Low budget Bullet-time effect

I think the most complicated part of making any rig of this sort is to create something that doesn't show up in the frame of the video. The White Rabit project solves this cleverly by having the rig over the actors heads. Having a rig a meter off the ground would mean that unless the camera was pointed upwards it would shoot the rig itself all the time requiring massive post frame-by-frame masking and editing to clean up the footage afterwards.

I think the easiest thing to do this would be to make a variation of the White Rabit project where the camera somehow "hangs" down from the main track. Another posibility is to use a bent tube or something like that attached to the centre of the rig that bends down much like the arms in an umbrella with the camera attached on the end. Then you just have to figure out the counter balance.

another element you haver to consider is the motion blur and movement in the frame. spinning a camera fast around an object would cause a lot of interesting flickering, blur or smudge effects especially in the background.

It is a very tricky project for sure. doable but tricky. I think the easiest way to do it would probably be a circular doly track and some fast people to run it around. Take a look at the extra features on the "Saw" DVD and see how they did something simmilar. they did it in super fast motion but the exact same effect can be acheived by running fast and then slowing down the footage.
 
that's pretty cool but again... he wants something that's low budget
this guy used 24 cameras... 229$ each

i think the green screen idea is the most practical when keeping the budget in mind... the programs are around 100$ or if you try really hard they could be found elsewhere for free
 
Every camera is controlled via an electrical circuit, I'm sure some ultra cheap digital cams could have their innards soldered so that you could connect them all to one central button to take a single still from multiple cameras at once.

Let's say you took 30 1.3 megapixel cameras and soldered them all up to one button...between solder, wire, button, and cameras you'd be looking at prolly about 800-900 USD.

If you wanted more motion, I think you could skip the button, and just line up all the connections on a board and just pass a wire over all the connections...it wouldn't be an accurate sync, but I'm guessing no one is trying to have an accurate 24fps to have dialogue.

also, as much as most people hate ewe boll, he helped to create a very low cost bullet time effect in House Of The Dead, where he basically just spun a camera at high speed around a subject......theres actually a thread here recently that uses a large bearing to have rotating shots. reinforce it a bit, add a motor or hand crank, and bam you can just mount any ol' cheap mini-dv camera(I wouldn't trust any of my good cameras on such a rig) to the rig and start rollin. http://www.dvxuser.com/V6/showthread.php?t=174327

I think a large two bearing system that has three triangular supports that connect outside of the center, and which each has an equal ballast(decided by weighing camera and it's support system) would be able to be spun at quite a high speed, while avoiding having the support systems viewable by the camera....although you'd need to carefully choice focal length.

but I digress.....I think this is very possible on the cheap.
 
I thought I'd study a bit more into this, and made a small prototype for doing bullet time:
http://www.youtube.com/watch?v=cAdWdZTXbUM

This one is just a simple bearing with some various hardware and PVC, but a better model would be an electric motor which is ceiling mounted(or tree, building, etc) that attaches to a board or some such thing which has the camera mounted at one end and ballast on the 180. You spin the rig while running at the highest frame rate possible(this video is at 60fps in a 24fps timeline) with the talent in the center and have them do their action...to accentuate the effect you can slow down the action in post....of course mine was quickly slapped together and you can see the rig itself, so I'll work on tilting the camera a bit so that you can't see the PVC, and in a final model this tilt would also help to avoid seeing the ballast.

I'll update this with a vimeo link once it's finished uploading, although it'll be awhile til they allow the HD cut, as I've allready uploaded an HD clip this week:
http://vimeo.com/5499254

One last thing...to get a good estimation of how this kind of shot would look give all your attention to my fist in the center of the frame.
 
Last edited:
I was thinking further into this, and how the rig would basically be pretty limited by
1. The speed of rotation
2. Speed of film

There is an inverse relationship here, because the faster you spin the rig the less frames you capture around a subject...I myself am using an HVX-200, so I'm limited to 60 frames per second. I believe for a half decent amount of pictures on a 180 you'd need to spin the rig at 1 revolution per second for a total of 30 frames....the problem here is that this is over a quite extremely long 1/2 second, and alot of movement can happen in that amount of time. You could rotate the rig twice as fast and only depend on 15 frames on a 180 at a total of .25 of a second, but I'm unsure if that is enough material to work with.

Looking further into this for more effective and low cost options I ran into the Casio EX-F1. It costs about $800-1000 USD and can capture at a resolution of 512 × 384 at a jaw dropping 300 frames per second! The res isn't all that amazing, but I think this can be forgiven while in a super cool matrix shot...people would ignore resolution(hopefully) and concentrate on the novelty of the shot.

I hope I did my math right, but if you could spin a rig at 5 revolution per second while running this camera at 300 frames per second you could get a total of 30 frames on a 180 in the amazing time of just 1/10 of a second!

All that at a price tag between rig, camera, etc of a bit over a grand...not bad.
 
I believe I'm putting these ideas towards a rig.....my first rig will spin at 30rpm,eventually be able to be run at 250rpm, and probably use the cheaper Casio EX-FH20(210fps).

One of the biggest problems is the increased amount of G's at higher speeds....if the stuff attached to a motor weighed ~10+ pounds, the ends of it would be putting up with somewhere around 1-2 tons of force(10 foot circumference).

These problems make one need to beefen up the stuff attached to the motor, thus increasing weight, and again increasing the amount of force.

One solution I've come up with is a circular track which is powered by a blade sticking out from the motor...this blade will push/pull a device with wheels, mount, camera, and since it isn't physically attached to the blade/motor the surrounding track and camera mount assembly can be beefed up as much as needed to sustain the force of the rotation....of course I can't go to overboard on the camera mount assembly, as this IS the point at which the most force will be created.

The first rig shouldn't be all that hard to accomplish...30rpm. The forces will be pretty low, and I should be able to build the first prototype out of wood. If my math is right I should be able to get 210 images on a 180 over a 1 second period...this isn't ideal, but it should look fairly cool...I'm unsure what the breaking point of the wood rig will be, so I'll most likely do alot of tests with dummy weight to find the breaking point, and then only run at 60-70% of this breaking point. Hopefully I can even get to 60rpm, as while this still isn't my ideal speed, I think stuff will start to look REALLY cool with 105 images on a 180 at .5 seconds

Any advice, thoughts, etc. would be greatly appreciated.
 
Last edited:
You must log in or register to reply here.
Back
Top