[Gabriel] is making 3D movies using only one camera. This should be impossible because true 3D needs to be stereoscopic, with images from different perspectives for each eye. He’s worked this out by mounting the camera on a CNC gantry and programming it to make two passes along slightly different paths. He’s plotting the camera paths using SketchUp and a plugin that exports paths as CamBam files, automatically adjusting for perspective. The two videos are then merged using Stereo Movie Maker.
We’ve embedded both a 3D video as well as behind-the-scenes filming video after the break but you’ll need the red and blue 3D glasses to view the former. It’s not too much of a stretch to tweak his methods and use this for stopped motion video where one button press takes a frame for each eye. Now, who will be the first to bring us a Star Wars remake filmed in stopped-motion 3D using the original action figures?[youtube=http://www.youtube.com/watch?v=DIp5kyBB6JY]
[youtube=http://www.youtube.com/watch?v=KKjOpRsZiHE]
Wow this is an amazing way to over complicate something! i love it!
It’s pretty funny to try to imagine a person with free access to a CNC machine to play with, but only one camera and no way to get another.
I would rather see someone wrap two cameras together with a duck tape.
this only works if the object is not in motion, also lol @ only having one cam, yet having a CNC??
cool way of getting it done I guess….
I’d rather see the Matrix done with action figures. Bullet time!
Not accurate. I can see that both passes were not identical the stop points and some of the middle points were not 100% identical.
Neat idea, cheaper to simply buy 2 cheap camcorders.
Of course they were not identical, if they were it would not be 3d. The 3d effect is generated by giving each eye something slightly different to see.
He obviously had two cameras. If not, what was he using to film the ‘making of’ vid? Maybe he was using a still cam and I missed it.
I suspect this is as much a ‘can it be done’ type of project. But either way, he did it. He had his own reasons for doing it the way he did. If you are going to judge, judge on the fact that IT WORKS! He set out to do something that was non-trivial, and stuck with it until it did what he wanted. THAT is the point of this site, to celebrate [success | tinkering | learning | what ever you want to call it].
I will continue to feed the trolls: I have been a near daily visitor to this site for years. Back when there were few comments, and even fewer posts (less than one a day on average) I felt the site’s comments were almost worthless. Then H.A.D. Experienced a surge of growth. At that point the comments approched the intelectual level of slashdot comments. Sure, there were people who did not fully understand the subject of the post, but their comments were posted in good faith and it helped spur discussion. Now the comments are truly worthless. I am very close to ignoring the comments much the way I ignore Digg comments.
If you don’t agree with the project or the way it was done, or even if you think you could do better, I think I speak for many others when I say, I DON’T CARE. It it isn’t meant to continue the conversation (and no, putting a project down for any reason does not help advance the conversation. Now if you were to offer a solution to a precieved problem that would be different) I don’t want to read it.
This was typed on my iPhone. Please ignore any spelling errors.
“This was typed on my iPhone. Please ignore any spelling errors.”
Don’t you mean smelling errors? :-)
I like how you use a CNC but you are using a sub 200$ Sony Cybershot point and shoot camera that wasn’t even intended for high quality filming or anywhere near it…
While you are at it, butter your toast in the morning with your face and send that to us, I would like to see that too!
Cool idea, but sitting here with my 3D glasses it does not seem to be accurate.
this is a pretty cool idea, you can maybe even actively change the distance apart at certain points to give it some kind of focusing effect. I wonder what the results would be like.
I’m loving these comments, you trolls are becoming more entertaining than the posts. When it’s too simple, you cry about it, when its overly complicated, you cry about it. There’s no winning with you chumps.
Ok, what I want to see is this hack combined with the previous hack of being able to change the page characteristics so that there’s a 2D and a 3D version of hackaday (at least for the pictured hacks).
that is one cool way of doing it and if he had used 2 cameras how much of a hack would that be?
@Daniel, I could not agree more.
๏̯͡๏
Regardless of how well it works or whether or not he had another camera, it is an interesting concept.
No Arduino. Big fail. Sponsors don’t like these types of posts…
This is similar to how Coraline was filmed. The camera takes one picture moves one side by ~5 inches and take another picture. Animator does his thing and process starts over.
Not the best 3-d. That would make one hell of a stop-motion video rig though.
wow great hack?
@M4CGYV3R
I built a cnc machine (three of them, actually), and I’m planning on using what I’ve learn with it to make a camera control system for stop motion. I suck at programming, so it will probably take more time to make a decent “moving point” tracking system (so I can, for example, rotate the camera around a moving character while he’s moving) than to actually build the rig.
The possibilities are endless. Most independent animators can’t afford to buy a complex camera control system like the ones used in “big” movies; it’s good to be a CNC hobbyist :D
There’s a group already doing this stuff WITHOUT the camera paths and in realtime! The technique is called monoslam (monocular simultaneous localisation and mapping), see youtube.com/activevision
very very cool stuff!
Cool stuff indeed!
Anyway, I’m not planning on building an automatic tracking system (would be too unstable, and my programming isn’t that good anyway); I was thinking on tracking a predefined (by software) moving point, so I can get a smooth approximate of the camera movement (and, more importantly, make fast camera tests before each take, since it would be really, really bad if the camera moved unexpectedly after 2 hours of moving the puppet around). Another idea I have is making a system to record the manual movement of the camera, so it can later reproduce it exactly but step by step (optical encoders FTW!).