Projection mapping is terrifying
Admit it. The first few bunch of times you did projection mapping gigs, they were terrifying. I don’t know anyone who thinks otherwise. There are so many moving parts, so many variables, and such a huge lack of iteration and testing that can be done on many projection mapping projects that it can paralyze even the most experienced developer. But once you get up and running, that fear pretty much leaves. Why is that?
In a recent conversation I had about projection mapping, I came up with a few solid strategies for developers of any experience level that can take the terror even out of first-time projection mappers. These tips can apply to most kinds of projection mapping, but they really kick in for complex projection mapping involving 3D models (buildings, cars, sculptures, etc).
Understand the playing field and separate the jobs
One of the tricky things with projection mapping is the amount of steps involved in the process. There are large and discreet steps including:
- Making/dealing with the 3D model of the object being mapped (including doing all the UV unwrapping)
- Creating content for the actual projection mapping (including all the perspective tricks)
- Dealing with projectors and the actual build out of whatever the setup needs to be
- Rolling out the system, software projection mapping, content warping, playback, etc
One of the overwhelming parts of projection mapping is that all these things and more need to be dealt with thoroughly. Trying to hold all of it in your mind at once can be completely overwhelming for developers new to projection mapping. Being able to list all the steps separately and tackle them with focus one and a time will make your projection mapping project much more manageable.
Simplify geometry for UV unwrapping
With complex projection mapping jobs, you’ll often get some like the architecture’s 3D models. These models can be slow to load, overly complex, and impossible to UV unwrap. They’ll include every little detail of the building or object, most of which are not useful in any stage of the projection mapping process. Especially if you’re working on your own and trying to get through a whole projection mapping project, UV unwrapping may be on of the toughest parts of the process if you don’t have a strong 3D background.
In these situations, one of the best things you can do is start by deleting all the extra elements from the model. Really spend a little bit of time trying to lighten it up like deleting unnecessary details (like tiny ridges in the building that wouldn’t be seen during projection anyways) or extra objects that won’t take projections (like small poles or glass elements). It’ll make it a lot easiest to work with.
You can also take the object as a whole and divide it up into a bunch of smaller objects that would be easier to manage and UV unwrap. When you bring them back into the software you can keep their world space coordinates from the 3D exports so that everything lines back up easily.
I don’t come from a 3D model (or 3D at all!) background, so I can be honest when I say I conceptually get what it means to UV unwrap something, but in practice I have no idea what all the buttons in C4D do. I know the ones I need and the rest are just there. But even with that lack of knowledge, I still get my work done because I know that everything in the 3D space is relative. If i use the “wrong” unwrapping method (box vs cubic vs cylindrical vs etc) as long as it can be worked with in the other stages of the process then it’s not a show stopper. For example, you’d get used to creating your content in After Effects using the weird UV map, then you’d just keep passing it along the chain. It might not be the ideal perfect UV unwrap job, but for all intents an purposes it could work, and it will importantly allow new projection mappers to keep chugging along down the process.
This is because once you unwrap the UV and hitting all your button in C4D, you’re essentially creating a new UV map for the model as well. If you then take that 3D model and export it, it’ll take it’s UV information to the next software it gets imported to. So if you then import that weirdly mapped 3D model into TouchDesigner, you still have your (weirdly) mapped 3D object intact. Thus, things will work and, more importantly, you’ll be able to avoid getting stuck and bogged down with some random one thing you don’t understand much.
Now I’m not condoning that just randomly doing things is good for a project with a team and money on the line, but if you’re learning to projection map or studying immersive media or in a lower-stress job situation, you can just bash your way through.
Iterate a ton
The nature of projection mapping is that you often won’t actually have your canvas till right at the last minute (when it’s much too late to correct any significant issues). It is invaluable to do whatever you can do to make a scale model using paper or a 3D printer and rigging up a small project to it to test out what you’re doing. The content you’re making or the projection plan go terribly wrong if you don’t have a ton of experience. The nice thing about iteration is that clients also would love to see iteration and samples and tests! It’s a win-win! Propose to increase the budget slightly to add the model you need to build and a small rental projector and you’re off to the races.
Projection mapping can be daunting to say the least, especially when you’re just getting started. Using these strategies during your experiments, learnings, and even small projects can get you to finish line when you’d otherwise be blocked. Enjoy!