lecture18

# lecture18 - Visibility Culling Visibility David Luebke...

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Visibility Culling Visibility David Luebke Computer Science Department University of Virginia <[email protected]> Recap: General Occlusion Culling • When cells and portals don’t work… – Trees in a forest – A crowded train station • Need general occlusion culling algorithms: – Aggregate occlusion – Dynamic scenes – Non­polygonal scenes D2 Recap: General Occlusion Culling • I’ll discuss two algorithms: – Hierarchical Z­Buffer • Ned Greene, SIGGRAPH 93 – Hierarchical Occlusion Maps • Hansong Zhang, SIGGRAPH 97 D3 Recap: Hierarchical Z­Buffer • Replace Z­buffer with a Z­pyramid – Lowest level: full­resolution Z­buffer – Higher levels: each pixel represents the maximum depth of the four pixels “underneath” it • Basic idea: hierarchical rasterization of the polygon, with early termination where polygon is occluded D4 Hierarchical Z­Buffer • Idea: test polygon against highest level first – If polygon is further than distance recorded in pixel, stop—it’s occluded – If polygon is closer, recursively check against next lower level – If polygon is visible at lowest level, set new distance value and propagate up D5 Hierarchical Z­Buffer • Z­pyramid exploits image­space coherence: – Polygon occluded in a pixel is probably occluded in nearby pixels • HZB also exploits object­space coherence – Polygons near an occluded polygon are probably occluded D6 Hierarchical Z­Buffer • Exploiting object­space coherence: – Subdivide scene with an octree – All geometry in an octree node is contained by a cube – Before rendering the contents of a node, “render” the faces of its cube (i.e., query the Z­pyramid) – If cube faces are occluded, ignore the entire node D7 Hierarchical Z­Buffer • HZB can exploit temporal coherence – Most polygons affecting the Z­buffer last frame will affect Z­buffer this frame – HZB also operates at max efficiency when Z­pyramid already built • So start each frame by rendering octree nodes visible last frame D8 Hierarchical Z­Buffer: Discussion • • HZB needs hardware support to be really competitive To date, hardware vendors haven’t bought in: – Z­pyramid (and hierarchies in general) unfriendly to hardware – Unpredictable Z­query times generate bubbles in rendering pipe • But there is a promising trend… D9 Hierarchical Z­Buffer • Hardware beginning to support Z­query operation – Allows systems to exploit: • • Object­space coherence (bounding boxes) Temporal coherence (last­rendered list) – Systems I’m aware of: • • HP Visualize­fx graphics SGI Visual Workstation products – An aside: applies to cell­portal culling! D10 Hierarchical Occlusion Maps • • A more hardware­friendly general occlusion culling algorithm Two major differences from HZB: – Separates occluders from occludees – Decouples occlusion test into an depth test and a overlap test D11 Hierarchical Occlusion Maps • Occluders versus occludees: Blue parts: occluders Blue occluders Red parts: occludees occludees D12 Hierarchical Occlusion Maps • Depth versus overlap: View Point Z X Y Depth + Overlap = Occlusion D13 Hierarchical Occlusion Maps • Representation of projection for overlap test: occlusion map – Corresponds to a screen subdivision – Records average opacity per partition • Generate by rendering occluders – Record pixel opacities (i.e., coverage) D14 Occlusion Maps Rendered Image Occlusion Map D15 Occlusion Map Pyramid 64 x 64 32 x 32 16 x 16 D16 Occlusion Map Pyramid D17 Occlusion Map Pyramid • Analyzing cumulative projection: – A hierarchical occlusion map (HOM) – Generate by recursive averaging (once per frame) – Records average opacities for blocks of multiple pixels, representing occlusion at multiple resolutions – Construction can be accelerated by texture hardware D18 Overlap Tests • Query: is projection of occludee inside cumulative projection of occluders? – Cumulative projection: occlusion pyramid – Ocludee projection: expensive in general • • Overestimate ocludee with 3­D bounding box Overestimate projection of 3­D bounding box with 2­D bounding rectangle in screen­space D19 Overlap Tests • Hierarchical structure enables some optimizations: – Predictive rejection • Terminate test when it must fail later – Conservative rejection • The transparency threshold – Aggressive Approximate Culling • • Ignore objects barely visible through holes The opacity threshold D20 Aggressive Approximate Culling 0 1 2 3 4 D21 Hierarchical Occlusion Maps • Not discussed here: – Depth test • • Depth estimation buffer Modified Z­buffer – Selecting occluders • For more details, see attached excerpt from Hansong Zhang’s dissertation D22 HOM: Discussion • Provides a robust, general, hardware­ friendly occlusion culling algorithm – Supports dynamic scenes – Supports non­polygonal geometry – Not many hardware assumptions D23 HOM: Discussion • • Efficient coding, careful tuning a must Fairly high per­frame overhead – Needs high depth complexity, good occluder selection to be worthwhile – UNC’s MMR system: D24 Visibility Culling: Discussion • When is visibility culling worthwhile? – When scene has high depth complexity • • Examples: architectural walkthroughs, complex CAD assemblies, dense forest Non­examples: terrain, single highly­tesselated object (e.g., a radiositized room) D25 Visibility Culling: Discussion • How does visibility culling compare to: – Level­of­detail: • • Reduces geometry processing Helps transform­bound apps – Visibility culling: • • Reduces geometry and pixel processing Helps transform­ and fill rate­bound apps – Texture / Image representations: • • Reduces geometry and pixel processing Incurs texture/image processing costs D26 Visibility Culling: Discussion • How does visibility culling interact with level of detail? – Fairly seamless integration; generally a win – One issue: visibility of simplified model may differ from original model; requires some care – LODs can speed up occluder selection and rendering D27 Visibility Culling: Discussion • How does visibility culling interact with texture and image­based representations? – Texture/image reps generally replace far­field geometry • • • Involves an implicit occlusion culling step Reduces scene depth complexity, decreasing the utility of visibility culling If near­field geometry still includes complex heavily­ occlusive assemblies, still a win D28 Visibility Culling: Discussion • How much culling effort is appropriate? – Cells and portals: relatively cheap, with large potential speedups – Hierarchical occlusion maps: relatively costly, carefully weigh potential gains – Multiple processors allow much more aggressive culling calculation • • Pipelining culling calculations, Performer­style, allows cull time = render time Tradeoff: one frame increased latency D29 Summary • The basic, very powerful idea: – Rapidly compute a potentially visible set – Let hardware handle the rest • For many scenes, visibility culling is a simple way to get huge speedups – View­frustum culling always a must – For scenes with high depth complexity, occlusion culling can be a big win D30 Summary • Architectural models: visibility is practically a solved problem – Cells and portals work well • • Cull­box portal culling: simple, fast Line­stabbing: elegant, powerful D31 Summary • Occlusion culling of general models: still a largely open problem – Important issues: • • Dynamic scenes Aggregate occlusion effects D32 Summary • General occlusion culling algorithms: – Hierarchical Z­buffer: • • A simple, truly elegant algorithm But doesn’t seem amenable to hardware – Hierarchical occlusion maps: • • • • More complicated, difficult to code & tune Better suited to current hardware Lends itself well to aggressive culling Fairly high overhead, only worthwhile with high depth complexity – Promising trend in hardware D33 ...
View Full Document

## This note was uploaded on 12/09/2011 for the course CS 561/661 taught by Professor Lubke during the Summer '11 term at Virginia Tech.

Ask a homework question - tutors are online