EG2014 - Short Papers

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/7706

Browse

Browsing EG2014 - Short Papers by Subject "I.3.7 [Computer Graphics]"

Now showing 1 - 5 of 5
  • Thumbnail Image
    Item
    Deferred Shading for Order-Independent Transparency
    (The Eurographics Association, 2014) Hillesland, Karl E.; Bilodeau, Bill; Thibieroz, Nicolas; Eric Galin and Michael Wand
    Rendering many layers of transparency presents difficult challenges with respect to performance. Most previous work focused on the sorting problem, paying the full shading cost for all fragments. We present a method that defers shading until fragments can be classified as less important to the final pixel color, allowing us to switch to a lower-cost, approximate shading function. We apply this idea to TressFX, which is a state-of-the-art hair rendering technique used in video game production. For hair rendering, we switched to low-quality shading in all but the front eight fragments per pixel. This gave us a 75% speedup without noticeable loss in visual quality.
  • Thumbnail Image
    Item
    Latency Considerations of Depth-first GPU Ray Tracing
    (The Eurographics Association, 2014) Guthe, Michael; Eric Galin and Michael Wand
    Despite the potential divergence of depth-first ray tracing [AL09], it is nevertheless the most efficient approach on massively parallel graphics processors. Due to the use of specialized caching strategies that were originally developed for texture access, it has been shown to be compute rather than bandwidth limited. Especially with recents developments however, not only the raw bandwidth, but also the latency for both memory access and read after write register dependencies can become a limiting factor. In this paper we will analyze the memory and instruction dependency latencies of depth first ray tracing. We will show that ray tracing is in fact latency limited on current GPUs and propose three simple strategies to better hide the latencies. This way, we come significantly closer to the maximum performance of the GPU.
  • Thumbnail Image
    Item
    Merging Live and pre-Captured Data to support Full 3D Head Reconstruction for Telepresence
    (The Eurographics Association, 2014) Fleury, Cedric; Popa, Tiberiu; Cham, Tat Jen; Fuchs, Henry; Eric Galin and Michael Wand
    This paper proposes a 3D head reconstruction method for low cost 3D telepresence systems that uses only a single consumer level hybrid sensor (color+depth) located in front of the users. Our method fuses the real-time, noisy and incomplete output of a hybrid sensor with a set of static, high-resolution textured models acquired in a calibration phase. A complete and fully textured 3D model of the users head can thus be reconstructed in real-time, accurately preserving the facial expression of the user. The main features of our method are a mesh interpolation and a fusion of a static and a dynamic textures to combine respectively a better resolution and the dynamic features of the face.
  • Thumbnail Image
    Item
    Regional Time Stepping for SPH
    (The Eurographics Association, 2014) Goswami, Prashant; Batty, Christopher; Eric Galin and Michael Wand
    This paper presents novel and efficient strategies to spatially adapt the amount of computational effort applied based on the local dynamics of a free surface flow, for classic weakly compressible SPH (WCSPH). Using a convenient and readily parallelizable block-based approach, different regions of the fluid are assigned differing time steps and solved at different rates to minimize computational cost. We demonstrate that our approach can achieve about two times speed-up over the standard method even in highly dynamic scenes.
  • Thumbnail Image
    Item
    A User Study: Is the Advection Step in Shallow Water Equations Really Necessary?
    (The Eurographics Association, 2014) Kellomäki, Timo; Saari, Timo; Eric Galin and Michael Wand
    Heightfield methods, such as the pipe method and shallow water equations (SWE), have often been used to simulate large areas of water. Of these, the SWE are often preferred due to being more realistic, but they are also more complex and demand more computational resources than the pipe method. These two methods were presented to over 40 subjects in both a gaming and a video context to see whether they report noticing the advantages of SWE compared to the pipe method. No significant differences were observed in any of the categories measured (hedonic valence, flow, spatial presence, realism). Therefore, at least considering using the pipe method instead of the SWE is recommended. Also, varying the time step between 5 and 20 ms did not affect the user experience.