A hardware architecture for surface splatting

Abstract

We present a novel architecture for hardware-accelerated rendering of point primitives. Our pipeline implements a refined version of EWA splatting, a high quality method for antialiased rendering of point sampled representations. A central feature of our design is the seamless integration of the architecture into conventional, OpenGL-like graphics pipelines so as to complement triangle-based rendering. The specific properties of the EWA algorithm required a variety of novel design concepts including a ternary depth test and using an on-chip pipelined heap data structure for making the memory accesses of splat primitives more coherent. In addition, we developed a computationally stable evaluation scheme for perspectively corrected splats. We implemented our architecture both on reconfigurable FPGA boards and as an ASIC prototype, and we integrated it into an OpenGL-like software implementation. Our evaluation comprises a detailed performance analysis using scenes of varying complexity.

Motion Blur for EWA Surface Splatting

“…One of the possibilities to address this issue would be an implementation which exploits the GPU for general purpose programming. Another approach could compute the visibility function much more accurately by using specialized hardware extensions to current GPUs which support a reduced A‐Buffer or by employing an extended dedicated hardware architecture such as [WHA*07]. As our current implementation uses a constant number of samples per frame, the sampling could be varied adaptively for each point based on its motion and shading changes over time in a future work.…”

“…EWA splatting has been implemented on conventional GPUs either by employing textured quads to approximate the points [RPZ02] or by making use of advanced features of GPUs [ZRB*04, BHZK05, ZP06, GBP06]. Recently, a dedicated hardware architecture for EWA surface splatting has been presented [WHA*07, HSA*08]. Reconstruction methods for rendering point sets [ABCO*01, GG07] have been investigated as well, either through raytracing [AKP*05] or a dynamic upsampling approach which directly splats the generated samples [GGG08].…”

OMiCroN – Oblique Multipass Hierarchy Creation while Navigating

“…Finally, the splats in the rendering queue are used as input for the traditional two-pass EWA filter described in [38]. Several methods for computing the sizes of the projected splats were tested [38,39,40,41]. The splat bounding box computation algorithm described in [41] resulted in the best performance-quality relationship and all results reported in this paper applied it.…”

Efficient image reconstruction for point-based and line-based rendering

“…Apart from these multipass splatting methods, several alternative point-rendering techniques have been presented; in particular, a ''single-pass'' splatting technique [36,37], dedicated splatting hardware [38], and ray tracing of points [39][40][41]. The ''single-pass'' splatting technique proposed by Zhang and Pajarola [36,37] has to compute groups of non-overlapping splats (in object space) and is therefore less suitable for large, dynamic point sets.…”

“…The ''single-pass'' splatting technique proposed by Zhang and Pajarola [36,37] has to compute groups of non-overlapping splats (in object space) and is therefore less suitable for large, dynamic point sets. While the dedicated hardware published by Weyrich et al [38] achieves promising performance results, the prototypical hardware cannot provide the rendering performance of splatting techniques implemented on today's GPUs. Ray tracing of points was first published by Schaufler and Jensen [39] and led to work by Adamson and Alexa [40], and Wald and Seidel [41].…”