From: snowdond@cs.man.ac.uk (Dave Snowdon) Subject: INFO: VR devlopers survey - RESULTS! Date: 17 Mar 93 15:14:30 GMT Organization: Dept Computer Science, University of Manchester, U.K. Here (at last!) are the results of my survey on VR developers. Sorry, its taken me so long to finish it - life has been hectic recently. I got a total of 13 replies which is not big enough to make it worthwhile trying to extract any statistics. However, I think there is much useful & interesting information contained in the replies I received. I hope others will find this information useful as well as myself. Thanks to everyone who responded. (Bob & Mark - feel free to put this in the s.v-w archives - if you think its worth keeping that is.) cheers Dave -- Dave Snowdon Advanced Interfaces Group Department of Computer Science Tel: (UK) 061 275 5717 Manchester University E-mail: snowdond@cs.man.ac.uk M13 9PL U.K. or snowdond@uk.ac.man.cs Don't worry - there's madness in my method ----------------------------------------------------------------------------- VR Developers Survey -------------------- compiled by Dave Snowdon (snowdond@cs.man.ac.uk) Firstly, I would like to thank all those who repsonded to my survey for their kind help and for taking the time to answer. Number of replies: 13 Contributors: Shaun Bangay (g88b7912@beta.ru.ac.za) tad@ROCK Diego Montefusco (montefus@ghost.dsi.unimi.it) Reignier thierry (d89-tre@nada.kth.se) Bernie Roehl (broehl@sunee.UWaterloo.ca) Robert Grant (Robert_Grant.LAX1B@xerox.com) Hans Jense (Hans.Jense@fel.tno.nl) Mark Koch (wirehead@thedeck.Eng.Sun.COM) Paul Emerson (paul@gta.com) David Szego (dszego%srvtech@CANREM.COM) Andrew Connel (Andrew.Connell@arrc.salf.ac.uk) Ken Pimental (kpiment%sense8.uucp@well.sf.ca.us) John Harrison (jharriso@acs.ucalgary.ca) I have listed the applications mentioned, the types of VR used, input devices used, software toolkits and the hardware platforms. The number following each entry is the number of mentions each entry received. Following this this I have listed the (edited) text of the replies. Applications mentioned: - building walkthroughs: 2 - virtual checkers (draughts): 1 - 3D CAD: 1 - games: 3 - space walk: 1 - missile launcher trainer: 1 - visualization: 2 - art: 1 - business: 1 - telepresence: 1 Types of VR used: - immersive: 6 - desktop: 5 - either: 1 Input devices: - powerglove: 4 - VPL dataglove: 3 - joystick: 2 - mouse: 3 - keyboard: 2 - Division 3D mouse: 1 - Gravis Mouse stick: 1 - Logitech Red Baron 6D tracker: 1 - CyberGlove: 1 - Teletact glove: 1 - Teletact commander: 1 Toolkits: - VES (tad@ROCK) - DIVE - REND386 - Division dVS - SUN XGL - Sense8 WorldToolkit - Dimension VRT - VPL Swivel - iGRIP - Alias Upfront - AutoCAD - Multiverse (Robert_Grant.LAX1B@xerox.com) Hardware used: - transputer array: 1 - IBM PC (or clone): 5 - Amiga: 1 - SUN workstations: 2 - IBM RS6000: 1 - Division ProVision 200: 2 - Silicon Graphics: 1 - Macintosh: 1 ------------------------------------------------------------------------------ The Survey Here are the full responses that I received. I have edited the replies to produce a nore or less consistant format. Any mangling of the answers is therefore probably my fault. As a reminder, the orginal questions were: - a brief description of the application(s) you developed - which hardware platform did you use? - what `type' of VR are you using (eg immersive VR, desktop VR etc) - did you use a software toolkit ? If yes: - what toolkit? - who produces this toolkit? - any references which describe the toolkit - what you you feel are the good features of the toolkit? - what you you feel are the bad features of the toolkit? - any general comments about the toolkit - any general comments about VR application design - did you encounter any pitfalls? - what input device(s) did you use? How effective were they for the application. ------------------------------------------------------------------------------- From: Shaun Bangay (g88b7912@alpha.ru.ac.za) - a brief description of the application(s) you developed I am developing a Virtual Reality system at Rhodes University as part of my Masters degree. The system is being developed on a cluster of transputers and attempts to make use of the parallelism offered by the hardware. Two areas are being concentrated on: producing reasonably fast graphics without extra hardware assistance by making use of the parallel processing, and designing a "virtual reality operating system" to support easy development of applications. The system is designed to support multiple users in a world, with multiple worlds simultaneously. Applications that have been written to date include a rough walkthrough of part of the department and a game of virtual checkers (draughts) where the user manipulates the pieces via a powerglove. - which hardware platform did you use? Transputer cluster (31 nodes) with G300 graphics cards, and a powerglove interfaced via a PC. >- what `type' of VR are you using (eg immersive VR, desktop VR etc) Desktop in that no HMD is present yet, although a glove may be used as input device. >- any general comments about VR application design Still working on it. >- did you encounter any pitfalls? Not yet. >- what input device(s) did you use? How effective were they for the > application. Powerglove - Tiring to use. Keyboard - easier to control but less intuitive. Shaun Bangay Computer Science Department Rhodes University Grahamstown 6140 South Africa g88b7912@beta.ru.ac.za ------------------------------------------------------------------------------- From: tad@ROCK - a brief description of the application(s) you developed Currently, we are in the final stage of design, and the initial stage of implementation of a distributed system virtual world. Briefly, what this implies is a virtual "world" distributed over any number of systems on a network. We decided on the distributed system design so that we would be able to distribute the number crunching load of a complex world. Basically, there are two major portions of our project. The first is the VES, or Virtual Environment Server. The VES is the distributed system in charge of maintaining the world. Its made up of four primary servers, each Each primary server is responsible for major portion of the virtual world. The Inter-Base Daemon (base==local world) is responsible for setting up connections among base's and also the various servers in each base. It is responsible for passing control of objects between bases as well. The Interface Daemon is responsible for handling all active objects. This includes the "user" object. Its also responsible for making sure that autonomous objects get run. The Collision Daemon is responsible for tracking all objects in the world and notifying the Interface daemon upon collisions between objects. At that point, the Interface daemon is responsible for passing that message on to the objects. Finally, the DataBase Daemon. This server keeps track of all the objects in the world. When the world first begins, it sends a list of objects, and all the relevant information about them to the collision daemon, also, whenever a new object is created, it is responsible for telling the Collision and Interface daemons about it. Thats a brief (somewhat muddled) overview of the VES. The second major portion of our project is the VDS, or Virtual Display Server. The way we have designed our VES, it is completely hardware independent. All it needs is a network, and some machine(s) to do number crunching. However, the VDS is a completely different story. This is the display portion of the virtual world, whereas the VES was the abstract idea of the world. The VDS is responsible for taking the generic world descriptions passed to it by the VES and displaying they appropiately. It is designed to be hardware orientated, but at the same time, completely open to interpatation. For example, Joe User might have a nice HMD with surround sound, etc, etc. His VDS will be able to take advantage of that and do sound and 3d graphics, and what have you. It will be "immersive" VR. Sue Smith might be running a shutter glasses/powerglove based homebrew display server. It is up to whoever wants to write a display server as to how they want to display the world data sent to it. In this way, its much like Xwindows. The clients (VES) know nothing about the hardware, but they do have a standard language to talk to each other with. The X server (VDS) knows how to take that generic information and display it so that it makes sense to you. - which hardware platform did you use? for the Environment Server, it is primarily machine independent, although our initial implementation will be unix and TCP/IP based. As to the display server, there are a few groups working on various PC servers, an Amiga one and a DOS one. - what `type' of VR are you using (eg immersive VR, desktop VR etc) Ultimately, we will support both, as the design stands, it does support pretty much anything you could want. Our initial implementation will be somewhat immersive, with Mattel Powerglove and Sega Shutter-Glasses support, but thats not to say that someone couldn't implement a desktop VR display server. - did you use a software toolkit (eg WorldToolkKit, MRtoolkit, PROvision etc)? Nope, We designed from the ground up. - any general comments about VR application design VR applications using our design will be very simple. There will be the ability to build applications out of pre existing ones (a-la the unix shell and pipe "|" command ) In this way, one would graphically "build" their program in the world space. Also, there will be standard libraries so that anyone can write an object (application) in either their favorite programming language or a script like language that will be built into the environment server (undecided at this time). Hopefully, by designing it this way our Virtual World system will be able to serve as a standard base for many applications. - did you encounter any pitfalls? Several in the design stage, and I'm sure we will find flaws with our design in the implementation stage. However, nothing that we weren't able to solve without a bit of arguing and compromise. - what input device(s) did you use? How effective were they for the application. We plan on using shutter glasses and gloves. ------------------------------------------------------------------------------- From: Diego Montefusco (diego@minerva.st.dsi.unimi.it) - a brief description of the application(s) you developed A prototype for an 3D interface for a surface modeller: you could grab the whole surface and rotate it, or pick the single control points of the surface (it was a B-Spline or a Bezier) and move them in space. You were using a VPL DataGlove as input device. - which hardware platform did you use? A PC486. - what `type' of VR are you using (eg immersive VR, desktop VR etc) Desktop, but I am planning to make it stereo (I think with LC shutter glasses) because you can't actually perceive well the spatial relationship between your cursor (a hand) and the surface. - did you use a software toolkit (eg WorldToolkKit, MRtoolkit, PROvision etc)? I did everything from scratch. - any general comments about VR application design They strongly involve human perceptual issues: we NEED strong competence and research in this field - did you encounter any pitfalls? Every kind of... The biggest was a design error: we though it would have been easier to understand the relationship between hand and surface positions. We had to make a trick to avoid using stereo (which we wanted to do, but had no HARDWARE for...) - what input device(s) did you use? How effective were they for the application. We used, as I said, a VPL standard DataGlove. It was effective as 3D input device. Maybe a 3D mouse could do the same, but with the glove we could exploit "natural" gestures for the instructions (a fist to grab, a pick - index and thumb touching- to pick the points, open hand to open menu, pointing to highlight a menu item...). I don't actually know if this last point really cares, or was just funnier... (maybe a 3D mouse with some buttons could be as good - surely this way was more poetic and beautiful) The project was done for an university exam (Automatic Design Techniques) by Valerio Raggi and me, Diego Montefusco. A third person did the math algorithms to compute the surface but was not involved in the design of the interface/interaction, which was the interesting point of the whole project. Hope this helps, Diego Montefusco --------------------------------------------------------------------- | Diego Montefusco | Computer Science Department | | Via Pirano 4 | University of Milan - Italy | | 20127 Milano | | | Italia | P L E A S E R E P L Y M E T O : | | +39 (0)2 27001467 | montefus@ghost.dsi.unimi.it | --------------------------------------------------------------------- ------------------------------------------------------------------------------- From: Reignier thierry (d89-tre@nada.kth.se) I am a french exchange student at the Royal Institute of Technology (KTH) in Stockholm, Sweden. During this last year of studies I am pursuing my thesis work in Virtual Reality at the IPLab (Interaction and Presentation Laboratory) at KTH. The subject is 'Exploration of interaction methods with a dataGlove'. This study is part of a larger swedish program: - a brief description of the application(s) you developed The Swedish Institue of Computer Science (SICS) virtual world or telepresence effort is a part of Multi_G, a swedish effort to develop high speed computer networks and advanced distributed multimedia applications and platforms. The SICS Distributed Interactive Virtual Environment (DIVE) is an experimental platform for development of televirtuality, user interfaces and applications based on shared 3D syntethic environments and other tentative systems using the illusion of being in alternate spaces. By its distributed architecture, DIVE is especially tuned to multi-user applications, where several participants may interact even though they may be geographically dispersed. DIVE is a loosely coupled heterogenous distributed system based on UNIX and internet networking protocols within local and wide-area networks. Consistency and concurrency control of common data is achieved by active replication, reliable multicast protocols and distributed locking methods. - which hardware platform did you use? Here at KTH we have on SUN workstations with a graphic accelerator. At the SICS they have 3 Silicon Graphics workstations INDIGO, IBM PC and SUN Sparc stations SLC. - what `type' of VR are you using (eg immersive VR, desktop VR etc) it will be immersive. But I'm more taking care of the DataGlove part. - did you use a software toolkit (eg WorldToolkKit, MRtoolkit, PROvision etc)? If yes: - what toolkit? DIVE (see above) - who produces this toolkit? SICS - any references which describe the toolkit see above - what you you feel are the good features of the toolkit? As far as I'm concerned I think it's the best one... seriously I'm just colecting paper for the moment and I don't have so much time to work with it. But a student team is actually using it to simulate a network manadgment. - what you you feel are the bad features of the toolkit? I havn't got so much experience. - any general comments about VR application design My work consists in collecting papers concerning the use of that tool (DataGlove) and I must try to define the criteria upon which tools like gloves, mouse, spaceball ... are better with different applications. I should determine what kind of interface is optimal for a dataGlove in different applications? Which applications are best served by dataGlove interaction, which would do better with some other interaction tools? - did you encounter any pitfalls? the banckrupt of VPL Inc. - what input device(s) did you use? How effective were they for the application. a VPL DataGlove, a Bird and at the SICS the VPL eyephone. ------------------------------------------------------------------------- - Reignier thierry - - email: d89-tre@nada.kth.se - ------------------------------------------------------------------------- - Kai-Mikael Jaa-Aro (advisor) - - Interaction and Presentation Laboratory - - Department of numerical Analysis and Computing Science - - Royal Institute of Technology - - S-100 44 Stockholm - - SWEDEN - - fax: (+46) 8 790 09 30 (thesis work student) - ------------------------------------------------------------------------- ------------------------------------------------------------------------------- From: Bernie Roehl (broehl@sunee.uwaterloo.ca) - a brief description of the application(s) you developed A simple virtual world that you can walk around in and manipulate objects. - which hardware platform did you use? 486/50 - what `type' of VR are you using (eg immersive VR, desktop VR etc) Desktop and homebrew HMD (built by Dave Stampe) - did you use a software toolkit (eg WorldToolkKit, MRtoolkit, PROvision etc)? No. We used our own (REND386). - any general comments about VR application design It's tedious without tools. Our objects and worlds are currently all built using text editors. - did you encounter any pitfalls? None in particular. - what input device(s) did you use? How effective were they for the application. Joystick, mouse, keyboard, PowerGlove. The University just recently got a Logitech 6D tracker which we plan to interface soon. Bernie Roehl, University of Waterloo Electrical Engineering Dept Mail: broehl@sunee.UWaterloo.ca BangPath: uunet!watmath!sunee!broehl Voice: (519) 885-1211 x 2607 [work] ------------------------------------------------------------------------------- From: Robert Grant (Robert_Grant.LAX1B@xerox.com) I'm not a formal VR developer, I'm just working on a system in my own time, that maybe I'll be able to release to the net. - application(s) you developed: The application I'm developing (provisionally called Cybernet, until I found out that there was a company already called that! :-() is a multi-user, networked, non-immersive VR system, that provides an API for VR application developers. The concept is that all the object management and communications is handled by the system and the programmer just adds their twist to the whole thing by providing the code for any special object features and the actual object definitions - structure, colour etc. I'm using two driving problems to test my architecture, one is a rip off of Dactyl Nightmare - i.e. people running around on platforms shooting each other (pondering adding the option of one user being the pterodactyl! :-)). The other is a dogfight game where people fly around in planes shooting each other up again, not that I'm really into violent games, its just that I can get my friends to test it out if there's a chance of blowing each other up! My current status is that both applications are working through the API, but I need to improve the code that manages the Dactyl movements. I'm also adding feedback to the user about score etc. through some client independent calls. After that it's code review and debug time! - hardware platform: IBM RS6000s and 16Mb token ring network - what `type' of VR: Non-immersive, desktop. - did you use a software toolkit: No, built my own! - any general comments about VR application design: Very difficult to develop a general purpose solution to the problem. You may have read some of my posts, but it is very difficult to abstract the problem so that any world specific code can be isolated, and the server/client code kept as general as possible. Various problems I've had are: automated object collision detection, automated user/object management, flexible ancilliary output (i.e. user feedback outside of the world - menu selection and textual feedback etc). - encounter any pitfalls: Still trying to come up with a solution to keeping an object within a certain range of values without hard coding those values. - what input device(s): Mouse and keyboard, pondering PowerGlove, but the problems I have are at a more abstract level than just interaction. - How effective were they for the application. OK, though some interfaces are a little dubious. Robert. RGrant.lax1b@xerox.com ------------------------------------------------------------------------------- From: Hans Jense (Hans.Jense@fel.tno.nl) G. J. Jense and F. Kuijper, Virtual Reality Research at TNO-FEL, in: Proc. 2nd ESA Workshop on simulators for European space programmes, Noordwijk, Nov. 1992, ESA publ. nr. ESA-WPP-039. G. J. Jense and F. Kuijper, Virtual Environments for Advanced Trainers and Simulators, to be presented at the International Training Equipment Conference (ITEC), London, May 1993. These articles describe the (prototype) applications we have developed so far. If you like I can send you photocopies. - a brief description of the application(s) you developed 1. A space-walk demo, based on a scenario from ESA. Basically it allows an astronaut to fly around the Hermes spacecraft and the Columbus Free Flying laboratory, using hand-held thrusters for propulsion. In addition to this, simple On-orbit Replaceble Unit exchange procedures can be performed. 2. A simple VE for training target acquisition and firing procedures for the shoulder launched Stinger anti-aircraft missile. It allows the participant to intercept incoming, low flying aircraft by following the correct aiming and firing procedure. - which hardware platform did you use? Division ProVision 200. - what `type' of VR are you using (eg immersive VR, desktop VR etc) Immersive. - did you use a software toolkit ? We use dVS, which is more of a VE operating system than a toolkit. > If yes: > - what toolkit? > - who produces this toolkit? Division Ltd., Bristol, UK. - any references which describe the toolkit See for instance: Charles Grimsdale, dVS -- Distributed Virtual environment System, Proc. Computer Graphics '91 Conference, London. - what you you feel are the good features of the toolkit? - modular approach - inherently supports a parallel process model - object oriented application programmers interface - what you you feel are the bad features of the toolkit? - documentation could be better - some minor problems to do with synchronization between concurrent processes, supposed to be solved in upcoming new release - hierarchical construction of complex objects now rather kludgy, also expected to be improved in new release. - any general comments about the toolkit In general we're quite happy with dVS. - any general comments about VR application design So far, our rule-of-thumb which states that development time is roughly evenly split between object modelling and software development, has always been confirmed. - did you encounter any pitfalls? No. - what input device(s) did you use? How effective were they for the application. Division's 3D mouse has been very effective in both applications. We have so far had no need to use a DataGlove. Kind regards, - Hans Jense +---------------------------------------------------------------+ | Hans.Jense@fel.tno.nl | | | | Fysisch en Elektronisch TNO Physics and Electronics | | Laboratorium TNO, Laboratory, | | Postbus 96864, P.O. Box 96864, | | 2509 JG Den Haag 2509 JG The Hague, The Netherlands | | | | Tel: 070 - 326 4221 Phone: +31 70 326 4221 | | Fax: 070 - 328 0961 Fax : +31 70 328 0961 | | | +---------------------------------------------------------------+ ------------------------------------------------------------------------------- From: Mark Kock (wirehead@thedeck.Eng.Sun.COM) Note: My research in VR is not affiliated with my job(yet) at Sun Microsystems. - a brief description of the application(s) you developed VeX - Visualization environment for X-windows. - which hardware platform did you use? Sun SPARCstation 1 - what `type' of VR are you using (eg immersive VR, desktop VR etc) Desktop VR. - did you use a software toolkit ? - what toolkit? XGL - who produces this toolkit? Sun Microsystems - any references which describe the toolkit None. - what you you feel are the good features of the toolkit? Fast rendering in X-windows. Works within OpenWindows enviroment. Includes features like: Vertical retrace sync. Supports hardware double buffering. Supports stereo in certain hardware. - what you you feel are the bad features of the toolkit? Not designed specically for VR applications. I had to write my own API "VeX" on top of XGL. - any general comments about VR application design My main goal was near real-time graphics. I am tired of seeing these "VR" systems with frame-rates that are around 5 frames per second. The minimum rate should average around 16 frames per second. I have had to limit the detail of the objects that I am rendering to achieve my goal of real-time animation of the objects. - did you encounter any pitfalls? I feel very alone in this field. There aren't many people in this field who know a whole lot about what they're doing. I figured there would be more experts abound than I have found but it seems like we are all learning something entirely new. I still find most newletters, books and magazine articles to be dissapointing in the information that they provide. I usually scan through them but I find that they don't have much to say. I get more info out of the average stuff in Computer Graphics World. - what input device(s) did you use? How effective were they for the application. I use a regular mouse. I would like to get a Logitec 3D mouse but they still cost around $1000. I am looking into some ideas I have for my own 3D mouse but don't have much time to persue it. Another device I use is the Sega LCD glasses. I have written my own drivers for them and they seem to be fairly effective for my use. They do make my eyes tired rather quickly though. I would like to get some StereoGraphix Crystal Eyes but they are just as pricey as the 3D mouse. - More about VeX I devolped VeX because I found that a low level rendering package like XGL does not handle 3d objects very well. It handles lists of 3d points. I wanted to handle groups of objects which happen to be made up of lists of 3d points. VeX objects contain position and orientation information and can be grouped into a tree hierarchy. I single call to my rendering function will recursively render all of the objects in the tree. All of the primitives are eventually rendered by XGL though. VeX is object oriented but written in C. I have plans to rewrite it in C++. This should make the system easier to use from a programmer's point of view. The main advantage of VeX is that I can hack up a virtual world within a few hours and start playing with it. It would take me days to weeks to do it any other way. I have not looked into ay of the commercially available VR packages. Most I have found are only available on the PC, are too expensive or are less than 10 frames-per-second rendering. ------------------------------------------------------------------------------- From: Paul Emerson (paul@gta.com) - a brief description of the application(s) you developed We are developing a demonstration VR world to promote custom world services that we are offering. The demonstration consist of a building that has an atrium in the center and 4 wings (North, South, East and West). Access to each wing is via a pair of automatic sliding doors. Each wing consist of a main hallway with 6 rooms, (three on each side). To enter a room you must must in font of a door and press the activation button on the joystick. The door the slides open revealing the room and you may then enter. Each wing is devoted to a particular area: Games, Factory, Art, Business. Each room in a particular wing contains an aspect of some VR concept in regards to the wing it resides in, (eg. an alien shoot-em-up game in the game wing). We have also implemented a robotic arm model for one room that allows the user to pickup an manipulate objects in the room. Our demo world also implements some sound output. - which hardware platform did you use? 2 486/50 Clones Each confiugred as: 16MB RAM 250 MB HD IDE VGA card and Display 2 serial ports 1 parallel port 1 Joystick port The machines differ as: Machine A: 1 Gravis Mouse Stick with adapter card 1 Standard Joystick Machine B: MediaVision PAS 16 Sound card 2 Standard Joysticks 1 VR flight helmet 2 Harmonic Research RGB-to-NTSC convertors 2 Sony PVM 1341 NTSC Video Monitors 1 Polhemus 3SPACE Tracker (connected to machine B) 1 Sense8 Special Design Paralell cable for connecting the two systems 2 SPEA Fire Boards (i860) video cards. - what `type' of VR are you using (eg immersive VR, desktop VR etc) Immersive - did you use a software toolkit ? Yes. If yes: - what toolkit? Sense8 World Toolkit (i860 version) for SPEA Fire board - who produces this toolkit? Sense8 - any references which describe the toolkit Their promotional literature. - any general comments about VR application design The time spent designing and creating the model in a CAD package is almost equalled to the time spent coding. - what input device(s) did you use? How effective were they for the application. Polhemus 3Space: for head tracking. No problems and works quite well. But the sensors must be within 30" of each other. All metal must be out of the sensing area. Garvis Mouse Stick: We went with this because most users can't fly a SpaceBall. And since our is a demo for many people we thought this was simpler and cheaper. Standard Joystick: Sense8 didn't support this so we wrote our own driver. Works as well as the supported Gravis Mousestick. We haven't had much problem with accuracy either. Our driver now make it possible for us to use any PC type game port hardware. It's not portable to other hardware platforms, but these devices are too numerous and cheap to ignore. We use the 1st joystick for flying and the 2nd for controlling actions, like the robotic arm, etc. -- Paul Emerson | Global Technology Associates, Inc. President | 7198 Harbor Heights Circle Email: paul@gta.com | Orlando, FL 32835 CIS: 72355,171 | Tel 407-296-3636 FAX 407-295-1954 ------------------------------------------------------------------------------- From: David Szego (dszego%srvtech@CANREM.COM) - a brief description of the application(s) you developed Our company, The Servants Of Technology, is developing what we call MindSlip, an integrated package consisting of glove, hmd (stereo) and head/hand tracking, which can interface with a standard VGA and parallel port. It will however, require custom software. We are aiming at the games arena for starters. - which hardware platform did you use? 80x86 systems. As well, we are designing a co-processor card (long term plans) to interface to the ISA bus. - what `type' of VR are you using (eg immersive VR, desktop VR etc) It would be immersive, but not necessarily real-time generation of worlds. It could be considered Desktop VR because of the system it runs on. We like to think of it as PC VR - did you use a software toolkit ? If yes: - what toolkit? I am working with Bernie Roehl to adapt his Rend386 libraries. As well, we are designing our own set of VR routines. I cannot say any more about the software we are designing at this time. - did you encounter any pitfalls? Time and money. We are a group of students, although we all have extensive professional experience. This limits us in terms of when we can do our R&D, and how much we can spend on it. All money is out-of-pocket for us. - what input device(s) did you use? How effective were they for the application. We started by using a PowerGlove, but are designing our own as the PGlove is rather noisy. As well, we are looking at the Logitech Red Baron 6D tracker. It seems to be a very stable unit. Hope this was of some help to you, please send email if you have any specific questions. David Szego (DSzego%SrvTech@canrem.com) --------------------------------------- //// The Servants Of Technology oo / Reality Engineering \/ We Think, Therefore You Are. ------------------------------------------------------------------------------- From: Andrew Connell (Andrew.Connell@arrc.salf.ac.uk) At ARRl we seem to be getting a bad reputation for never responding to surveys since our mail server refuses to believe sci.virtual-worlds exists, so I hope this brief answer is some help. - a brief description of the application(s) you developed Telerobotics using VR/Telepresence for both mobile and manipulative robots. Nanotechnology (NanoPresence?) Linking VR with real time Scanning Tunnelling Microscope(STM) image acquisition. Simulation Visualisation for dynamic process design. Maintainence testing for aero designs to avoid physical mockups. + others ( architechtural, mining, hazardous environment). - which hardware platform did you use? Various Silicon Graphics (Onyx RE2 arriving soon). Division - have developed on most platforms. (e.g. vision, provision, PCVision, still ocasionally use SuperVision). Macintosh PC - what `type' of VR are you using (eg immersive VR, desktop VR etc) Immersive Desktop, Large Screen Stereo Projection. - did you use a software toolkit ? We have used most toolkits, for both design and visualisation, such as, Dimension VRT, Division Dvs, Sense8 WTK, Swivel (VPL), iGRIP (Deneb robotics), Alias upfront, AutoCAD as well as many in-house developed tools. - any general comments about VR application design Don't give it all away to companies for free. - did you encounter any pitfalls? Yes, never believe a spec sheet. - what input device(s) did you use? How effective were they for the application. Polhemus isotrack and fastrack , flock or birds (ascension), logitech ultrasonic, DataGlove , CyberGlove , Teletact gloves/commander. * In response to a question asking for further details Andrew responded: * Dave Dave, Most of the work we are doing is in collaboration with other companies, so I'm not able to go into too much detail. However, when we have got our current VR initiative off the ground ( 1-2 months ) we will probably start talking more, though some of the work has already been documented. The STM work ,in conjunction with Salford University, has been included in VR News and a few papers, and we are hoping to get something into Physics review letters about that soon. Some early ( and unfortunately quite poor quality ) footage of the aero maintanence and dynamic simulation work was shown on the BBC 9 o'clock news last month, if you know someone who videod it that might be some help. Though as I said we will probably get quite talkative about it once all the funding has been finalised. The types of VR we use depend on the audience, executive level poeple prefer desktop or 3D projection, whereas for enthusiats or experiments we use immersion. I think we went in to this a little more in an article for VR Report last year. As for which toolkits we use, it depends on the application. Our favourite modeller is iGRIP, which is used also for dynamic simulation, it is has a great selection of powerful solid modelling tools. For running immersive worlds WTK seems to be winning with Dimensions VRT best for designing worlds. As for gloves, I still personally don't use them on a daily basis, preferring the ease of use as a handgrip such as the commander. Gloves are also bad news at exhibitions, the last time we took a Teletact glove to one it was destroyed by 1500 spaniards who wouldn't leave the hall without trying it on. Hope this is the information you wanted, Andy. ---------------------------------------------------------------------- Andy Connell National Advanced Robotics Research Centre Tel: (+44)61 745 7384 University Road FAX: (+44)61 745 8264 Salford UK M5 4PP Email:apc@uk.ac.salf.arrc ------------------------------------------------------------------------------- From: Ken Pimental (kpiment%sake%sense8@well.sf.ca.us) * Ken replied to my survey asking if he could be of any assistance. I * asked him if we would answer a few questions about Sense*'s WorlToolKit. * He very kindly took the trouble to send this interesting reply. * Dave - What do you perceive as being the strong points about WorldToolKit? 1. multi-platform support -- Sun, SGi, PC-860 (we also have TIGA and VGA versions working in house). We are the only VR toolkit to provide this level of platform independence. 2. scalability: start developing with our soon to be release VGA product and port to SGi Reality Engine when you have saved enough pennies. 3. texture-mapping: all of our competitors say they *will* have this, but we have been shipping it for over a year on the PC and SGi (Soon on the Sun too!). 4. elegant and intuitive language for 3D simulation: This is our strongest feature. Industry proffesionals have purchased our product simply because, as they say, "it is an elegant approach for VR development". 5. customer base: the quality of our customer base is outstanding. Some of these are experts in the VR field, and they have made an informed decision based on analyzing all the possible tools they could buy. Customers such as: British Telecom Nottingham Univ. NTT Jet Propulsion Lab Rockwell NASA Ames (a site-wide license!) Institute for Simulation and Training Samsung Lockheed Mitre Corp. Banff Center for the Arts Carnegie Mellon University and many more research labs. 6. Feature set. Version 2.0 has over 400 functions to allow rapid development of VR environments. Features such as: Terrain following Interpolated Paths Multi-level collision detect Interactive texture manipulation Heirarchical objects Accurate display of intersecting objs 3D Sound/MIDI sound Network capable Parallel rendering DXF/NFF in/out Dynamic object/polygon creation Support for Gouraud/textured surfaces Support for antialiasing and perspective-correct textures on PC 3D Animation functions 2D Foreground fns (text,etc) Support for SGi GL and Performer within WTK renderer (SGi product) Support for hw devices like: Ascension, Polhemus, Fake Space Labs, Crystal River Engineering, Logitech, StereoGraphics CrystalEyes, Spaceball, Voice recog. hw, Touch screens, plus other devices - How do you feel it compares against similar products such as Dimension International's Superscape or VREAM? Not in the same class. We offer far broader support for high performance rendering on multiple platforms. Both Superscape and VREAM are limited to PC/VGA solutions. For the price of the Superscape VR toolkit (L5,400) you can buy WTK plus a SPEA Fire board (L5,000). VREAM is a nice product for VGA, but simply doesn't have the functionality or performance that proffessionals require. The same can be said for Dimension's product. - What sectors in the market is WorldToolKit particularly targeted at? Currently, the proffesional VR developer. Someone with at least a $25k budget. All the misc. hw for VR is so expensive that these are the only people doing serious VR work. We are just announcing a new product code-named "Mercury" that will be available for hobbyists. This will be a sub $1,000 VGA-only DOS toolkit. This is for the experimenters and tinkerers out there that want to start exploring. Only WTK offers them a path to higher performance if they outgrow their VGA system. - From your contact with users of WorldToolKit have created any guidelines for developing VR applications? Any DOs and DON'Ts? Tough one. Budget 50% of the effort for modeling and 50% for code development. Optimizing for performance is a tricky balancing act. Customers also need to understand the limitations of today`s technology and also be flexible with the rapid pace of change. - Are you aware of any unusual applications which are using WorldToolKit which have not yet been accepted as `standard' applications for VR? Hines VA Hospital Wheelchair accessibilty study Research marketing tool for analyzing product presentation (can't say who). Intercontinental networked VR Museum (CMU) Robotic control (NASA Ames) Jenny Holzer art project "Spirit Lodge" by Canadian artist, (was in National Gallery) ------------------------------------------------------------------------------- From: John Harrison (jharriso@acs.ucalgary.ca) Sorry I didn't answer your first posting; I've been travelling without access to News for some time. Hope this info will help you; I don't have a lot of time to reply, and we haven't published any papers yet (though a quick summary of my experiences will be published in the CIPS Informatics '93 proceedings - the conference is in Calgary, Canada at the end of April '93). - a brief description of the application(s) you developed We are running a project entitled "Art and Virtual Environments" here at the Banff Centre for the Arts, and are in the midst of developing several VR pieces with several groups of artists. As well, in Fall '91 we ran the "Bioapparatus" project, creating several VR pieces with several individual artists on lower-end hardware. So, I'll answer each question in two parts; with our current, high-end (A&VE) project in mind, and then our former, low-end (Bioapparatus) project. It is difficult to do justice to the artists' work with quick descriptions of the various projects, so perhaps I should just summarize that the projects cover a range of VR possibilities with architectural walkthroughs, explorations of anatomy, interation with 3D "paintings", a Salish longhouse with interactive spirits, recreations of landscape, interactive narratives, and exploration of 3D mathematical formulae. I can send you a copy of our press release, which has more detail on the projects, if you like. We've done about 12 interactive VR environments so far. - which hardware platform did you use? A&VE: 2 SGI VGX 310 Powerseries computers We are currently ordering a SGI Onyx system with RealityEngine2 graphics. We are using NeXT IRCAM workstations for 3D sound. Bioapparatus: IBM '486 clone DVI graphics boards Macintosh IIfx with Focalpoint cards for 3D sound. - what `type' of VR are you using (eg immersive VR, desktop VR etc) Mostly Immersive and projected VR (for performances), a little desktop VR. We are ensuring that the pieces have fully developed 3D audio as well as the 3D visuals - we find that enriches the pieces significantly. - did you use a software toolkit ? Yes. A&VE: If yes: - what toolkit? MR Toolkit - who produces this toolkit? University of Alberta - any references which describe the toolkit - what you you feel are the good features of the toolkit? Having sourcecode available Runs fast on SGI Not very hard to use. - what you you feel are the bad features of the toolkit? Not yet a "commercial" product, but getting better all the time. - any general comments about the toolkit We wrote our own custom software to do 3D sound with the NeXT computers and the IRCAM workstations. The computers all communicate and stay in sync over Ethernet. Our 3D models are developed using Alias software, by the artists and associates. Bioapparatus: If yes: - what toolkit? WorldToolkit - who produces this toolkit? Sense8 - what you you feel are the good features of the toolkit? Works great, lots of drivers, easy to program. Useful demo program included which we could modify to suit many of our applications. We received exceptionally good technical support from Sense8 - what you you feel are the bad features of the toolkit? - any general comments about the toolkit We wrote our own software on the Macintosh to allow Max to talk to the Focalpoint 3D audio boards; this facilitated audio development. The Mac and 486 talked to each other over a serial line to stay synchronised. Our 3D graphics models were developed using Autocad by an architect. - any general comments about VR application design See CIPS paper - did you encounter any pitfalls? See CIPS paper - what input device(s) did you use? How effective were they for the application. We've used Logitech trackers, Ascension birds, Polhemus sensors and a Dataglove as input devices. We also developed a simple button and joystick device to allow for a 3-month installation of one piece at the National Gallery of Canada. Logitech trackers work o.k., though sensitive to ambient noise and slow. Had trouble navigating with the bird as a 3D cursor. Polhemus sensors work great, though we have the old Isotrak sensors with a limited range and slower update rates than the Fasttraks. We had to redesign our lab somewhat to get the Polhemuses away from all the metal in the walls/ceiling/floor to avoid tracking errors. The Dataglove works fine for simple gesture recognition and navigation by pointing with the hand. Sorry for the brevity of this response; some of these issues are covered better in my CIPS paper. We're in the middle of a heavy production time here now, so perhaps I can supply more information in a couple of months when the A&VE pieces are closer to completion. -John Harrison Senior Research Analyst Art and Virtual Environments Project The Banff Centre for the Arts Banff, Alberta, Canada jharriso@acs.ucalgary.ca ------------------------------------------------------------------------------- Dave Snowdon Advanced Interfaces Group Department of Computer Science Tel: (UK) 061 275 5717 Manchester University E-mail: snowdond@cs.man.ac.uk M13 9PL U.K. or snowdond@uk.ac.man.cs Don't worry - there's madness in my method