How to Scale a Niche Live Video Content Business with AIR Camera Robotics and AI Autotracking

A Conversation with Chris Durkee, Owner of 4 Legged Flix

4 Legged Flix (4LF) is a video production company that specializes in dog agility competitions. We livestream events, and also capture video of individual dogs so that owners can buy it later to use for training purposes or bragging rights.

As a small business, we’re always looking for ways to make things more efficient. We’ve done a few upgrades to our equipment over the years, and moving to AIR camera robotics is one that was definitely worth the expense. 

 

We do about 20 events a year all over the United States, and sometimes internationally. There are three people in the company, including myself. The competitions are held mostly in equestrian centers or fairgrounds, where dogs compete for time across a set course. Events are anywhere from 3 to 5 days long, and we need to cover dog runs in multiple rings simultaneously. Infrastructure varies at the venues, so we usually arrive a day in advance to set up our own network.

 

We use 6 AIR One robots running the Z CAM E2-M4, Stream Deck and vMix over LAN to switch between cameras and add graphics, and then stream out to YouTube. We also record in-camera for individual competitor videos. We mount the robots about 15 feet (4.5 meters) up on a tripod, and will bring them down to about 7 feet (2.2 meters) when we’re recording podium ceremonies. Most of the time, we can manage the robots wirelessly, which saves us a bunch of time when changing positions.

AIR camera robotics and control technology has made a major difference in the quality of content we can produce for our clients and the way we manage our business:

Networking + Setup: AIR’s plug-and-play networking and self-calibrating robots let us do our daily setup in about 30 minutes, literally a 75% reduction in setup time. We can just plug in the AIR One robot, and we’re done! When we move to a new venue, or move the camera from the ring to the podium ceremony, it’s a simple rinse and repeat process. Before AIR, we had to spend hours networking the cameras for each show, and running cable for SDI. We’d have to copy and paste SRT request codes into HTML, which would take us hours.

App Camera Control: The iPad app is brilliant! We used joystick controllers with our previous pan-tilt heads, which created a jerky movement because our guys would push it and stop, push it and stop. There was no software control, so an operator would need to manually manage speed of movement. The AIR app is super intuitive, the head movement is superior to any PTZ, and the software speed adjustment makes it easy to use. Just hold the stylus to the screen and follow the dog! When an op needs to track a dog faster, they just move the stylus farther from the dot, and when they need a more gentle movement, they just move it closer to the dot. Even my seasoned ops will laugh that they can do moves with the touchscreen app that they could never do with a joystick. It’s also convenient that we can access all camera controls through the app without having to bring the camera down. 

Portability: We serve competitions all over the United States, and sometimes internationally. The AIR One rig and all the gear slides in and out of one small case that easily packs away. And when we need to change camera locations during an event, it’s very easy to simply pick it up and move it.  

Efficiency: AIR makes it possible to manage 4 to 6 cameras simultaneously from a central position with a team of three people. I can control the cameras remotely, and my techs can address what’s needed out in the venue.

Customer Service: AIR’s service is great! Response time to issues is fast, and we’ve even had them fly in to service a robot so that we didn’t miss the competition.

AI Autotracking: AIR is responsive to customer requests, and you can see that their mission is to make this system easier and easier to use. 4 Legged Flix was among the first AIR customers to test their AI autotracking, AIR AutoPilot. We found it to be incredibly useful. The AI can immediately pick out differences between objects and all I need to do to train it is drag a box around the dog I want it to follow. It could even track two objects at once – a dog and a trainer – and tell the difference between humans, dogs, and types of dogs. We thought the AI might lose the dog when it went through a tunnel obstacle, but it followed the animal right through to the other side and on to the end of the course! When the dog is done, we just deselect it in the UI, and draw a box around the next animal to run. This saves us an immense amount of time and effort.

Watch AIR AutoPilot in action tracking a border collie running an agility course. We see the AI follow the dog precisely. At the 0:16 mark, the camera operator selects both the dog and the trainer, and we see the camera continue to follow, keeping two objects in frame. At 1:06, the dog enters a tunnel, becoming fully occluded, and the AI picks the dog right back up again after it emerges from the other side of the tunnel. At the end of the video, we see the camera operator deselect the dog and trainer (objects 86 and 87), and select a new dog to follow before the course restarts.