26 June 2024

In the belly of the beast

We sat down with Jason Elliott, Director of Photography for MrBeast. Jason has been intimately involved in many of the channel’s wildly successful videos, and revealed insights about how his team brought Jimmy’s vision to life on hundreds of millions of screens around the globe.

By Atomos

From humble beginnings in a small town in North Carolina, Jimmy Donaldson (AKA ‘MrBeast’) has built an unparalleled empire in the sphere of YouTube content creation. His primary channel boasts 272 million subscribers as of this writing — not to mention the millions of additional subscribers that follow his secondary project BeastPhilanthropy. Jimmy credits his rise to prominence to his unique blend of ‘IRL’-style content, ad hoc comedy, and consistent reinvestment of profits into the brand that often manifest as overwhelming gestures of philanthropy towards everyday people.

Perhaps his most recognizable trademark, however, is the massive production behind his ultra-viral YouTube videos. From replicating the set of the hit Netflix show Squid Games in his North Carolina studio to crushing Lamborghinis in hydraulic presses, everything about MrBeast says “go big or go home.” While many marvel at the scale and cost of such stunts, the video technology behind capturing them is just as impressive.

A recent MrBeast video involving such a stunt, titled ‘Ages 1 – 100 Decide Who Wins $250,000’, was one of the most complicated in the channel’s history. The video covers a multi-day competition where 100 individuals, each representing a different age, live inside clear plastic cubes and participate in various tests to determine who gets to go home with a quarter of a million dollars. The game tests the players’ strategy, loyalty, and determination on a hilarious and emotional journey as they attempt to survive Jimmy’s relentless elimination challenges.

Here’s what Jason Elliott, the Director of Photography for MrBeast, had to say about the dedicated work that takes place behind the scenes.

What was the scale of the project from a technical perspective?

This project was massive. Luckily, we already had a bit of foresight about how to set it up from our previous ‘Ages 1-100’ video, but this time we knew we wanted to push the limits even further. We rigged every single contestant cube – all 100 of them – with 3 cameras each, plus a tablet with a custom game app. That’s 300 cameras and 100 tablets, just as a baseline. We used a combination of GoPro Hero 12s with media mods and Marshall POV cameras. We also placed a few cameras in the ceiling and perimeter of the play area, along with several handheld operators that captured the interactive content.The storage needed was just mind blowing. Even recording everything at 1080 30p, we still managed to record several petabytes of data.

Can you describe the workflow?

Our team is accustomed to giant projects, so we’ve really dialed in our process. Once we get the approved video ideas, we get straight to work laying out our camera plans. This includes things like what cameras we will use, how many cameras, camera placement, signals routing, recording options, data estimations, lighting, installation schedules, among several other factors. We start in broad strokes and refine as we get closer to the shoot date. Then once the set is ready, we get to work with setup and testing.

After the set is fully rigged out, a story is created in real-time. The creative is there as a guidepost for us to build our tech and capture plan, but what happens during filming is completely up to whoever is in front of the camera. Nothing is scripted, nothing is planned. It’s all very in-the-moment, and we have to be ready to adjust our strategy at the drop of a hat. Being prepared to capture those unexpected events as they occur is crucial.

Then, from a post perspective, a lot of effort goes in to how we tell that story. Given that we record every single moment from every single angle, we rely heavily on our story team to take detailed notes on content to give the editors. A lot of credit goes to our amazing post team as well. They absolutely grind through processing the footage in increments, slowly chopping things down into manageable bits before sending it off to our finishing editors. To say I’m glossing over their contribution is an understatement; they collectively put in hundreds of hours per video to make the content really shine.

“This time we knew we wanted to push the limits even further.”

What problems did you face, and how did you solve them?

The method of how we cover such massive videos can easily lead to complications. Basically, we use a LOT of cameras and generate a LOT of data. Because a lot of MrBeast content is made up of those IRL moments that are unplanned and unscripted, all our cameras must be actively recording for 24hrs a day until the content is over. In the case of a challenge like ‘1-100’, this inherently creates several challenges wherein the solution ultimately boils down to what cameras we use and how we record them.

Our very first ‘1-100’ challenge ran for almost 8 days. As you can imagine, that would create quite a lot of additional labor if we had to continuously go in and change the batteries and memory cards. While that could technically have been done, it simply wasn’t a feasible option based on the ongoing disturbance it would cause the players. Imagine a contestant who is exhausted the next day because we kept them up all night constantly changing batteries. Our camera team has a policy of non-interference or influence when it comes to contestants. No matter how we go about physically filming something, it cannot interfere with game play. We keep it as real and fair as possible. So, we made the decision to hard power everything and record it all remotely using the Atomos Shogun Studio 2s.

The power part was relatively simple, but recording presented a mountain of problems to solve. We ended up running nearly 56 miles of cable for that video — and that was only for the 200 cameras (we only did 2 per cube in that video) and some ethernet cables! Then came the actual recording hardware. We had previously discovered that recording tons and tons of cameras presented a very real media shortage issue. Often there would be hundreds, if not thousands of SD cards floating around. Back then I made the call to use the Atomos Sumo 19 whenever possible in its 4 ISO mode.

While this offered a nice way to quickly switch between and view 4 different cameras on a large monitor, the ability to record remotely (and all 4 cameras with one button), the biggest benefit was the overall physical media reduction on set. 4 different camera feeds, all recorded onto one single piece of media. Not to mention the affordability of the media, especially when you consider it at scale. The Sumos were a great solution for us at the time, and we still use them quite often.

Once camera numbers really started to grow as we planned for our second ‘1-100’ video, we knew we needed to become even more efficient. This is where the Shogun Studio 2 came into play. Having 2 monitors on board, being rack mountable, having front-load SSD slots, accepting timecode, and not taking up a ton of room, it was a clear-cut choice. We doubled down and brought in racks and racks of them — 46 units in total on the second ‘1-100’ video. In short, they gave us a super-efficient way to record and monitor vast quantities of cameras, as well as being modular, allowing us to easily expand or reduce our racks based on the video needs. Plus, as an added bonus, it looks really cool on camera when they’re shown in any control room or BTS shots. Naturally, there were many other technical challenges, but I’ll have to save those for another time.

“(Shogun Studio 2) was a clear-cut choice. We doubled down and brought in racks and racks of them.”

Was there anything you wanted to do but couldn’t?

In a perfect world we’d want to record all 300 feeds in UHD, but the cost simply doesn’t make sense just yet. Additionally, we use PTZ cameras a lot in these types of situations and, well, the manufacturers mostly focus their products on the farther end of focal ranges. In other words, we’re keeping an eye out for wider options in the PTZ world than what is currently available, ideally one that we can fit in tighter spaces and give our viewers a more dynamic experience from up close. Beyond that, our camera team is made up of some truly talented individuals who always dream big, so we’re always thinking up crazy new ways to film things, and we often do figure out a way to make it work. So, while there are usually several things we want to do but can’t — we almost always find a solution eventually. There is certainly a lot that we’d like to do in the future, although I don’t want to give anything away just yet.

Where did the idea for ‘Ages 1-100’ come from?

I believe “Ages 1-100” was one of Jimmy’s earlier ideas from when the channel was smaller, but he realized then that they didn’t have the resources to pull it off at that time. So, he waited, and when the moment came where we had enough resources to not only make the video, but make it well, we finally set things in motion. After the success of the first version, we decided to bring the concept back while also expanding on it creatively. The big difference here was putting the challenges and gameplay in the hands of the contestants. Instead of individually competing in challenges or doing so as groups, it was about truly letting them decide who a winner would be from within. That, and many more cameras to see it all unfold.

Has it given you ideas for future projects?

Absolutely! There’s always so much to learn from each video, from both a technical perspective and a story perspective. As with all future ideas, I can’t reveal them… but I know that we as a camera team, and as a channel, always aim to do our best and push the limits of what is possible to make our viewers’ experience better and better with each new video.

Keep Exploring

Case Study

Shooting anamorphic with Tito

Case Study

The Creator: shooting a Hollywood blockbuster in challenging locations

Creator Content

LA Clippers: producing social media with Ninja Ultra