Mapping a dome with TH2G and soft edge blending???

Scratchpole's picture

Hope some of you can help and someone may be interested in using this if I can get it polished off.

I started this (simple) composition a couple of years ago to exercise my quartz skills but never quite got there.

I am hoping it will be possible to use three projectors with 0.8 to 1 lenses in order to fill an entire dome using the Matrox triple head to go.

I have utilised Matthias Oostrik' CI filter but am having trouble with the settings for the overlaps, please have a fiddle and see if you can crack it for me. Matthias has tried to help but unfortunately he does not have K3D installed.

I think I have got it right for the joins in the middle but it needs overlaps for the texture at each end and the more I fiddle the less I understand!

Note: this would be a cheap-ass way to do dome projection. I know there may be issues with the projectors focal length (DOF), and it may have to be changed to six projectors, but I don't have a client for this and it's just a proof of concept.

Thanks in advance. Joe

Dome mapping.zip176.09 KB
TH2GSEBDome.jpg67.84 KB

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

Scratchpole's picture
Re: Mapping a dome with TH2G and soft edge blending???

OK I think it's done :) Big respect to Matthias for the overlap/tiling technique, which is not built into the filter.

Check it out and tell me if you think it will work. It is set up for internal front projection but could easily be used for projecting onto the outside of a dome by rotating the model.

Need to make a line-up image as I know it's going to be hard to get all the projectors aligned correctly.

Cheerio, Joe

Dome Mapping Final 2.zip177.83 KB

dust's picture
Re: Mapping a dome with TH2G and soft edge blending???

if you and matthias say it works i will take your word for it. matthias should know if it will work he does massive projections.

i know one of my teachers is trying to get matthias on help him project onto i giant oil cauldron. i guess something similar to this would need to be done. instead of spherical it would have to be cylinder shape. this is exciting though i always wanted to do a spherical projection since i have seen paul bourkes stuff.

how is this any cheaper than paul bourkes method ? i think he is just using a mirror then warping which would require 6 to 2 less projectors ?

to bad i don't live in my old studio. i had built spherical cyclorama for photography. i called it the cove. i could have tested this for you. oh well that was a previous life.

what you need is a friend at a planetarium to test.

Scratchpole's picture
Re: Mapping a dome with TH2G and soft edge blending???

I think it will work. But it could be a bit messy at the peak of the dome where all three corners meet.

Hey you can send your teacher my direction if he needs help, a cylinder is easy in comparison;)

I meant cheap-ass as in it could be done with consumer equipment on a smaller scale than you get in planetariums. Also as it is designed for 4x1 content, quite a few VJ's are already working at TH2G/4x1 ratio so they could just chuck in their clips and be ready to go. When it comes to planetariums the content is very expensive.

Paul Bourke has written about this method on his site from what I understand, but I don't think he has implemented it.

I do have a friend who works on planetariums: the problem is they all have different systems. This method could be suitable in less scientific situations.

Next step might be to make it suitable for 4000pix diameter images, how would I chop up the images to fit?

dust's picture
Re: Mapping a dome with TH2G and soft edge blending???

I have 3m pocket projector can't really test this but some day I plan to make spherical setup. I know how to build a wall like that. I used to build skateboard ramps. I suppose a small one wouldn't be expesive but also wouldn't need a triple head.

Scratchpole's picture
Re: Mapping a dome with TH2G and soft edge blending???

I may have spotted a flaw. The model I have used is composed of 1/3 sections of a hemisphere, this does not take into account the overlapping areas. I think I need to add the edge blended areas to the 1/3 sections.