http:^^www.tc.cornell.edu^visualization^contrib^cs490-94to95^shuang^bm.html
来自「This data set contains WWW-pages collect」· HTML 代码 · 共 245 行
HTML
245 行
Date: Sun, 01 Dec 1996 19:08:01 GMTServer: NCSA/1.5Content-type: text/htmlLast-modified: Mon, 22 May 1995 12:38:50 GMTContent-length: 11448<html><title>Blobby Modeler</title><body><h1><img src="hole.gif" align=left> Blobby Modeler</h1>Version 1.0 May 18, 1995<br>Szu-Wen Huang<p><hr><h3>Acknowledgements</h3><dl><dd>Anu, for lending me the Tcl/Tk book;<dd>Asif, for giving me good ideas every now and then;<dd>Dr. Bruce Land, for the guidance and patience;<dd>Mom and Dad, for the money it cost to come here and so much more; and<dd>Mabs, for making everything worthwhile.</dl>My most heartfelt gratitude.<p><hr>This document contains information relevant to the development of <b>BlobbyModeler</b>, including a summary of previous research leading to theactual project. Instructions on the installation and operation of thesoftware are likewise provided.<p><img src="ball.gif"> An <a href="#intro">Introduction</a> to <b>BlobbyModeler</b>: What it is, what it does, and some theoretical foundation onhow it came to be.<p><img src="ball.gif"> <a href="#goals">Goals</a> I hope to achieve in thedevelopment of this software.<p><img src="ball.gif"> <a href="#design">Design</a> methodologies andimplementation stages.<p><img src="ball.gif"> <a href="using.html">Installing and using</a> <b>BlobbyModeler</b><p><img src="ball.gif"> <a href="#concl">Conclusion</a> and a summary offeatures accomplished, and perhaps some future directions.<p><img src="ball.gif"> <a href="#refs">References</a> to some relevantpublications.</h3><hr><a name="intro"></a><h3>Introduction:</h3><blockquote><em>"Established geometric modeling techniques exist to handle most engineeringcomponents, including `free form' shapes such as car bodies and telephones.More recently, there has been a lot of interest in modeling natural phenomenasuch as smoke, clouds, mountains, and coastlines where the shapes aredescribed stochastically, or as fractals. None of these techniques lendsitself to the description of <em>soft</em> objects. This class of objectsincludes fabrics, cushions, living forms, mud, and water."<p>- <a href="#refs">Geoff Wyvill, Craig McPheeters, and Brian Wyvill</a> 1986</em></blockquote>As early as 1982, <a href="#refs">James Blinn</a> began exploring methods tomodel molecular structures, which was often done with ball-and-stick orspace-filling-sphere models. In the interest of both artistic variety andscientific accuracy, a new model that can appear more like a real electrondensity cloud for a covalent bond.<p>The conventional approaches to model such a shape via the familiar bicubicor quadric surfaces is difficult for elaborate molecules. For this reason,we seek to simulate an actual electron map with quantum mechanic representationof atoms as a density function of the spatial location. For example, ahydrogen atom might be represented as:<br><dl><dd>[1] <em>D(x, y, z) = exp(-ar)</em></dl>where <em>r</em> is the distance from a location to the center of the atom.If several atoms are present in the scene, the effects of their density fieldare simply added. Extracting a surface of constant value across the field,the molecular density cloud can be formed. Blinn named this the <em>blobbymodel</em>.<p>Years later, Wyvill, McPheeters, and Wyvill decided to apply the technique more generally to the modeling of <em>soft</em> objects, which aretraditionally awkward at best to render. They presented a modification toBlinn's exponential function with one which is computationally cheaper.Additionally, their function has the desirable quality of being able toignore an atom which is too far from the point in question. They firstdefined a bicubic function:<dl><dd>[2] <em>C(r) = 2r³/R³ - 3r²/R² + 1</em></dl>where <em>C</em> is the intensity of influence, <em>r</em> is the distancebetween the point in question and the atom, and <em>R</em> is the radiusof influence of the atom. An almost equivalent but more efficient versionof the function is:<dl><dd>[3] <em>C(r) = ar^6/R^6 + br^4/R^4 + cr²/R² + 1</em></dl>where the values of <em>a</em>, <em>b</em>, and <em>c</em> are found bysolving with the following constraints:<dl><em><dd>C(0.0) = 1.0<dd>C(R) = 0.0<dd>C'(0.0) = 0.0<dd>C'(R) = 0.0<dd>C(R/2) = 0.5</em></dl>yielding:<dl><em><dd>a = -0.444444<dd>b = 1.888889<dd>c = -2.444444</em></dl>In 1991, <a href="#refs">Shigeru Muraki</a> took it one step further, andapplied blobby models on range data to visualize very interesting objectssuch as human faces. Immediately obvious from his research is that toachieve a reasonable degree of realism modeling anything other than simplemolecules, a blobby model should be composed of a large number (typicallyhundreds) of atoms.<p><b>IBM DataExplorer (DX)</b> provides primitive modules that can be used tocreate blobby models. However, since the creation of the model involves theplacement of many atoms, manual design ceases to be a good option quickly. The<b>Blobby Modeler</b> was designed to ease the task of the creation of models,as well as compute the resulting density field off-line to prevent overloading<b>DX</b>.<p>Blobby Modeler was written in C and Tcl/Tk in the spring semester of 1995 aspartial fulfillment of requirements toward CS 790 (Master of EngineeringProject) under the supervision of Dr. Bruce Land, Cornell University Departmentof Computer Science.<p><hr><a name="goals"></a><h3>Goals:</h3>The following are design goals of <b>Blobby Modeler</b>:<ul><li><b>Ease of use</b> - the software should adhere to familiar user interfaceconventions to minimize the learning curve and increase productivity quickly.<li><b>Simplicity</b> - the software should be easy to use and maintain,because it should run alongside <b>DX</b>. It should fit in a modest amount ofmemory.<li><b>Speed</b> - the software should be fast enough to increase productivityrather than decrease it.<li><b>Portability</b> - the software should be portable across Unix platforms,and should be usable in the Cornell Theory Center <b>IBM AIX</b>, ComputerScience Department Undergraduate Laboratory <b>HP-UX</b>, and on <b>SunOS</b>. I will attempt porting to <b>Linux</b> as a test of <b>System V</b>-basedmachines.</ul><p><hr><a name="design"></a><h3>Designing <b>Blobby Modeler</b>:</h3>A three-dimensional modeler running on a conventional computer has thedifficulty of unsuitable input devices. Some attempts were made to allowfor the mouse to control two of the dimensions and the keyboard the third,but I personally think that is clumsy and awkward to use. Instead, Iadapted the approach taken by many personal computer graphics tools as wellas graphics workstation software. I provided three editing windows, throughwhich all degrees of freedom can be encompassed. <b>Blobby Modeler</b>hosts 3 editable windows fixed with the <em>xy</em>, <em>yz</em>, and<em>xz</em> views. In addition, it provides a <em>free window</em> in whicha camera can be placed in 3-space and the model previewed. The renderingin the free window is simplistic and in fact, crude. However, a complicatedpreview window can slow down the response of the software considerably, anddefeats our goal to be lean and mean.<p>By virtue of proximity, <b>Blobby Modeler</b> was written on a SunOS platform,specifically one running SunOS 4.1.1. The target platform would ideally beany platform that <b>DX</b> would support. Though I have not been able totest over many operating systems and versions, I'm fairly confident that thecode should port quite easily.<p>To facilitate the development of the user interface, I selected the <b>Tcl/TkToolkit</b> versions 7.3 and 3.6, respectively. Since this is an interpretedlanguage and offers only a complicated link to C, I also selected an add-oncalled <b>Embedded Tk (ET)</b> version 1.1 by D. Richard Hipp<samp>(drh@world.std.com)</samp>. <b>Tcl/Tk</b> is available at various sitesincluding <samp>ftp.cs.berkeley.edu</samp>, <samp>ftp.x.org</samp> amongothers. <b>ET</b> is available at <samp> ftp.std.com</samp>.<p>Roughly speaking, I divided the tasks among <b>Tcl/Tk</b> and C by writing in<b>Tcl/Tk</b> code that isn't very CPU intensive. The C portion turned outto include only the functions that generate the field (including a horriblequadruple-nested loop) and those that manage the free window (which implementthe 3-D transforms and 2-D projections).<p>In the C portion, a notable feature is the liberal use of global variablesto store partial results. This is done to improve performance by reducingredundant recomputations. Most values involving trigonometric functionsare cached in these variables, and recomputed only as necessary. All in all,the C fragments are simple and straightforward, composing only some one-fourthof the total number of lines of code.<p>The <b>Tcl/Tk</b> portion of the software handles most of the user-machineinteractions, handling drawing, selecting, windows, refreshes, and others.About three-fourths of the program is involved here. Since this was myfirst <b>Tcl/Tk</b> program, it's performance can likely be improved byrewriting some routines.<p>The <a href="using.html">help</a> text is worth mentioning, though. Iimplemented a small widget in <b>Tcl</b> that can display a very small subsetof HTML. There are subtle differences between my implementation and HTML. Unlike HTML, carriage returns are significant in the help text; nested formatsdon't work; and formats cannot span different lines. Despite that, conversionbetween the two formats should not be difficult.<p><hr><a name="concl"></a><h3>Conclusions and Future Directions:</h3>At its present state, <b>Blobby Modeler</b> is a complete 3-D editing platformready for beta-testing. Undoubtedly as it undergoes actual usage, problemswill arise and enhancements will be made. Glaringly absent, however, is theability to move the atoms in an animation. With that capability, <b>BlobbyModeler</b> can probably be used to model scenes such as boiling water.Animation will definitely be a desirable addition.<p>Aside from spherical atoms, it will also be interesting to incorporate atomsof other shapes to aid in the construction of the scene. Cubes will certainlybe easy to work with, for instance.<p>The free window can of course be vastly improved to include its own renderer.It will prevent the small trouble of having to go over the <b>DX</b> in orderto render the scene.<p>Despite the possibilities, this project has achieved its goal of providinga fast, easy-to-use interface for the designer of 3-D blobby models. It ismy hope that actual users will determine the usefulness of the program.<p><hr><a name="refs"></a><h3>References:</h3><ol><li>J. Blinn. <em>A Generalization of Algebraic Surface Drawing.</em>,ACM Transactions on Graphics. 1982.<li>G. Wyvill, C. McPheeters, B. Wyvill. <em>Data Structure for softobjects.</em> Visual Computer. 1986<li>G. Wyvill, C. McPheeters, B. Wyvill. <em>Animating soft objects.</em> Visual Computer. 1986<li>S. Muraki. <em>Volumetric Shape Description of Range Data using"Blobby Model".</em> ACM SIGGRAPH. 1991.<li>J. Ousterhout. <em>Tcl and the Tk Toolkit</em>. 1994.<li>A. Watt. <em>3D Computer Graphics.</em> 1993.</ol><p><hr><address>Last Modified: May 18, 1995<br><a href="mailto:szuwen@cs.cornell.edu">Szu-Wen (Steven) Huang</a></address>⌨️ 快捷键说明
复制代码Ctrl + C
搜索代码Ctrl + F
全屏模式F11
增大字号Ctrl + =
减小字号Ctrl + -
显示快捷键?