Bullet Time Video Booth

preview_player
Показать описание
When my cousin asked me to build a video booth for his wedding reception I might have gone a step too far...

00:00 Intro
00:20 What is a bullet time video booth?
02:01 The cameras
03:01 Camera arrangement and processing
04:30 The components
05:00 The quarter circle stand
05:25 Mounts
05:54 Power supplies
06:44 USB hubs
07:58 Camera trigger
09:04 Sony a5000 and HDMI grabber
09:24 Big push buttons
10:00 Lights
10:43 Conclusion
11:46 Outro and examples
Рекомендации по теме
Комментарии
Автор

I bet when your cousin asked for a photo both for his wedding, he was not quite expecting this! What an absolutely amazing project with so many smart hidden solutions!

tolstoj_
Автор

the effect looks 100% like The Offspring’s „The Kids Aren’t Alright” music video, love it!

fortekkk
Автор

This is executed extremely well—what an amazing video, thank you so much for sharing, and I'm glad you were able to relay the pitfalls and learnings so clearly (like upgrading from a Pi 3 to an older Dell laptop so guests didn't have to wait so long). I could see a company building out a more reliable version of this and renting it out for events, it looks great!

JeffGeerling
Автор

Brilliant, you are the desired guest at every wedding or party!

DenisShiryaev
Автор

You could also use a laser lever to line up all the camera lenses during setup, just keep the lenses cover on while you do it to not damage the optical sensors.
To prevent guests from bumping into it, ask the venue if they have spare velvet rope to setup around the area or bring your own.

_B.C_
Автор

What an underrated channel. Fantastic job! The capability of cheap "old junk" tech is awesome.

AnthonyCotto
Автор

I'm also a big fan of gphoto2 + canon xti! Cool to see someone else still using them. Awesome project.

haun
Автор

And _THIS_ is why I have the notifications on for you. Great video!

yuGtahT
Автор

I love the quality of your videos, especially the 3D work... They are awesome.

masamarrr
Автор

this is just too amazing. sometimes I'm glad i live in this timeline

Tamonduando
Автор

Wow, this is such a fantastic project. I love hearing about all the long-tail problems these engineering projects run into. What a final result, though!

PaulHobbs
Автор

Nice Project and what a unique wedding activity!! - We're currently developing a 3D-Photogrammetry-Scan-Rig at the University in Kaiserslautern and I see a lot of parallels to your Rig! Especially finding a proper solution for powering and triggering the Cameras (we're currently use 18 Canon 2000Ds with the prospect/aim for a number of 32 and more!). I really dig your channel and the well done content foremost your enthusiasm, storytelling and the animations! Keep it up :)

michelvaillant_
Автор

this is some rock star shit homie good job!

iquemedia
Автор

They had something like this setup at the science museum when I was a kid.

fdg
Автор

Excellent work! What a thoughtful and meaningful gift for your cousin, and a great project. :-)

I resonate with your concerns on USB, and potentially switching to SBCs with ethernet connections instead. I'm planning for a similar transition with one of my own projects, but that comes with it's own challenges huh! Keep up the great work!

flibidydibidy
Автор

Awesome content! I’m building something similar but for photogrammetry of people for VFX and I faced the same problems! 😅

santennio
Автор

Ooof! Me still shooting pictures with my Canon Rebel xti 😆

joeldezwaan
Автор

Amazing. I saw recently someone doing "bullet time" by swinging an insta 360 around their head. Not sure if would work so well at a wedding but might be another option.

aspzx
Автор

How did you do that the video Camera takes the video only for a certain time after triggering ?

cedricziller
Автор

Googles NeRF Neural Radiance Fields method should dramatically reduce the number of cameras needed to do this effect.

You can probably get away with 4-5 cameras and still get a smooth artifact free bullet-time effect. And the exact of cameras doesn't matter either, as long as they all capture approximately the right thing.

Processing is very slow, so you'd still need something else for the preview. Perhaps the "Direct Voxel Grid Optimization: Super-fast Convergence for Radiance Fields Reconstruction" paper, which has published code, and can do a NeRF-like reconstruction in 8 seconds to reasonable fidelity?

olivermattos