A Comparison of Rendering Techniques for 3D Line Sets With Transparency

Abstract

This paper presents a comprehensive study of interactive rendering techniques for large 3D line sets with transparency. The rendering of transparent lines is widely used for visualizing trajectories of tracer particles in flow fields. Transparency is then used to fade out lines deemed unimportant, based on, for instance, geometric properties or attributes defined along them. Since accurate blending of transparent lines requires rendering the lines in back-to-front or front-to-back order, enforcing this order for 3D line sets with tens or even hundreds of thousands of elements becomes challenging. In this paper, we study CPU and GPU rendering techniques for large transparent 3D line sets. We compare accurate and approximate techniques using optimized implementations and a number of benchmark data sets. We discuss the effects of data size and transparency on quality, performance and memory consumption. Based on our study, we propose two improvements to per-pixel fragment lists and multi-layer alpha blending. The first improves the rendering speed via an improved GPU sorting operation, and the second improves rendering quality via a transparency-based bucketing.

Interactive Focus+Context Rendering for Hexahedral Mesh Inspection

“…This enables opacity-based blending, i.e., α-blending, of fragments in the correct visibility order. For sorting, we use a GPU-friendly implementation of priority queues [32].…”

“…[31] decrease the opacity of nonimportant foreground lines using per-frame opacity optimization. A summary and evaluation of different GPU transparency rendering techniques for large line sets is given by [32]. [33] build a line hierarchy to continuously decrease the density of less important lines.…”

“…For instance, for the largest hex-mesh used in our experiments about 340 million fragments are generated per frame. Therefore, we use a GPU version of priority-queues using a binary tree implementation as search structure [32], which reduces the time required for sorting to slightly more than half of the overall frame time.…”

Real-Time Hair Filtering with Convolutional Neural Networks

“…The method by Münstermann et al also estimates a visibility function but with power, or trigonometric, moments [Münstermann et al 2018]. This results in a low frequency visibility function and is found by Kern et al to frequently over or under estimate visibility [Kern et al 2021]. Maule et al present a comprehensive survey of OIT methods [Maule et al 2011].…”

“…Similarly, lightvisibility can then be evaluated with Shadow Map [Williams 1978] techniques. However, alpha compositing traditionally requires fragments to be processed in back-to-front order, and most Order Independent Transparency (OIT) techniques are either very expensive or give insufficient quality for the high depth complexity of hair [Kern et al 2021;Münstermann et al 2018;Salvi et al 2011].…”