Projecting onto a geodesic dome

modified's icon

Good day fellow Jitterererers.

I am an exceptionally new Max/Msp/Jitter user; though I've just moved over from Pure Data & Vvvv so have some idea of graphical programming for AV.

I am a 2nd (going into 3rd) year degree student studying Digital Music & Sound Art and I'm about to enter my final year and start my final project which I intend to be an immersive audiovisual installation.

Long and short of it I guess is how in Jitter would I approach the task of projecting onto a geodesic dome from the outside?
I want to make a space that people can explore and interact in without having the discomfort of projector light in eyes.

I've been informed on the Vvvv forums that I'll be looking at 3 projectors and to use this to run all 3 from my computer.

In Vvvv I have an idea on how to get to grips with this; but to be honest Vvvv seems (so far) to be a bit more intuitive when it comes to projection and setting up mappings.
I'm sure it's not, but the thought of relearning everything I've learned in the last year or so to the same standard is a little daunting and Jitter seems fairly complex... Especially in that Vvvv uses spreads, which are very powerful & I'm not sure how you would go about replicating in Max.

Basically I bought Max a week ago, and I've about 8 months to get this task sorted.
I've been programming audio in Pure Data for 2 years and visuals in Vvvv for about a year; I'm no expert but I'm clued up enough to be able to hit the ground jogging so to speak. That said; a little direction would be incredibly useful.

If anyone has any advice then I'll appreciate it very much.

Here's a model of the type of dome I want to use; I'm told this is helpful as there's definitive triangles which make the mapping a little bit easier.

I've also read that Jitter has it's roots in the video realm whereas Vvvv has a more prototyping & generative history. This aspect appeals to me more about Vvvv and so I wonder if Jitter will actually be capable of doing what I want it to do; which is generative and custom visuals. Think Ryoji Ikedia, John Whitney, Alva Noto/Carsten Nicolai.

I am not all that into playing with actual video; and my project is about synchronicity between audio & visual materials, thus a video can't provide that. It needs to be realtime generative AV synchronicity.

Thanks in advanced for any help

kodamapixel's icon

I've found that projection mapping onto a dome is quite tricky with Jitter. I'm not familiar with VVVV, but from what I've been told, it deals with blending between multiple projectors a little better than Jitter does. Mapping to the triangular surfaces, rather than blending between projectors, will definitely be much easier, however you might find yourself trying to reinvent the wheel, just to stay entirely within Max - I know I did.

From your post, I can't tell whether you're wanting to project onto the outside surface for any technical reasons. If you can project internally, I found this resource helpful. It explains how to use a mirror to project onto a hemispherical surface. One of the benefits of this method is that it seems to deal with focus across a curved surface much better.

modified's icon

Hi KodamaPixel. Thank you for the reply.

I should have made it clear that I wish to project on to the dome from the outside. The idea is that people are free to explore the space without having the issues of shadows and projector lights in the eyes. This is why I want to use a dome (or a teepee) type structure with some lycra for the surface of it which can be projected on from the outside.

Vvvv is fantastic for blending between projectors, and it seems at the moment that it may be a little more in depth than Jitter might be.

I don't really mind how the final result comes, whether it be entirely in Max or with other software. But the issues I had with PD and Vvvv were serious.
(OSC connection between the two programs was not fast enough for what I want to do. There's synchronization errors & data dropouts & also Pure Data uses WDM drivers which don't work with Focusrite multichannel soundcards).

I can try and use Vvvv with Max, though I fear that there may be the same errors with the OSC (Localhost too).

kodamapixel's icon

Sounds familiar; I had the same reasons for attempting to project onto the outer surface.

My suggestion would be to test your projectors on a dome that's as close to full-scale as you can get. Once you get past the blending/multi-projector hurdle, you might find that sharp focus is your biggest problem - remember that the depth of field will need to be relatively great to keep focus across a curved surface.

antialias's icon

Mesh Warp Server looks like a great resource for projection mapping in Jitter:
http://meshwarpserver.org/?p=50

maybites's icon

AntiAlias is perfectly correct. But beeing the developer of meshwarpserver I am obviously biased.

Since I am not a VVVV developer I cannot speak about the advantages of VVVV versus Jitter when it comes to dome projection, but my guess is it is similarly hard. using meshwarpserver helps you to solve the hard problems of mapping and softedges. I have never done a dome, the closest mapping to a dome was a 180 degree projection onto a flat bubble shape with 6 projectors, so I think meshwarpserver can help you solve your mapping problem.

however, meshwarpserver only solves the mapping and softedge part (even though looking at your model, I dont think softegde is needed, you can map each triangle by one designated projector, and your supportstructure will cast a shadow in the edge, so softegde will be overkill), the content creation (if interactive) has to be done inside a different application. Meshwarpserver, though opensource, is a standalone application, but on OSX you can use syphon to pass video-textures from another application (this can be a jitter patch if you wish, but openframeworks, processing or any vj-tool would do, too) to MWS.

before you start you need to know how you wanna map your content onto the dome-shape. MWS works with custom meshes with custom UV-Maps. The UV-map defines how the videotexture is spread over the mapped-mesh. If it is going to be an immersive experience where you want to create the illusion of standing inside a virtual space that extends into all directions, then it gets complicated.

ah, and I forgot: there is a commercial alternative: http://www.blendydomevj.com

_src's icon

I was the sound designer on a geodesic dome show with Media Designers Matthew Ragan, Alex Oliszewski, and Daniel Fine. Check out how they worked here:

They ran Touch Designer and Izzie (which was seriously fantastic for all sorts of reasons)- I ran max, ableton, and cuelab. It all played nicely.

I'd go into it more, but I just popped on because I saw geodesic in the title of the thread and thought I'd point you toward them.

Best of Luck.

oh, and beware the acoustics.
it was a nightmare.