Since the release of The Matrix over a decade ago, we've all seen the incredible "Bullet Time" segments done by many different people, including an incredible surfing video using the technique, but something I've never seen before is this technique used on something as large as a car. The Splice Boys, a team comprised of Richard Kendall and Tom Brandon created this massive camera moves using dozens of cameras all rigged to go off at specific intervals.
The shoot was a tough one, taking almost 10 hours to set up some shots in freezing cold weather on the tops of mountains or getting nearly burned to a crisp in the desert, but it came together well and I think that the finished product is absolutely beautiful.
So what do you think? Is the effect overdone? or have The Splice Boys come up with a new and interesting way of using a well known technique? Let us know in the comments below!
very time consuming and tedious but that's usually when the best work is produced
My thoughts exactly. What a bitch to set up... but the results are really impressive
stunning!
Incredible!
I notice that the cameras are triggered sequentially (not at the same time) as they did in MATRIX. As you you the spin, the car keeps moving a bit. Much better than the "static spin" imo. Congrats.
I should try that next time I have 75 cameras, 75 triggers and the appropriate controller... And a nice car. And a bunch of friends. And two full days of fair weather.
I'm always happy to see good use of that effect. Well done ! Also check out this new video from the Big Freeze guys, with 96 cameras !!! Outch ! https://www.facebook.com/photo.php?v=564481366920006
Cool, but conceptually, I think there should be even more movement of the Rover to be optimally effective in the context of an ad. I want to see it tearing up the terrain. Great concept though.
The frame rate made this less awesome. It's crummy because so much work and the idea is beautiful, but stills are just a crappy choice for this specific movement. High speed cameras (yes I know much more expensive) would have been nicer to get that movement. The shots were great and the execution looked nice but I just need a better frame-rate to enjoy this.
Do you realize how fast you'd have to physically move a single camera to get those shots? Even if they had a Phantom Flex (or even a RED Epic) to do the high speed capture, they'd have to be able to whip the camera around a motorized dolly track of what looks to be nearly 100' in a fraction of a second. Lets be generous and give them 1 second to make the camera move of 50' - that's 3,000 feet per minute. Think about that. You'd have to move the camera on a curved dolly track, in a remote location at 35 MPH. And since Those shots took place over a longer distance and in a shorter time. it's probably closer to 75' in about 1/10 of a second. (75*10*60*60)/5280 is 512 miles per hour. Good luck with that. At 2570FPS (max for the Phantom Flex) you'd only get 257 frames - not even 3 times as many as with the Splice Boys rig above (about 90 cameras from what I counted) so it would still look stacatto, although not as much. with a RED, they'd get 120 FPS, so 12 frames over that track, so almost 1/7 of the number of frames.
You mistook my comment. I guess it wasn't very specific. I meant it would be better to have 75 or 90 or whatever high speed video cameras, same set up, but then you can take the frames from all the cameras and gradually move. Although 250 frames would probably still be a lot nicer :P. I'm just saying the staccato completely ruins what could have been a fabulous look for me. Instead of elegant, which may not be what they were aiming for, it looks amateurish.
Or it may have been the editing. I'm not sure, but all I can say is that the frames look too separate
OK I take it all back. I think it's the most likely case that some of the cameras for on slightly off angels. The ball heads they used could have easily been slightly off resulting in a weird shift between frames. The problem so it seems would be a lack of continuity between cameras and a more absolute system of camera placement should be considered for future work.
I agree with you, I really feel the same way about the smoothness (or lack, thereof) of the movement.
What we should also realize is that the bullet time smoothness we see in a movie like the "Matrix" is due in part to CG frames being interpolated, and stuck in between, the actual frames in the final rendered scene
David, you seem to have an insanely keen eye for how 90 cameras may not be on the exact same plane. Having watched this video several times now, I only see this in about 3-4 frames. If the solution to the problem is so simple ("a more absolute system of camera placement"), perhaps you should go out and do it better?
There is a gap between each camera so the image will appear to jump between each frame. The only ways to fix this are A) spend hundreds of thousands of dollars (probably significantly more) hiring a team of VFX professionals to interpolate frames between each real frame or B) Spend hundreds of thousands of dollars (again, probably more) developing a camera dolly motor that can take a Phantom Flex (and all of its accessories) around a flexible track, in the wilderness, at over 500 miles per hour.
Do you know what I see when I watch this video? I see an insane amount of hard work and expertise going into a project that takes a fairly common technique to a new level. I see 90 cameras that had their shutters sync'ed on a delay to allow for super high frame rate video to move over a long distance. I see power and equipment problems and solutions for tough locations. I see a visually stunning ad for a car.
You see a few frames out of place. I guess you win. Have an Internet Point.
But 3-4 frames is more than enough to make it less than perfect. Just as 1 part per million oil in water is enough to make the water taste bad, 3-4 frames of 90 is way more than enough to ruin this process for me. I would be happy to do this better if you'd like to commission the project. Although as I am not an idiot I will not invest so heavily proving you wrong.
The physical gap is not likely the issue. After review it is certainly the direction of a select camera(s) being slightly off. Which sucks considering how much work the rest of the shoot was. However your estimates on fixes are still insane. 10K would be enough for interpolation of frames, as I've hired VFX teams to do this in the past, although it's truthfully not always great, and in this case may not be necessary. Additionally I am confident I could build a dolly/track for 50K. 35 MPH is doable and you would not notice it as being much faster, in fact this video probably could have been not as slow in order to show a bit more life in the vehicle, but that's another issue. The rig is possible, and affordable. This I know.
You know what I see when I watch this video? I see a mistake. Syncing 90 cameras isn't very difficult either, perhaps a thousand dollars of hardware and a little programming and a solution can be found for not so much time/money/work, power and equipment problems are no more intense than any location shoot, so no biggy there, and if these guys are so brilliant, I see a stupid mistake in having a camera pointed in a slightly off direction making them appear a lot less brilliant to me, which is a shame.
My comment is purely "This could have been perfect, I'm sad it wasn't as I enjoyed the idea so much."
Most people are not as anal as I am about perfection, and indeed 99.8% of viewers wouldn't care one way or another about 3 frames being out of place, but I noticed it and I care. With these guys calling themselves specialists, I would imagine they're not 100% happy with the end product either.
- Dave
I have to say, I agree. Not that the end result isn't great but there are a few frames that are slightly off and they probably see it also. It's not something that makes the whole thing suck and it's actually pretty hard to see, so I doubt most people would notice, but it's there. I know personally I have done things like this where I got done with everything and noticed a small mistake and it just killed me. I'm sure they feel the same way.
sure there is a gap between each camera - but the lens's field of view covers that gap.
How would this be any different than what was done with the still cameras? You'd just have the same "jumps" between the cameras (actually, much larger, more jarring jumps due to the larger size of the cameras). The only difference would be that you could watch a few frames from the same angle, then jump to the next camera. There's no net benefit of a "smoother" image here. Plus the insane cost.
but this wasn't done a with a single camera... so you're entire comment makes no sense. David Dvir simple said he would've liked a higher frame rate - so in the case of this shoot they could've used twice as many cameras to have a smoother playback
I think the issue for me is because of the lack of 180 degree shutter everything appears too staccato. If it was possible to have a rig that shook to emulate the camera movement that's when I would look really good. Either that or a super fast track.
In this case, the staccato nature of the video is from the physical distance between the cameras, not the shutter angle. For the super fast track concept, look at my reply to David Dvir's comment.
But the shutter angle and lack of physical camera movement therefore creating no motion blur play a part in the stacatto effect.
I think they should have gone for a track as slow as they could have gone.
Obviously I realise it's not the same scale but this looks so much better: https://vimeo.com/33408157
Before matrix there was ACCEPT
http://www.youtube.com/watch?v=9el2lg2olpE
I watched it twice. I'm not feeling well ::)
thanks for pointing that out...