Rendering Techniques 2001: Proceedings of the Eurographics Workshop in London, United Kingdom, June 25–27, 2001

355 Pages·2001·19.656 MB·English

by  Annette Scheel, Marc Stamminger, Hans-Peter Seidel (auth.), Prof. Dr. Steven J. Gortler, Professor Dr. Karol Myszkowski (eds.)

Series Editors
W. Hansmann
W. Purgathofer
F. Sillion
S. J. Gortler
K. Myszkowski (eds.)
Rendering Techniques 2001
Proceedings of the Eurographics Workshop
in London, United Kingdom,
June 25-27, 2001
Eurographics
Springer-Verlag Wien GmbH
Prof. Dr. Steven J. Gortler
Institute of Computer Science
Harvard University
Cambridge/Mass., USA
Professor Dr. Karol Myszkowski
Max-Planck-Institut für Informatik
Saarbrücken, Germany
This work is subject to copyright.
All rights are reserved, whether the whole or part of the material is concerned, specifically those
of translation, reprinting, re-use of illustrations, broadcasting, reproduction by photocopying
machines or similar means, and storage in data banks.
Product Liability: The publisher can give no guarantee for all the information contained in this
book. This does also refer to information about drug dosage and application thereof. In every
individual case the respective user must check its accuracy by consulting other pharmaceutical
literature. The use of registered names, trademarks, etc. in this publication does not imply, even in
the absence of a specific statement, that such names are exempt from the relevant protective
laws and regulations and therefore free for general use.
© 2001 Springer-Verlag Wien
Originally published by Springer-Verlag Wien New York in 2001
Typesetting: Camera-ready by authors
SPIN: 1084676
With 198 partly coloured Figures
ISSN 0946-2767
ISBN 978-3-211-83709-2 ISBN 978-3-7091-6242-2 (eBook)
DOI 10.1007/978-3-7091-6242-2
Preface
This book contains the proceedings of the lih Eurographics Workshop on Rendering,
which took place from the 25th to the 27th of June, 2001, in London, United Kingdom.
Over the past 11 years, the workshop has become the premier forum dedicated to
research in rendering. Much of the work in rendering now appearing in other
conferences and journals builds on ideas originally presented at the workshop.
This year we received a total of 74 submissions. Each paper was carefully reviewed
by two of the 28 international programme committee members, as well as external
reviewers, selected by the co-chairs from a pool of 125 individuals. In this review
process, all submissions and reviews were handled electronically, with the exception
of videos submitted with a few of the papers.
The overall quality of the submissions was exceptionally high. Space and time
constraints forced the committee to make some difficult decisions. In the end, 29
papers were accepted, and they appear here. Almost all papers are accompanied by
color images, which appear at the end of the book. The papers treat the following
varied topics: methods for local and global illumination, techniques for acquisition
and modeling from images, image-based rendering, new image representations,
hardware assisted methods, shadow algorithms, visibility, perception, texturing, and
filtering. Each year, in addition to the reviewed contributions, the workshop includes
invited presentations from internationally recognized experts. This year we were
pleased to have Ed Catrnull (Pixar) and Michael Cohen (Microsoft Research) as
invited speakers. As in previous years, we expect these proceedings to become an
invaluable resource for both rendering researchers and practitioners.
We wish to thank organizing chairmen Yiorgos Chrysanthou and Mel Slater and their
colleagues at the Department of Computer Science, University College London, for
their help in the production of the proceedings, and for taking care of all the local
organization aspects of the workshop. We also want to acknowledge Sun
Microsystems, SOl-Silicon Graphics, and Electronic Arts for contributing financial
support.
Finally, we wish to thank all the authors who submitted their work to the workshop,
and the programme committee members and external reviewers for all the time and
energy they invested in the review process. We were impressed with both the quality
of the submissions and the quality of the reviews evaluating the papers. We are
honored to present the results of this process in the form ofthis book.
Steven Gortler
Karol Myszkowski
June,2001
PROGRAMME COMMITTEE CO-CHAIRS
Steven J. Gortler (US)
Karol Myszkowski (G)
ORGANIZING COMMITTEE CO-CHAIRS
Yiorgos L. Chrysanthou (UK)
Mel Slater (UK)
INTERNATIONAL PROGRAMME COMMITTEE
Ronen Barzel (US) Eric Lafortune (B)
Kadi Bouatouch (F) Dani Lischinski (IS)
Per Christensen (US) Michael McCool (CA)
Yiorgos Chrysanthou (UK) Nelson Max (US)
Daniel Cohen-Or (IS) Sumant Pattanaik (US)
Paul Debevec (US) Bernard Peroche (F)
Oliver Deussen (G) Xavier Pueyo(E)
Julie Dorsey (US) Werner Purgathofer (A)
George Drettakis (F) Holly Rushmeier (US)
Simon Gibson (UK) Christophe Schlick (F)
Wolfgang Heidrich (CA) Peter Shirley (US)
Erik Jansen (NL) Francois Sillion (F)
Henrik Jensen (US) Brian Smits (US)
Alexander Keller (G) Jos Starn (US)
LOCAL COMMITTEE
Alan Chalmers Sharif Razzaque
Simon Gibson Maria Ruiz
Andreas Loizides Bernhard Spanlang
Celine Loscos Franco Tecchia
Contents
Thrifty Final Gather for Radiosity. ............................................................................................. 1
Annette Scheel, Marc Stamminger, Hans-Peter Seidel
Reflected, Transmitted Irradiance from Area Sources Using Vertex Tracing .......................... 13
Michael M Stark, Richard F. Riesen/eld
Simulating Non-Lambertian Phenomena Involving Linearly-Varying Luminaires ................. 25
Min Chen, James Arvo
An Illumination Model for a Skin Layer Bounded by Rough Surfaces .................................... 39
Jos Stam
Real-time, Photo-Realistic, Physically Based Rendering
of Fine Scale Human Skin Structure. ........................................................................................ 53
Antonio Haro, Irfan Essa, Brian Guenter
Efficient Cloth Modeling and Rendering. ................................................................................. 63
Katja Daubert, Hendrik P. A. Lensch, Wolfgang Heidrich" Hans-Peter Seidel
Decoupling Strokes and High-Level Attributes for Interactive Traditional Drawing. ............. 71
Fredo Durand, Victor Ostromoukhov, Mathieu Miller,
Francois Duranleau, Julie Dorsey
Artistic Composition for Image Creation .................................................................................. 83
Bruce Gooch, Erik Reinhard, Chris Moulding, Peter Shirley
Shader Lamps: Animating Real Objects With Image-Based lllumination ............................... 89
Ramesh Raskar, Greg Welch, Kok-Lim Low, Deepak Bandyopadhyay
Image-Based Reconstruction of Spatially Varying Materials ................................................. 103
Hendrik P. A. Lensch, Jan Kautz, Michael Goesele, Wolfgang Heidrich, Hans-Peter Seidel
Polyhedral Visual Hulls for Real-Time Rendering ................................................................. 115
Wojciech Matusik, Chris Buehler, Leonard McMillan
The Wavelet Stream: Interactive Multi Resolution Light Field Rendering ........................... .l27
Ingmar Peter, Wolfgang Strasser
Differential Point Rendering ................................................................................................... 139
Aravind Kalaiah, Amitabh Varshney
Interactive Sampling, Rendering for Complex and Procedural Geometry. ........................... .l51
Mark Stamminger, George Drettakis
Point-Based Impostors for Real-Time Visualization .............................................................. 163
Michael Wimmer, Peter Wonka, Francois Sillion
VIII
Opacity Shadow Maps ............................................................................................................ 177
Tae-Yong Kim, Ulrich Neumann
Interactive Rendering of Trees with Shading and Shadows ................................................... 183
Alexandre Meyer, Fabrice Neyret, Pierre Poulin
Combined Rendering of Polarization and Fluorescence Effects ............................................. 197
Alexander Wilkie, Robert F. Tobler, Werner Purgathofer
Hardware-Accelerated from-Region Visibility Using a Dual Ray Space. .............................. 205
Vladlen Koltun, Yiorgos Chrysanthou, Daniel Cohen-Or
Real-Time Occlusion Culling with a Lazy Occlusion Grid .................................................... 2l7
Heinrich Hey, Robert F. Tobler, Werner Purgathofer
Perceptually Driven Simplification for Interactive Rendering. .............................................. 223
David Luebke, Benjamin Hallen
Measuring the Perception of Visual Realism in Images ......................................................... 235
Paul Rademacher, Jed Lengyel, Ed Cutrell, Turner Whitted
A Perceptually-Based Texture Caching Algorithm for Hardware-Based Rendering. ............ 249
Reynald. Dumont, Fabio Pel/acini, James A. Ferwerda
Path Differentials and Applications ........................................................................................ 257
Frank Suykens, Yves Willems
Interleaved Sampling. ............................................................................................................. 269
Alexander Keller, Wolfgang Heidrich
Interactive Distributed Ray Tracing of Highly Complex Models. .......................................... 277
Ingo Wald, Philipp Slusal/ek, Carsten Benthin, Markus Wagner
Realistic Reflections and Refractions on Graphics Hardware
with Hybrid Rendering, Layered Environment Maps. ............................................................ 289
Ziyad S. Hakura, John M Snyder
Texture, Shape Synthesis on Surfaces .................................................................................... 301
Lexing Ying, Aaron Hertzmann, Henning Biermann, Denis Zorin
Real-Time High Dynamic Range Texture Mapping. .............................................................. 313
Jonathan Cohen, Chris Tchou, Tim Hawkins, Paul Debevec
Color Plates ........................................................................................................................... .321
Thrifty Final Gather for Radiosity
Annette Scheelt, Marc Stamminger:/:, Hans-Peter Seidelt
tMax-Planck-Institut for Computer Science www.mpi-sb.mpg.de
:l:iMAGIS/GRAVIR-REVES -INRIA Sophia Antipolis www-sop . inr ia . fr / reves /
Abstract. Finite Element methods are well suited to the computation of the
light distribution in mostly diffuse scenes, but the resulting mesh is often far
from optimal to accurately represent illumination. Shadow boundaries are hard
to capture in the mesh, and the illumination may contain artifacts due to light
transports at different mesh hierarchy levels. To render a high quality image a
costly final gather reconstruction step is usually done, which re-evaluates the il
lumination integral for each pixel. In this paper an algorithm is presented which
significantly speeds up the final gather by exploiting spatial and directional co
herence information taken from the radiosity solution. Senders are classified, so
that their contribution to a pixel is either interpolated from the radiosity solution
or recomputed with an appropriate number of new samples. By interpolating this
sampling pattern over the radiosity mesh, continuous solutions are obtained.
1 Introduction
In the past 15 years, much research has been concentrated on improvements of the
radiosity method. The computation of radiosity solutions of more than one million
patches is now possible on contemporary standard PCs. However, these solutions most
often do not meet high-quality demands of many commercial applications. The problem
is inherent to the method: because lighting detail is generated in object space, very fine
tessellation is necessary to capture fine lighting detail. Thus quality is not only a time,
but also a memory issue. Furthermore, the widely used linear Gouraud interpolation is
prone to Mach banding, so that a human observer perceives the tessellation structure
easily. In addition, long thin triangles can lead to frayed shadow boundaries, and finally
the different levels at which light is transported to a patch may lead to artifacts. To some
extent, these issues can be addressed by adapting subdivision to lighting discontinuities
[11,21] or by using higher-order interpolation (e.g. [6]).
Ray-based Monte-Carlo methods usually compute the illumination at independent
sample positions in image space. This point sampling allows exact lighting computa
tions, but also makes the exploitation of coherence more difficult. As a result, illumi
nation from big light sources or indirect light make the lighting computation expensive.
For stochastic sampling as a rule 500 or 1000 samples per pixel are needed in such
cases, and in some cases noise is still present.
The idea of final gathering is to combine both approaches. In a view-independent
preprocess a not necessarily perfect radiosity solution is computed. Then in a second
view-dependent step ray tracing is performed that recomputes parts of the illumination
considered critical. One common technique is to only recompute the direct light in
this pass and to add the indirect light from the finite element solution (e.g. [16, 5]).
Alternatively, the radiosity solution can be interpreted as a light source definition, and
the last light bounce towards the eye is recomputed with ray tracing (e.g. [2]).
S. J. Gortler et al. (eds.), Rendering Techniques 2001
© Springer-Verlag Wien 2001
2
1.1 Previous Work
Compared to the huge number of publications on radiosity methods, the number of
publications on a high quality rendering postprocess is surprisingly low. Often, a high
quality rendering step is only a final chapter for a new radiosity algorithm, e.g. [11, 17,
2, 12], but only very few papers address the issue in detail.
In [10], Kok et al. describe a method to decrease the number of shadow samples for
area light sources exploiting information from a progressive radiosity pass. Their algo
rithm subdivides senders adaptively in order to detect completely visible or occluded
parts that can be processed quickly. With a clever shadow pattern scheme, the number
of samples is reduced further. To some extent, this radiosity postprocessing step is done
implicitly in the radiosity computation itself by the idea of Hierarchical Radiosity [8].
Several authors proposed the use of (hierarchical) radiosity for indirect light only,
and to regenerate the usually more detailed direct illumination by ray tracing. More
information from the radiosity solution is exploited in the local pass by Linschinski et
al. [11]: Form factors are recomputed for each pixel and visibility for direct illumina
tion only. Their approach is extended by StUrzlinger [20], where the number of samples
is adapted to the relative contribution of the sources and visibility is computed in a
stochastic approach. Christensen et al. [2] used a final gather step for reconstructing
their radiance solutions. They gather all links with a fixed number of visibility samples,
but also use the cluster hierarchy. It is pointed out that the final gather requires signif
icant computation time and that despite their conservative resampling scheme artifacts
still remain visible. Smits [17] uses a fixed number of visibility tests and analytic form
factor evaluations for each point. For links carrying a relative error below a certain
threshold, the estimate of the link itself is taken. To avoid bias which might occur due
to the threshold, Russian Roulette is used with links whose relative error is below the
threshold.
The final gather step which is very briefly described for the algorithm presented
in [12] extends these ideas. For each receiver, critical senders are determined, whose
contribution is to be computed exactly by final gather.
Bekaert et al. [1] use the radiosity solution for defining an importance function for
a Monte-Carlo ray tracing step. A similar idea later followed in [14], where a radiosity
solution is used to guide a Monte-Carlo path tracer.
Finally, the photon map approach [9] is relevant, because it is essentially a final
gathering approach, however it is based on a global photon tracing pass as preprocess.
From the photon hits an approximate lighting solution can be computed quickly. The
high quality images, however, are then obtained from a well optimized final gather step.
1.2 Idea
The idea of this paper is to better exploit the information from the global radiosity step
for a more 'thrifty' final gather. Consider the example in Fig. 1, middle, showing a
coarse radiosity solution from a candle, illuminated by a big light source on the left and
a smaller one on the right. Due to the different sizes of the light sources, the shadows
are very different. The left shadow has a clear outline that blurs with the distance to the
object; the right shadow is blurry and completely washed out after a short distance.
The radiosity solution is obtained in a few seconds and thus of poor quality, but
it is clear that for a final gather step valuable information can be obtained from it.
In addition to the radiosity mesh which already captures the most important changes
in illumination, the link structure contains per patch hints about shadow boundaries
and the light sources responsible. Furthermore, during the radiosity computation more