This is a project which will literally pull your mind out from your head.
The whole idea is to view ourselves live in 3rd person. A camera will be mounted to ourselves looking at us from behind. A video goggle setup will then allow us to only see ourselves in 3rd person.
We will explore how to make a setup like this for relatively cheap. We'll look at other permutations, such as transmitting the data, using the setup in paintball with a marker mounted camera, and so on.
[edit: check out step 10 for the 3rd person view]
Step 1: The Camera's
As the setup goes, really any camera with a video output will do. I have a small video CCD camera with the component video and audio output. That ran me about $60. You can get cheaper ones if you'd like. If you really want to play with this, I would suggest getting a good camera. After having done it, I would prefer a higher quality camera.
Some digital cameras or camcorders can do video out as well. That will work, as long as it is live.
Step 2: The Video Goggles
This is the more expensive part. There are a few options out there when looking for video goggles. I have a headplay setup. It gives me a 800x600 resolution, which is one of the better ones I have seen. These run around $350. You can get some 520x480 res googles for $250 if you look. I have other projects in mind for these goggles, so I wanted something better, but it is expensive.
We did look at other options, but to really immerse yourself into the video, a goggle setup is best. Further more, in the future, if I can get some 3d cameras, the goggles I have will support that.
It is possible to get a small LCD and stick it in front of your face, but that just doesn't seem that it will work as well. And when you consider that a nice sized screen will run you $100 to $200 bucks, plus you need a way to mount it, it just doesn't add up.
Step 3: Radio Transmisson
This part is entirely optional. We placed a video splitter on the camera and ran it to a 800mW 900Mhz radio transmitter. Now, a word of caution. You do need an Amateur Radio License to legally operate something of this power. It is true that you can buy them anyways, but it is good to stay legal. The nice thing is, it is easy to get your license.
The information needed to get your license is located here:
Step 4: Recording the Video Stream
We have the radio setup transmitting the video and audio. At the other end we have the receiver which is connected to the video -> USB converter. This allows a remote viewer to watch what the camera is seeing. It also allowed us to easily record the video stream.
I am using a KWorld converter. They are cheap, it seems to work well. There are a ton of other options.
Step 5: Constructing the Camera Mount
The camera mount system is made of 1/2" PVC. A frame is constructed and attached to a vest. If you don't have a vest, no worries. I just had one around. You can easily strap the frame to your chest with some straps. This may actually be better as it will hold the frame more closely to your back, whereas the vest had some play.
Extending from the frame is a T junction with the boom. The boom need more support so initially we added some string, and then some wire, but finally we decided we needed some more rigid support and added more PVC pipe.
Only the T-junction at the base of the boom was glued in. All the rest of the pieces are hand tightened.
It is important that the whole frame is rigid and well attached to your body. As we will see later, the motion of the camera relative to your body can make or break the experience.
The camera itself was attached by screwing the mount that the camera came in into a PVC cap. This work very well and it allowed us to adjust the camera angle which was very important.
Later we used some digital cameras to record the action at a higher res. Because the cap was just fit on, with no gluing, we were able to switch it out for a cap with a 1/4-20" bolt which screws into the standard camera mount. We cut a slot into the cap so that the camera viewing angle was adjustable. With the boom fixed it is important to be change the camera angle so that you can see yourself. Also, with different people wearing the suit, we needed to sometimes change that angle based on their height and preference.
Step 6: Gimbals Mount & Other Failings
Initially we were dead set on building a gimbals system in order to stabilize the camera. There are a ton of instructables on gimbals, and we more or less followed their ideas.
We found in the end, that the gimbals did not reduce the motion. We wanted to dampen out the jostling motion of walking or running. However, the wobble of the camera in the setup was far too disorienting. Finally, after testing a camera fixed to the boom we decided that it was really not all that bad and we stuck with it.
After walking and jogging we soon learned the value of having a rigid frame and made adjustments there as well. Any wobbling or jostling of the camera was very disorienting.
Step 7: System
So, now as we have discussed all the parts, we need to put it all together. Really, it is simple. The camera output goes to the goggles and the radio transmitter. The radio receiver is connected to the USB converter.
The goggles have their own battery, the camera and transmitter work of 12V. We tried two different power sources. The simple, yet more expensive in the long run, method was to use to lantern batteries in series. This provided the 12 volts needed. The second option was to use a 11.1 V lithium battery. The equipment worked fine at this voltage and the battery was substantially smaller, lighter, and rechargeable.
Step 8: In Practice: Testing
After getting the whole setup working we did some tests just standing. We didn't yet have the batteries hooked up, so we were tethered in a sense. This allowed us to test the camera and goggles and gimbel to get our initial reactions.
First reactions: It worked well. It felt very odd and I can already sense the effect of a moving point of consciousness (we'll talk more about that later) the gimbel is working, but needs some adjustment.
Step 9: In Practice: Walking & Running
After getting the battery connected we were able to walk up and down the hallway. Right away the gimbel was a problem. It was swaying far too much. Were able to make it better by changing the mass and adding some damping, but in the end we settled on a fixed camera. It was hard enough to walk around in this thing. With the camera moving, it made you feel sea sick.
Depth perception is an issue. It is also hard to remember that you have this thing strapped to your back, so no quick turns indoors and everyone else needs to watch out.
Step 10: In Practice: Depth Perception
Now we took the setup outside where we had some room to run. Objects look farther away than they really are. This is due to the wide angle lens on the camera.
We were able to cautiously run, but after 5 minutes in the suit you started to feel comfortable. Because you only have one camera, you lose a lot of your depth perception. But that is not the only way we sense depth, and soon my senses we adapting. It is amazing what the brain can do.
Looking back upon the tests I remember being 'sucked' into the camera. I felt like "I" was the camera. It was odd. I felt taller. In fact, because you were at a different level than the camera, it was easy to get your dynamics messed up, and sometimes we almost fell over.
After wearing the setup for a while and removing the goggles there is an odd feeling of your point of reference being adjusted. As if your consciousness were being sucked back into your head. Wild stuff.
Step 11: Advanced Setups: Multiple Cameras & Gun Mount
We recorded the video separately on the two cameras on the back. We transmitted the marker mounted camera signal to the laptop where it was recorded. The goggles viewed the marker camera. In the future we will add a switch so that you can quickly switch between views. I am sure that will mess with our heads even further.
Step 12: Future Work & Ideas
We'd really like the use a 3d camera. This will either require a 3d camera setup, or an additional camera and a way to interlace the two views as this is the format most 3d capable googles read the video in. Otherwise, if we can find some video goggles that can take in two video sources, that would be easiest.
Remote Control People
The original purpose of the radio transmitter was for another idea which hasn't been implemented yet. A human subject would wear the 3rd person camera and no goggles with their eyes covered. The video signal is transmitted to the controller. Using remote controls or some other signal, the subject would move according to the controllers wishes, much like a video game.
Move the 3rd person camera back
The 3rd person view we used was pretty close. It was challenging not being able to see your feet. Also, in a lot of 3rd person video games, you can see the entire character. This is difficult to do using our method as it will require a long extension rod. Also, you will be able to see the camera mount in the camera. We are experimenting with some ideas around this such as using a balloon.
It would be interesting to be able to match up the camera movements with the motion of your head. this was one odd part in using the pure 3rd person view. In order to look around you had to move your whole body. this felt unnatural. Using a gyro and a camera pan-tilt servo mount would allow us to have the 3rd person view which would move with your body and head. This would most likely increase the immersion as it is closer to reality.
Live video switching
And as I mentioned before. We are in the process of adding a video switch for the multi-view case allowing us to switch between views on the fly.
Upgrade our workshop
While we can make do with what we have, we could definitely get some new tools. Some things we have in mind are the following:
Craftsman 10'' Bench Drill Press with Laser Trac
Drilling through PVC with a hand drill and a vice isn't the most ideal...
Craftsman PVC Cutter
Well, with all the PVC cutting this would have helped just a bit.
If you have any other ideas, We'd love to hear them!