Immersive Virtual Environments (aka CAVE)

[aokeil]

Hi,

 

I have been using an Immersive Virtual Environment (aka CAVE) in teaching Architecture for a few months now. link: http://www.engg.uaeu.ac.ae/a.okeil/uaeu-cave/

 

It was mainly used to allow students walk through full size models of historical buildings, contemporary buildings and design proposals. I will be investigating using the facility as a design environment rather than as a presentation tool. In other words, the architect sits inside the immersive virtual environment and works on his design there. He/she is surrounded by the sketchy design and manipulates objects in order to develope the design further. This manipulation can take place directly on the images displayed on the screens or on conventional software running on a laptop inside the environment.

 

I would like you, specially those who are practicing architecture, to tell me your opinion about such a design setup. Potentials, anticipated problems, suggestions are all welcome.

 

Meanwhile, I would be very thankful if you could send me any of your interesting, high quality (with shadow, light and texture maps) VRML models for display in the facility.

 

Thanks.

[zamalkawy]

Dear Professor Ahmed ,

 

i,m very happy to see you here doctor .I'm developing a cd about alexandria old pharos for alexandria biblotique with Dr awad . Now i'm preparing the model to be viewed in a VRML environment and i'll be happy to send it to you once i finish it .

 

thank you

 

khaled adam

[mbr]

The problem I still see with all of this is the time needed to get it to look decent. Baking the textures and lighting the scene, then optimizing to play with current high end graphic cards is a chore. No architect would ever have near the budget to do this.

So if this is a theoretical project for the future, great, but for now it's just too expensive and too slow of a process.

 

I can design a building in wireframe and opengl without any problems. When I render I want to be able to test different materials, etc., which could not be done if the textures are baked.

 

Basically, it all sounds really nice, but the hardware is just not there yet. Again, if it's theoretical and proposing ideas about the future, then go for it. Someday it'll happen, just not for a long time.

[Richard McCarthy]

Hello guys!

I would like to invite you guys to post to the "Realtime/VR/ Interactive/ VRML" forum. (Scroll down... or click here : http://www.cgarchitect.com/vb/forumdisplay.php?f=66) There, we have many experts in the field to answer your question . I personally have worked with CAVE as well, and found it quite expensive and is not as immersive as I first thought. A lower cost effective way would be to purchase 6 DOF VR googles, which is also very space efficient.

So come post in the correct forum! We await you there

[aokeil]

Hi mbr, Richard,

 

1- In the design stage no baked textures is needed. As you said wireframe or simple shading might be enough but in this case in full size.

 

2- Caves are very expensive unless you build them yourself. The one I built cost 17000 US$ and if I am to build it again it might cost 14000 US$.

 

3- I predict that in a couple of years many architects would afford a CAVE but the problem is that the full potentials of immersive virtual environments as design tools (environments) could not be exploited using conventional design approaches and therefor there is great need for a new design paradigm. If we do not start experimenting now we will be waisting a lot of precious time.

 

4- Space for a cave will remain a problem. Our cave is in a room 8x12 m.

 

5- I doubt any architect will be happy to work several hours a day with a head mounted displays on his head.

[thomas+son]

Professor

Could you use SketchUp?? (http://www.sketchup.com) as your tool for “editing real-time” and just think of the cave as "a expensive monitor". You should be able to download the demo and try it to see if it would work. (It uses OpenGL). The price is reasonable and it is good for beginners. I am not a skectchUP employee, but I have seen in our own office people with no 3D skills take the program and run with it.

 

Please keep this forum updated. Your project is VERY interesting.

 

David Thomas

[Richard McCarthy]

Hello Professor Aokeil

 

I agree, CAVE is just way to expensive, it requires too many hardware to make it effective CAVE, also because it's 6 sides of wall to be projected, it means any 3D hardware needs to split it's processing time to 1/6th for each side of the wall. That means if the model is textured, you need to cache all those visible textures and models in ALL SIX WALLS! If you are doing urban visualisation, it become very inefficent and even impractical.

 

The best way I see it is to use Head Mounted 6DOF tracking VR googles (like Cyberglass or SONY's Glasstron with 6 DOF (degree of freedom) gyroscopic sensors). It is so much cheaper, Sony Glasstron is about $2000 USD each + DOF sensor it shouldn't be more than $2500 - which means that you can built SIX of these for the price of ONE CAVE. It is more space efficient as well, takes no more space than the size of a hat, and lastly it reqires much less dedicated hardwares than CAVE, and the 3D processing won't be split among the 6 projection walls, instead it will be just split between the closely aligned glasses. This is much better because using visibility culling alogrithm technique and backface culling, it means the CPU are not required to store and process as much textures and models because the view are always only facing ONE direction.

 

As you have pointed out, and I agree with you, that head mounted VR google also have it's short comings, it is still heavy after very long use (if it can be made light as a feather, then it will be no problem) , and many people who have work with it shows sign of motion sickness due to incorrect parallax settings (the left eyes sees slightly different thing to the right eyes because of the distance and angle to each other, and how different is the subject of a lot of adjustment and varies person to person) If we can solve those 2 main problem, I see no reason why head mounted VR google is not the choice of VR research.

[Richard McCarthy]

Professor

Could you use SketchUp?? (http://www.sketchup.com) as your tool for “editing real-time” and just think of the cave as "a expensive monitor". You should be able to download the demo and try it to see if it would work. (It uses OpenGL). The price is reasonable and it is good for beginners. I am not a skectchUP employee, but I have seen in our own office people with no 3D skills take the program and run with it.

 

Please keep this forum updated. Your project is VERY interesting.

 

David Thomas

 

Thomas, I have dream about that for a long time now as well

There should exist and low cost inexpensive VR program that behave like Sketchup. Actually, come to think of it, I think there did exist such thing on SGI decades before (early 90's)... I remember I read it but I can't quite remember the name of the program, I think the it might be called "DeVice" or "Dv8" or something along those line. My memory is hazy .......

The program basically let you use VR google to model and inspect the space. (Sort of like skethUP except you use a VR glove) If you point (with VR glove) on the ground and raise the ground, you get a fence! touch the wall and push and the wall is push inside... slap the walls and it will get destory/deleted.

 

So in essensce I think if there should exist a VR program, a new paradigm in user interface would need to exist to enable it's efficient usage. Gestural/voice recognition I think is most suited for this and will be most benefitial to the VR research in my opinion.

[jrsv]

Hello,

 

As has already been mentioned, cost, space allocation, and workflow integration do not favor current mass implementation of CAVE setups. At present, you can achieve a lesser degree of immersiveness (content not 1:1 scale and no viewer-centered perspective tracking) during (e.g. 3ds max7 walkthrough mode in camera viewport) and after (e.g. Quest3D) design stage.

 

To incorporate Immersive Projection Visualization (IPV) during design stage, the emerging field of Interaction Design can be of great help. Furthermore, to date, real-time interactive 3D (non IPV) productions are more abundant in urban design/planning projects, not architecture projects.

[seismograph]

interresting topic!

 

The problem with the cave is, that it's good for interior visits and studies, but not for designing or editing buildings. A Glove is a nice toy, but still not the haptic tool an architect will request.

Architects ..or at least most of the architects i know... still prefer pencil and paper for the first draft.. and when it comes to the first model they use everything laying around on the table to build a real model ..it's simply much faster then hanging around with the pc.

 

A quite good approach to the process of developing a building is augmented reality, which works with a really easy setup ,which cut costs ,don't need a 12x8m room and is a very mobile solution too. Also the haptic feeling of moving around real objects and seeing the real world together with the concept is really nice. ..at the moment it's best for urban design, but who knows what will come next, because you need very little tools to create a model IRL: some paper, a knife and glue -> three tools which shouldn't hard to code for some bitcracks.

 

anyway .. i hade a funny move in my experience with vrml last month with a project where i decided to make a vrmlscene instead a rendering:

It's ...for viennese conditons... vanguard architecture on an obscure location which needed a presentation for the governing major...they where impressed, because our beamer presentation cleared some irritations about the project and also dazzled them a bit *g* ..maybe ..and i don't want to move vrml in the corner of oddity.. ,you can still impress ppl with such presetations.

 

mail me when you want to see some screenshots.