding to the periphery. In this study, we contribute a straightforward foveated rendering method in Unreal Engine 4 (UE4) in order to gather insights on what may be holding back the adoption of these methods at a commercial level. Additionally, we contribute a conservative metric for calculating the foveal region size in the presence of system latency.
| Original language | English |
|---|---|
| Title of host publication | European Conference on Visual Media Production (CVMP) |
| Publication status | Published - 24 Nov 2015 |
Fingerprint
Cite this
Latency Aware Foveated Rendering in Unreal Engine 4. / Cosker, Darren; Swafford, Nicholas.
European Conference on Visual Media Production (CVMP). 2015.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Latency Aware Foveated Rendering in Unreal Engine 4
AU - Cosker, Darren
AU - Swafford, Nicholas
PY - 2015/11/24
Y1 - 2015/11/24
N2 - The human visual system is often assumed to be perfect despite limita- tions arising from a variety of different complexities and phenomena. Yet real-time rendering still operates on the assumption that a single render will be fully appreciated at any single point in time. With the increasing use of 4K-8K UHD displays and the push towards higher pixel densi- ties for head-mounted displays, the industry is pressured to meet market demands for intensive real-time rendering. The adoption of perceptually lossless rendering methods has never seemed more appropriate.A review of prior psychophysical and perceptual rendering literature suggests that perceptually lossless rendering, in which lossy renders are indistinguishable from their non-degraded counterparts, is a definite pos- sibility. Typically, perceptual rendering methods employ foveation, in which the region corresponding to the central field of view is rendered with higher fidelity than the region corresponding to the periphery. In this study, we contribute a straightforward foveated rendering method in Unreal Engine 4 (UE4) in order to gather insights on what may be holding back the adoption of these methods at a commercial level. Additionally, we contribute a conservative metric for calculating the foveal region size in the presence of system latency.
AB - The human visual system is often assumed to be perfect despite limita- tions arising from a variety of different complexities and phenomena. Yet real-time rendering still operates on the assumption that a single render will be fully appreciated at any single point in time. With the increasing use of 4K-8K UHD displays and the push towards higher pixel densi- ties for head-mounted displays, the industry is pressured to meet market demands for intensive real-time rendering. The adoption of perceptually lossless rendering methods has never seemed more appropriate.A review of prior psychophysical and perceptual rendering literature suggests that perceptually lossless rendering, in which lossy renders are indistinguishable from their non-degraded counterparts, is a definite pos- sibility. Typically, perceptual rendering methods employ foveation, in which the region corresponding to the central field of view is rendered with higher fidelity than the region corresponding to the periphery. In this study, we contribute a straightforward foveated rendering method in Unreal Engine 4 (UE4) in order to gather insights on what may be holding back the adoption of these methods at a commercial level. Additionally, we contribute a conservative metric for calculating the foveal region size in the presence of system latency.
M3 - Conference contribution
BT - European Conference on Visual Media Production (CVMP)
ER -