Temporal Anti-Aliasing State of the Art

Information

• Publication Type: Student Project
• Workgroup(s)/Project(s):
  • Rendering and Modeling
• Date: 2023
• Date (Start): August 2020
• Date (End): May 2023
• Matrikelnummer: 08900070
• First Supervisor:
  • Johannes Unterguggenberger
  • Michael Wimmer
• Keywords: Real-Time Rendering, Temporal Anti-Aliasing, Vulkan, TAA, Auto-Vk, Auto-Vk-Toolkit, Vulkan

Abstract

In this student project most of the temporal anti-aliasing (TAA) techniques described in the TAA STAR report by Yang et al. [2020] were implemented. The main goal was to write a testing application that allows to experiment with different methods and parameters in varying scenes. The different methods broadly fall into two categories: Acquisition of the temporal samples, and validation and rectification of history data. The topic of temporal upsampling was also briefly experimented with, but not followed up in depth. Additionally, TAA with adaptive ray tracing, as proposed by Marrs et al. [2019], was implemented.

Additional Files and Images

Weblinks

• TAA-STAR on GitHub
  Code repository that contains the implementations, based on Auto-Vk-Toolkit.
• Auto-Vk-Toolkit on GitHub
  Programming framework
• [Marrs2019]
  Adam Marrs, Josef Spjut, Holger Gruen, Rahul Sathe, and Morgan McGuire. Improving Temporal Antialiasing with Adaptive Ray Tracing. Apress, Berkeley, CA, USA, 2019.
• [Yang2020]
  Lei Yang, Shiqiu Liu, and Marco Salvi. A Survey of Temporal Antialiasing Techniques. Computer Graphics Forum, 2020.
  

BibTeX

@studentproject{cech-2023-taa,
  title =      "Temporal Anti-Aliasing State of the Art",
  author =     "Alexander Cech",
  year =       "2023",
  abstract =   "In this student project most of the temporal anti-aliasing
               (TAA) techniques described in the TAA STAR report by Yang et
               al. [2020] were implemented. The main goal was to write a
               testing application that allows to experiment with different
               methods and parameters in varying scenes. The different
               methods broadly fall into two categories: Acquisition of the
               temporal samples, and validation and rectification of
               history data. The topic of temporal upsampling was also
               briefly experimented with, but not followed up in depth.
               Additionally, TAA with adaptive ray tracing, as proposed by
               Marrs et al. [2019], was implemented. ",
  month =      may,
  keywords =   "Real-Time Rendering, Temporal Anti-Aliasing, Vulkan, TAA,
               Auto-Vk, Auto-Vk-Toolkit, Vulkan",
  URL =        "https://www.cg.tuwien.ac.at/research/publications/2023/cech-2023-taa/",
}