Fig 1.
Local and SSAO rendering of an atomic structure of a maltoporin protein (PDB ID 1AF6).
(a) Standard local lighting, which provides only a locally acceptable approximation of real world lighting. (b) Depth cueing, which unevenly shades the three (slightly tilted) channels and looks unrealistic at the bottom. (c-d) SSAO, at two different sampling sizes that emphasize spatial features according to user preference. The protein is shown in van der Waals mode in an orientation corresponding to Fig 2 in [16]. All molecular graphics figures in this paper were created with Sculptor version 2.1 [17, 18].
Fig 2.
Global illumination by ambient light rays (red) emanating from a hemisphere Ω.
The illumination is governed by a weighted sum of unblocked light rays (red) that reach the surface point p, where the weights are given by the cosine of the incident angle relative to the surface normal np. [16, 21, 22].
Fig 3.
Volumetric representation of a 2.2 Å resolution cryo electron microscopy density map of beta-galactosidase in complex with a cell-permeant inhibitor (EMDB ID 2984).
(a) Standard local lighting. (b-d) SSAO, at three different sampling sizes that emphasize tertiary (b), secondary (c), and primary (d) spatial features according to user preference.
Fig 4.
Illustration of buffer contents.
(a) Color buffer. (b) Corresponding depth buffer contents in grayscale (black is near and white is far).
Fig 5.
Illustration of SSAO approaches.
(a-b) Point-based SSAO for (a) concavity and (b) convexity. Red marks show sample points behind the surface, green ones in front of it. In (a), most points are behind the surface, resulting in high AO. In (b), most points are in front of the surface, resulting in low AO. (c) Line-based SSAO. The green parts of the lines are in front of the surface; the red parts are hidden. Computing the ratio of visible vs. hidden parts yields an AO factor of higher granularity than what can be achieved by the point-based method.
Fig 6.
(a) AO with strong banding artifacts. (b) Randomized sampling. (c) Smoothing applied after the randomization.
Fig 7.
Three-way comparison of sampling size effect.
Shown is a composite of CCMV images (PDB ID 1CWP) with three different heuristic sampling diameter settings. Adaption of the diameter highlights structural arrangements on different scales.