Cone-Traced Supersampling for Signed Distance Field Rendering

Andrei Chubarau, YANGYANG ZHAO, Ruby Rao, Paul Kry, Derek Nowrouzezahrai

Keywords: graphics, rendering, signed distance fields, antialiasing, supersampling, sphere tracing

TL;DR: We present cone-traced supersampling (CTSS), an efficient and robust spatial antialiasing solution that naturally complements the sphere tracing algorithm used to render Signed Distance Fields (SDFs).

Abstract: While Signed Distance Fields (SDFs) in theory offer infinite level of detail, they are typically rendered using the sphere tracing algorithm at finite resolutions, which causes the common rasterized image synthesis problem of aliasing. Most existing optimized antialiasing solutions rely on polygon mesh representations; SDF-based geometry can only be directly antialiased with the computationally expensive supersampling or with post-processing filters that often lead to undesirable blurriness and ghosting. In this work, we present cone-traced supersampling (CTSS), an efficient and robust spatial antialiasing solution that naturally complements the sphere tracing algorithm, does not require casting additional rays per pixel or offline pre-filtering, and can be easily implemented in existing real-time SDF renderers. CTSS performs supersampling along the traced ray near surfaces with partial visibility identified by evaluating cone intersections within a pixel's view frustum. We further devise a specialized sampling strategy to minimize the number of shading computations and aggregate the collected samples based on their correlated visibility. Depending on configuration, CTSS incurs roughly 15-30% added computational cost and significantly outperforms conventional supersampling approaches while offering comparative antialiasing and visual image quality for most geometric edges.

Track: Graphics

Revision: No

Revision Reviewers: No opinion

Summary Of Changes:

We considered all of the comments left by the reviewers and included as many possible suggestions as we deemed possible.
[Relevance to reviews indicated in square brackets]

Changelog:

Teaser figure
-Renamed scene "Simple Sponza" to "Sponza SDF" (also name changed in Table 1)
-Added in text reference to teaser figure

Abstract
-Moved footnote to the 2nd column to avoid splitting the abstract into two columns

Sec 3.3
-Split first paragraph [all reviewers: clarity]

Sec 3.4
-Split bitmask paragraph; added phrase "Pixel footprint is given by the traced cone with radius $R_c$, hence the mask spans $2 R_c$." [R1: implementation detail for visibility mask]
-Added short note on performance of visibility bitmask computations vs LUT approach [R1: performance]
-Added new figure: "Figure 5: Visibility bitmask encoded as a 32-bit integer." [R1: implementation detail for visibility mask]
-Split visibility mask implementation details into new subsection; New subsection "3.5 Visibility Bitmask Implementation" [R1, R2]
-Changed ordering/position of affected Figures

Sec 4
-Minor rephrasing to last paragraph (pseudocode) [clarity]
-Added hyperlink to footnote on page 1 with Shadertoy link [R1 did not see this]

Sec 4.1
-Removed spacing between subfigures and increased their size in "Figure 9: Affected pixels for CTSS and CTSS-R"
-Table 1, Renamed scene "Simple Sponza" to "Sponza SDF" (also name changed in Teaser fig)
-Table 1, added footnote for "Sponza SDF" to provide link to the used Sponza SDF implementation (https://github.com/mmerchante/sdf-gen-unity)

Sec 5
-Increased to line spacing in algorithm 1 [clarity]
-Decreased font size of comments in algorithm 1 [clarity]

Sec 5.2
-Clarified limitations of CTSS under "Antialiasing" [all reviewers]
-Added new figure: "Figure 10: Limitations of CTSS." [all reviewers]
-Added to the discussion under "Applicability" [all reviewers]
-Added short note on applicability to non-Eikonal SDFs [R2]

Sec 5.3
-Clarified future work directions [all reviewers]
-Added paragraph about CTSS and transparency [R1]