3D model?

[jamyn]

Hey everybody,
So, I'm attempting to work some magic in a pretty short time frame for a multimedia theater show.

Setup is as follows:

3 4*4 ft cubes in the center of the stage, upon which animated text and video clips will be displayed.

3 10 ft tall, 3 ft wide canvas covered flats behind, hung from the ceiling, to be used primarily for ambient clips.


I'm using one projector on the back, and one for the cubes.

The video clips contain characters in the play performing sign language that the live actors will interact with in real time.


My question is this::

I've seen a lot of different methods online for accurately spanning the video on the cubes. Some advocate useing AI or photoshop to draw on the projection surface and then using that to create a transparent .png to function as a mask in Resolume.

Others suggest creating a 3D model by taking a picture from the exact perspective of the projector(s) and importing it into Max3DS or Cinema4D, and then projecting a flat image that's already been pre-warped.


What are your opinions on this? Which method is better, quicker, etc? I'm not looking to create a lot of crazy animation--my primary objective is to get the video clips showing up correctly on the cubes.

All suggestions appreciated, this is my first shot at doing this

Cheers,
jamyn

[jamyn]

Also, any consensus on the best way to use 2 projectors? I might try using both my Macbook and my iMac and try to sync them in some way.... or is there an easier way to go without spending a ton of money?

[Basic]

I think it depends.

I did see this a while ago...

http://www.madmapper.com/after-effects- ... -building/

From mapping projects ive seen they have actually built little replica's of the structure and mapped that way. but I don't think any one workflow is correct. do you have much time to do it? sounds likea great project.

will it be the same image through the 2 projectors? you can use a splitter. but again depends on needs.

[jamyn]

hmm.... so should i be considering using madmapper instead of the onboard mapping in arena 4.1? I know you can syphon arena to madmapper, but is there any advantages to doing this?


the biggest question about that whole thing that i have is weather or not its necessary to create a 3D model and compose video specifically for it and then project a flat image without using any mapping software, or if I should just use the onboard mapping in arena (or madmapper)


in terms of the 2-projector situation, 1 projector will be at the back of the room, aiming high at the wal containing 3 10 feet tall 4 ft wide canvas-covered wood flats (the 2D background set). The other will be on a pretty short throw very close to the 3D mapped cubes in the center of the stage.


I'm not sure how to go about doing my composition stuff for this kind of a configuration... I would need two models because of the different perspectives of the projectors, but then that kinda screws my workflow for the content creation process. any suggestions?

[Basic]

hmm.... so should i be considering using madmapper instead of the onboard mapping in arena 4.1? I know you can syphon arena to madmapper, but is there any advantages to doing this?

Sorry I wasn't suggesting using madmapper I was just suggesting you use the general principle of how to map a structure/object

Others suggest creating a 3D model by taking a picture from the exact perspective of the projector(s) and importing it into Max3DS or Cinema4D, and then projecting a flat image that's already been pre-warped.

I didnt originally seen this. this seems to be what a lot of projects I've seen are based on

in terms of the 2-projector situation, 1 projector will be at the back of the room, aiming high at the wal containing 3 10 feet tall 4 ft wide canvas-covered wood flats (the 2D background set). The other will be on a pretty short throw very close to the 3D mapped cubes in the center of the stage.

I don't know sorry.

[4eyes]

Depending on the resolutions you will be outputting you should take a look at a DualHead2Go or a TripleHead2Go.

[DayVeeJay]

Gimp is better for tracing objects out. Photoshop likes to scroll when you get to the edge of the screen. You can also accomplish the same thing with resolume. Create a solid for each surface that you want to map. For example, if you where to map a cube with three visible sides then you would create three solids. Then you can use the advanced output to manipulate those solids to line up with your surfaces. Both work flows have their advantages and disadvantages. I personally like using models.

[LXConcepts]

I do a lot of video mapping/motion graphics.

For something simple like a building or what you're doing, you can do it with two blended projectors and a picture of what you're mapping.

Steps;

1. Take a picture of the object from the relative perspective of the projector(s)
2. Import the pictures into photoshop and create a mapping mask of the key features. Pick out things like hard edges, windows, doors, etc. make all the images taken into one picture that represents all the projectors from their relative point of view.
3. Use the background remover tool to eliminate everything in the image except for the object you want to create content for.
4. Import this now made mask into after effects and begin making content relative to your mask. Remember to turn this image off when rendering your video!

I do the same thing when pixel mapping a multi display system so I can tell exactly what's going to be in my screens when I do it.