Meeting
Agenda Meeting
Agenda
(July 8) 
(July 15)
(July 22)
(July 8)
(July 22)
L-systems (Lindenmayer systems) were developed as a way of describing both the development and the structure of plants. An alphabet is defined, and an initial string of symbols is supplied. A collection of rewriting rules is then iteratively applied to the initial string, transforming it into another string. The resulting string can then be interpreted in some fashion, to produce a graphical output.
I will describe L-systems in more detail, with pretty pictures, and then discuss my experience using L-systems in a commercial package.
John W. Tukey is one of the pre-eminent mathematical scientists of our era. He has spent most of his professional life at Princeton University, where he went in 1937 as a mathematics graduate student, joined the Department of Mathematics, and subsequently became Chair of Statistics and Donner Professor of Science. He has made fundamental contributions in mathematics and computation, and delineated and developed the area of exploratory data analysis, including its statistical aspects and themes. He has also served in many important consulting capacities, beginning with research and development in ballistics and other areas during the second world war, and later participation in the National Halothane Study, President Johnson's committee on environmental pollution, and other expert panels on most of the important issues of our time. He was for many years Associate Executive Director of Research for AT&T Bell Labs.
Data structures for combinatorial objects are traditionally designed to handle objects up to a certain size. We introduce the idea of threshold data structures: representations that allow a richer collection of operations on small objects than large ones. As illustrations of the general concept we discuss threshold data structures for sets and multisets, and show how the former can be applied to cache memory design.
Consider threshold representations for subsets of a universe of size n, supporting insertions and deletions at any level, and enumeration of sets whose size does not exceed the threshold t. We derive lower bounds on the space used by any such representation. When t is fixed and n --> infinity (the case of interest in memory design), any such representation must use, asymptotically, at least (t+1)*log2(n) bits of memory. Applying the theory of error-correcting codes, we design a structure, efficiently supporting the required operations, whose space consumption matches the lower bound. Similar results are proved for multisets.
This is joint work with Marcos Kiwi of Universidad de Chile.
Fifteen years ago, computer animation was revolutionizing the look of films, TV and videos - from spaceships rendered on a Cray supercomputer for the film The Last Starfighter, to chrome hockey pucks rendered on a DEC VAX for Hockey Night In Canada. Today, computer animation is a vital part of virtually every film, commercial, and game on the market. These images are the manifestation of a large, evolving computer animation industry: an industry subject to technological advances and paradigm shifts which create enormous business challenges for those involved. Through use of animation and images, this talk will provide insight into the technical and business challenges of the 3D animation (or digital content creation) industry, both past and present.
Kim Davidson is President and COO of Side Effects Software and President and Executive Producer of Catapult Productions, makers of the children's TV series, Monster By Mistake. After graduating from the University of Waterloo (BES,1977 and BMath, 1981), Kim worked at Human Computer Resources (now SCO). In 1984, he moved to the computer animation firm, Omnibus, where he spearheaded development on the proprietary in-house software that was to become PRISMS. Early in his career, Kim helped create over two hundred broadcast spots and logos for CBC, ABC and TSN. Last year, Kim was one of four Side Effects Software employees to receive a Technical Achievement Award from the Academy of Motion Pictures Arts and Science for his work in developing PRISMS.
He will receive the J.W. Graham Medal in Computing and Innovation at convocation at the University of Waterloo on June 19, 1999.
Side Effects Software, established in 1987, is a world leader in advanced 3D graphics and special effects software for use in film, broadcast and game development. Side Effects Software continues to lead the field of procedural animation with its unique visual network technology. The company's main product, Houdini, is used by a number of digital content creation leaders including Digital Domain, DreamWorks SKG, Rhythm & Hues and Sony Pictures Imageworks. Recent film projects created with Houdini include: The Matrix, eXistenZ, Prince of Egypt, and What Dreams May Come.
Sponsors for this seminar are the Faculty of Mathematics and the infraNET Project, University of Waterloo. The infraNET Project, initiated by the University of Waterloo in 1996, is a partnership to advance Web and Internet technologies. Its founding partners are: INFORIUM, MKS, Open Text, RIM, Sybase (Waterloo) and Waterloo Maple. We also gratefully acknowledge the assistance of the Institute for Computer Research, University of Waterloo.
With the rapid development of wireless communication hardware in recent years, wireless networks that provide mobile services anytime, anywhere are becoming more a reality than a mere possibility. Being able to communicate via data, voice, or even video at any place, at any time is a very attractive goal that many researchers were trying to reach. Powered by the latest processor technology, many mobile computers, such as laptops or personal hand-held devices, are capable of playing multimedia streams (video or audio) on local storage in real time. However, in the situation of video conferencing or live broadcasting, multimedia streams are fed through the wireless network to which the mobile device is connected. Thus, the limitations of the wireless network may severely downgrade playback performance. This is because it is generally well known that wireless networks, compared to wired networks, have lower bandwidth and higher error rate.
MPEG is one of the most popular digital media standards. Because of MPEG's highly effective compression algorithm, it is widely accepted and used in many areas. Inspired by the layered nature of MPEG syntax, filters can be applied to drop a layer of data without destroying the syntax and semantics of the overall bitstream. In a wireless environment, this approach can reduce the amount of data transmitted to achieve better overall playback performance.
This thesis explores the possibility of developing such an MPEG video filter that can provide data reduction for saving wireless network bandwidth. Based on a proxy architecture, Comma, developed previously by David Kidston, this MPEG video filter can be deployed to provide transparent filtering service.
A solid prototype of this transparent MPEG video filter, MFilter, is designed and implemented. MPEG video syntax and its encoding/decoding rationale are explained. The Comma proxy architecture that MFilter is based on is described. Design issues, MFilter's software architecture and transparent MPEG video filtering techniques are presented in details. Experimental results are also described.