We see virtual production stages everywhere, usually framed by discussions about LED panels, processors, and tracking systems. What gets less attention is what actually happens once the lights are on and clients walk in. Virtual production is not only for blockbuster series or studios with Hollywood-scale budgets, and not everyone is working on the next The Mandalorian. For everyday work, the realities are different and often more interesting.

To look past the tech diagrams and into daily practice, we spoke with the team at Door G Studios in East Providence, Rhode Island. Executive Producer Jenna Rezendes, VP technologist Joe Ross, and Unreal Engine Artist Sam Culbreth walk us through how virtual production functions on real commercial shoots, with real constraints, real schedules, and real clients.

Door G Studios is a creative production studio based in Rhode Island, specialising in commercial content, branded work, and campaigns that blend traditional production with virtual production. They operate an LED volume designed for practical, day to day shoots rather than cinematic excess. You can find them at door-g.com, explore their studio at door-g.com/studio, or keep up with them on LinkedIn.

DP: Hi Jenna, what is Door G?

Jenna Rezendes: Door G is a production studio based in Rhode Island. We have a versatile team of producers, creatives, and artists who love storytelling and keep pushing what’s possible in production. We work on branded content, commercials, and large-scale campaigns. We’re constantly exploring new ways to make production smarter and more efficient, whether that’s through virtual production or traditional filmmaking techniques.

Joe will share more about our physical stage and how it expands what’s possible for our clients, and Sam will get into how our production pipeline supports that. But from a big-picture perspective, Door G is a creative studio that stays ahead of what’s next in technology without losing sight of the human side of storytelling.

DP: What kind of customers use your stage? And what are the usual questions they have when they first come to you?

Jenna Rezendes: We work with a wide range of clients, from agencies and brands to production companies that just need the right space and support. Some clients come to us early on, looking for full creative ideation through production and post. Others already have their concept developed and just need production services, or even a studio rental. We’re set up to scale our support to meet their needs.

When people first reach out, the questions are usually about what’s possible on the stage: how virtual production actually works, what kinds of environments we can create, and how it impacts things like schedule or budget. We have a library of production-ready virtual environments that can be used as is, customised to fit the project, or fully custom-built when needed. It’s exciting to see clients realise that the stage isn’t just a technical tool; it’s a creative space that can open new ways to tell stories.

DP: How has your client communication changed since you started offering virtual production?

Jenna Rezendes: Communication has definitely evolved since we started offering virtual production. A big part of what we do now is education, helping clients understand how this technology changes the flow of a project. The conversations start earlier, and they’re more collaborative from the beginning.

With traditional production, clients are used to making creative decisions once they’re on set or in post. With virtual production, much of that decision-making moves to the front end. We spend more time in previsualization, running test sessions, and showing how scenes will look and feel before anyone arrives on set. That shift has actually deepened client relationships. They’re more involved in shaping the vision early, and they get to see ideas come to life in real time. Once they experience that, they realize how much more efficient and creatively freeing the process can be.

DP: When props are built for a shoot, say a few shelves or counters, can’t they be stored and reused for future productions?

Jenna Rezendes: Absolutely. One of the biggest sustainability advantages of virtual production is that so much of what we create lives in the digital space. Instead of building large practical sets that are used once and torn down, we’re designing digital environments that can be updated, customised, and reused across multiple productions. That dramatically reduces material waste and the carbon footprint that comes with traditional set construction.

For the physical elements, we do need to blend the virtual environments with the practical; we keep things minimal and modular. We’ll often store practical props or set pieces between shoots so they can be repurposed in future environments. It’s a smarter and more sustainable way to work, both creatively and operationally. We’re proud to be certified with the EMA Green Seal for sustainable studio practices, and that really guides how we think about production.

DP: What’s the most common misconception clients have about shooting in a VP stage?

Jenna Rezendes: One of the biggest misconceptions is that virtual production is only for creating worlds that don’t exist “on earth”, like post-apocalyptic landscapes, outer space, or big cinematic settings like The Mandalorian. And yes, it’s incredible for that kind of work, but it’s just as effective for everyday environments that are tricky or difficult to film in. We’ve used virtual production for medical facilities and retail spaces where on-location shooting can be disruptive, expensive, or constrained by logistics. With virtual production, we can easily recreate those environments with total control over lighting, timing, and camera movement.

Another misconception is the “fix it in post” mindset. In virtual production, so much of the decision-making happens on the front end. You’re designing and visualising the environment before you shoot, so there’s less guesswork later. Once clients experience that process and see the world on the LED stage in real time, it changes how they think about production entirely. It’s not about doing everything in post anymore; it’s about getting it right in the moment.

DP: When you prepare a shoot, what do you wish every customer already understood?

Jenna Rezendes: I wish every client understood how much smoother and more creative the process becomes when we’re all aligned early in pre-production.  There’s really no such thing as a perfectly prepared project, but the ones that set us up for success are those where clients share the right materials and collaborate with us early in the process. When communication is open from day one, and everyone is aligned on intent, technical needs, and delivery formats, it creates a strong foundation that saves a lot of time later. Clear references, organised assets, and early creative collaboration give our team what we need to prepare for production as effectively as possible.

Especially with VP, a little extra time up front for pre-production and previsualization makes a huge difference once we’re on set. Our post-production process is already streamlined, so the last thing you want to do is condense your pre-pro and pre-vis phases. That’s where we work through creative and technical challenges before they ever reach the stage, which ultimately leads to a smoother, more efficient shoot.

DP: And just to end on a fun note: you’ve spent quite a few minutes inside that stage. What’s the most amazing thing you’ve seen so far?

Jenna Rezendes: Nothing will ever top the first time I saw virtual production in action. That moment when it clicked for me that this technology wasn’t just for Hollywood but could completely change how we tell stories across all kinds of productions. Seeing a camera move through a digital environment in real time, with everything reacting naturally, was honestly jaw-dropping.

Since then, every project has its own amazing moment. Sometimes it’s watching a client experience it for the first time and seeing that same lightbulb go off. Other times it’s the thrill of seeing our team pull off something that felt impossible a few weeks earlier. The technology is incredible, but it’s those human reactions, the excitement, the surprise, and the creativity, that make it truly special.

DP: How do you manage creative decision-making between physical and digital departments?

Jenna Rezendes: Our production design team, Unreal artists, and DPs are in conversation from the earliest stages of a project so that creative ideas and technical realities stay aligned as things evolve. The goal is to ensure that what looks beautiful on screen also makes sense within the constraints of lighting, camera movement, and real-world space.

We approach it as one unified team rather than separate departments. The virtual art department might start developing the environment while our physical team experiments with lighting or set integration, and those discoveries flow back and forth until everything feels cohesive. That collaboration lets us balance creative freedom with practical execution, so nothing gets lost in translation between the real and virtual worlds.

Joe Ross, our technical director, and Sam Culbreth, our lead Unreal Engine designer, can speak more to how we balance creative with technology. They do a fantastic job creating the bridge between the two.

The Stage & Its Hardware

DP: So Joe, I hear you have a very big TV at work.

Joe Ross: Haha, I actually also refer to the volume as a big TV too! And the Brompton processors are just oversized HDMI adapters. The volume is 17m wide and 4m tall, so unfortunately, you’d probably get whiplash trying to watch TV across the whole thing. Those HDMI adapters I referred to are the Brompton SX40s.

We have three of them pushing pixels to our wall. 17,825,792 pixels to be exact, the full resolution is 8704×2048. We’re using a combination of InfiLED’s DBmk2 and X panels. The Xs are for the curve in the middle, since we have a curved L-shaped wall, which uses curved panels for the bend.

We went with that shape since it allowed us to make the best use of our studio size. It pretty much extends the full length of two sides of our studio. That allows us to shoot longer shots on one side of the volume and wider shots or trucking shots on the longer side of the wall.

Most often, we actually shoot into the curved corner of the wall, since it gives us the most flexibility to move the camera around without catching the edges of the wall, and it makes use of reflections and ambient light from both straight sides of the volume.

In terms of camera: We work with any sort. The refresh rate on our panels is fast enough that most modern cameras have no problem with any tearing or banding while shooting into the volume. We even have some people shoot with their phones with just default settings and have no problem. That said, we have an FX9 as our in-house camera, and mostly rent the Venice 2, RED V-Raptor [X], or Arri Mini LF. The RED Raptor in particular pairs nicely with the volume, since it has a global shutter, which helps even further with any refresh rate problems.

It also ties in with our Brompton processors and allows us to record two separate feeds from one camera: one with the 3D Unreal Scene on the volume and one with just green screen on the volume. The processors just alternate between the 3D scene and green, then the RED captures a separate feed for both, alternating each frame. It’s pretty wild! 

DP: How long does it take to set up and calibrate the stage for a new project?

Joe Ross: That’s definitely an “it depends” kind of question (laughs). If we’re doing a basic shoot with all of our standard in-house gear and well-tested Unreal levels, then everything is pretty much ready to boot up and go. So that may only take half an hour to an hour. We just need to reboot all the PCs to start fresh, connect them all to our Facilis server where the projects are stored, then launch a multi-user server for all of our Unreal machines to sync to.

If it’s a brand new shoot with rental gear and Unreal Projects that we haven’t shot in studio with before, then it can take much more time, up to a full day, to ensure all the pieces are playing nice with each other. The things that would require extra time are things like, calibrating the mo-sys camera tracking offset from the camera sensor to make sure the tracking is accurate, setting up and testing communication between our Unreal editors and render node and any on set tech that needs it, like a robot camera arm or DMX lights, then going through our shotlist and camera moves to make sure the Unreal level is playing nice with the full setup.

The robot arm is a big one, since our studio and Unreal Engine nDisplay config is set up by default to our Mo-Sys star tracker’s map. When we bring in the robot arm, we need to configure our nDisplay config to match the positioning and tracking system coming straight from the robot arm. We could just keep the Mo-Sys on the camera while it’s on the arm, but then we lose some of the benefits of having a robot arm tied into Unreal, plus our model of the Mo-Sys is older and pretty bulky, so it adds some decent heft to the rig.

DP: How do you collaborate with cinematographers who might be new to LED volumes?

Joe Ross: We, of course, do our best to “fix it in Pre” and have the DP and Director heavily involved while building out the virtual scenes, so they have a good idea of what things will look like when they step on set. From there, honestly, we keep things pretty close to traditional production from their standpoint. They just need to work with our virtual gaffer and the practical gaffer when lighting changes are needed in the scene.

That said, we have a few key pointers for first-time VP shooters. One is to try to keep the brightness up on the volume and stop down or ND down in the camera to get the best performance out of the volume in terms of colour and dynamic range. Another is to keep an eye out for any moire; if the focus plane is too close to the wall, you’ll start to see it. 

DP: A stage is more than a screen, a camera, and a server. Can you tell us about the rest of your setup?

Joe Ross: Yeah absolutely. For tracking systems, we use the Mo-Sys StarTracker, so our studio ceiling is covered with tracking dots. Probably more than we need, but that gives us very stable tracking, even when flags and lights get in the way of the Mo-Sys IR camera.

We have a wild wall as well, which we made a custom rolling rig for, so we can move it around the studio as needed for reflections and ambient light that matches our scene. That’s really all we have in-house, but often we will rent robotic camera arms that communicate with Unreal Engine. Those are a perfect pairing with the volume because they allow us to not only have very interesting, repeatable moves, but also to tell Unreal Engine where the camera will be before it gets there, cutting down on any delay between camera tracking and the perspective shift on the volume. 

View this post on Instagram

DP: You recently shot for a medical imaging company, and I heard there was sand involved. Please tell me you didn’t bring a bucket of it near your panels.

Joe Ross: Haha, yeah, that was my initial reaction when someone wanted to bring sand into the studio as well. There have been a couple of shoots with sand involved, actually. The one with medical equipment was a spot that cut from a hospital scene to a soldier stationed out in a desert. For that one, there wasn’t any physical sand used, but our virtual and physical art teams did a great job making it look very convincing!

The other instance of sand was actually for a music video shot in a sort of dystopian desert, city ruin type scene. For that one, we actually did bring real sand into the studio! I was definitely a little concerned about the grains of sand in the air but we kept it contained on a raised stage and kept it removed from the wall at least a few meters. We made sure no fans were used on set while the sand was there as well, and we had air scrubbers running in front of the volume to pull any particles down and away from the LEDs. 

In terms of keeping the volume clean, it’s tough to regularly clean the whole volume, but when needed, we’ll use compressed air or a vacuum with a gentle attachment to remove any dust or debris that may build up in the shader layer of the LED panels.

DP: When one of those LED panels goes dark, what happens next?

Joe Ross: Ah, the dreaded black panel… Well, it depends on where the issue stems from. So the first thing to do is diagnose whether it’s an issue with the panel/LED diodes, the internal PCB, or even software. If it’s just the LED panel itself, we will just pop it right out and replace it with spare panels that we got with our original order. It’s important to get extra panels with a volume like this because each batch of LED panels will be slightly different when they come out of the manufacturing process. So, having panels from the same batch ensures they match perfectly.

DP: What are the biggest maintenance challenges with LED volumes?

Joe Ross: The heat for sure! It can feel like a sauna in there at times if not managed properly. We definitely underestimated how much of a problem the heat from these panels would be. We’ve run our AC so hard trying to keep up with it that it keeps freezing over, and our facilities manager has had to literally chip the ice off it on the roof to get it back up and running again.

View this post on Instagram

We also have installed a secondary mini-split because the main AC wasn’t enough! Even now, we’re brainstorming better ways to remove heat from the studio instead of trying to pipe cold air in. We may be installing an exhaust in the ceiling to do just that!

DP: What’s the most creative or unusual technical solution you’ve had to invent on set?

Joe Ross: Ahh, this is the fun part! Don’t tell anyone, but I secretly love it when things break, and I need to scramble to fix them (Laughs). The old reliable workaround for many of our Unreal Engine hitches is to just freeze the image on the wall using our Brompton processors directly, while we work on rebooting or fixing issues or crashing in Unreal, without anyone noticing.

I think most of the more fun/interesting fixes happen in Unreal Engine. When things just aren’t working the way they are supposed to, you just do whatever works. One particularly odd instance was when we had a scene that traversed two rooms in two locations, stitched together in Unreal. The talent in the scene walked from a retail environment straight into their living room. For some reason, the lighting would not stay consistent, and lights from the living room would just pop on once the camera reached a certain point.

So after much head-scratching, we figured out that if we just stretch the frustum (basically the virtual camera’s vision) far out to the right side, so that it just thought that it was always in both locations at all times, then it would render all the lighting consistently throughout the move. Not the most elegant solution, but I always say, “if it’s stupid and it works, it’s not stupid.”

The Brain Bar & Pipeline

DP: Sam, with a job title like “Brain Bar Operator,” do you have the coolest business card in the world?

Sam Culbreth: So there are a few terms for it, and not all of them are as fun, but that’s actually only a portion of what I do. Most of my time is spent in pre-production building the environments and preparing them for a shoot.
Then, when we’re on set, it becomes the “central nervous system” of a virtual production stage. From The Brain Bar, we are loading our environments onto the LED wall, monitoring, and making adjustments to the scene. The scene’s virtual camera is connected with our tracking system, so it constantly updates its coordinates based on where we move the real camera.

DP: When you power up your Brain Bar, what’s actually howling in there?

Sam Culbreth: This is more of a question for Joe Ross, as he built our brain bar setup. But the specs are quite powerful to keep up with what we’re doing on the render node. If I had to make an upgrade.. Well.. I can always use more monitors!

Joe Ross: They are definitely pretty hardy. Most of our PCs were built in collaboration with Puget Systems out in Seattle. They’re always great to work with and are so diligent in testing which parts are best suited to your use case. At the Brain Bar, our machines run AMD Ryzen 9 CPUs and Nvidia 4090 GPUs. Our render node has an AMD Threadripper Pro CPU, which I’ve always wanted an excuse to use (Laughs), and two Nvidia 6000 ADA series GPUs. In terms of upgrades, we’re still running vanilla Unreal Engine and no real media server, so right now we’re looking into a Pixera, Disguise or Assimilate Live FX solution to help streamline and optimise some things.

DP: How do you handle colour management between Unreal, the LED wall, and camera output?

Sam Culbreth: This can be tricky, as we’re diving through multiple layers of colour between the wall, Unreal Engine, and the camera. Unreal Engine has built-in colour management tools for features like ACES. And we try to keep our wall settings consistent for each shoot.
But in the end, what really matters is what we’re seeing through that final camera output. Having that output visible in real time gives us plenty of opportunities to make adjustments before we film.

DP: When Jenna brings you a new project, what do you look for in the Unreal scene setup that a customer delivers?

Sam Culbreth: It all depends! I’m always happy to get references up front so I can start strategising right away. If there are preexisting 3D files, I can often convert them over to use in our scenes. As a general rule for textures, you want powers of twos. For any close or large textures, I like to stick to 4k. You can still get away with lower most of the time, especially if it’s far in the background. I’ll even go down to 256 if it’s a smaller, less detailed object.

DP: How do you document or archive the technical setup for each project?

Sam Culbreth: We have a standard template that we adapt with each new version of Unreal Engine. That way, I’m able to migrate any levels that have been built and keep all of our project settings consistent. I also tend to make a copy of the project before a shoot, in case anything changes dramatically while we’re on set. Once it’s over, we have a process for archiving old projects securely on-site in case we ever need to pull them back up.

DP: If you could write a wish list for your customers, what would it include?

Sam Culbreth: Come in with a solid plan and vision long before we’re on set.  We have some wiggle room to make changes, but your visuals should really be locked in before we’re shooting. If you’re uncertain about something, we can prepare a few options, but knowing that ahead of time gives us a chance to make preparations to adapt that won’t require on-set scrambling or potential crashes. Fewer changes on set means better stability and more precious time getting content.

DP: How do you coordinate real-time updates between Unreal, camera tracking, and lighting?

Sam Culbreth: That’s the neat part, you don’t! Once our motion tracking system is connected and the project is launched, the camera tracking handles itself. Most updates I make to the scene are typically immediately visible on the wall. Though occasionally I’ll have to save the scene to push the changes. As for latency, we test our scenes ahead of time to make sure they’re meeting our goals. We have a process for dividing up and disabling parts of our map when we need to improve our performance. 

DP:  What’s your favourite Alignment / test pattern?

Sam Culbreth: True Neutral, or Chaotic Good. – Oh test pattern, I don’t have one. But I tend to gravitate toward a synthwave aesthetic for screensavers.

DP: Gaussian splats are picking up momentum. Can you simply load one into your scene?

Sam Culbreth: We’ve experimented with a few different scan methods, and there is a bit of a process required to make it possible without totally overloading the system. Gaussian splatting is a very powerful technique that I’m hoping we can make more use of in the future, but it hasn’t been necessary for our builds. 

DP: What tools or plugins have become indispensable in your workflow?

Sam Culbreth: Of course, we have the necessary Virtual Production Tools like nDisplay, LiveLink, and Mo-sys. But I’ve really enjoyed using PolygonFlow’s Dash plugin and NVIDIA’s DLSS plugin. These tools really speed up my workflow when it comes to building environments, performance, and rendering. 

DP: And finally, the question every reader is thinking: can you game on the Door G stage after hours?

Sam Culbreth: I certainly hope to! We’ve tested a few things on it before, and it can absolutely serve as a massive monitor for maximum immersion. Furthermore, we’ve got plans to utilise our motion-tracking systems and VR environments to create some interesting options for virtual scouting and other experiences. 

The post Virtual Production Without the Disney Budget first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

For those who don’t know the tool: Assimilate is best known for Scratch, its colour grading and finishing system used in professional post pipelines for more than twenty years. Building on that experience, the company developed Live FX, a software-only real-time media server for virtual production. Live FX sits alongside Live Looks for live grading and Live Assist for multicam VTR and Play Pro Studio as the only professional Video Player you’ll ever need. Together, these tools cover on-set backup & offload, colour, capture, playback, and now LED stage control, completing a loop from wall to camera to grade.

Why Live FX Was Created

When virtual production started creeping into film workflows, nobody had actually built a media server for film sets. Crews borrowed tools from the live-event world or bent game engines like Unreal into shape. It worked, technically. It also meant a lot of compromises, hacks, and explaining to cinematographers why things behaved like stage tech.

Live FX was explicitly built to stop that nonsense. It is a software-only system designed for film production, not for programmers or touring engineers. Live compositing, projection mapping, and image-based lighting all live in one interface that makes sense to people who think in shots and lighting setups. There are no artificial limits on video outputs or DMX universes. It stays lean, runs fast, uses hardware efficiently, and scales cleanly from a single machine to multi-node setups. Playback, lighting, and metadata all reside in the same environment rather than being duct-taped together.

Take a driving plate shoot for a car commercial. Live FX can drive 360-degree plates across a curved LED volume, control real lighting fixtures based on what is happening in the image, and stay locked to camera tracking and lens metadata. As the virtual sun in the plate moves or shifts colour, the physical lights on set follow along in real time. The result is a much closer-to-final image before anyone even opens a compositing tool.

The best part is that Live FX does not demand a rack of servers or exotic hardware. With small or midsize LED volumes, it runs comfortably on a normal workstation. Think MacBook, think gaming PC. That makes LED workflows accessible to smaller studios without turning the hardware budget into a horror story.

Completing the Assimilate Ecosystem

Live FX also neatly finishes what Assimilate has been building for the past two decades. The company already knows a thing or two about media pipelines, and while Scratch covers real-time colour work and solid format playback. Live Looks brings grading onto the set. Live Assist handles multicam VTR duties. All familiar territory.

With Live FX in the mix, the ecosystem finally extends into projection mapping, camera tracking integration, and lighting control. In other words, the bits you need for LED volumes and still very much need for classic green screen stages. Crucially, Live FX keeps metadata intact. Lens data, focus, exposure, tracking information, scene, take, timecode – all of it can be captured on set and passed straight down the line into postproduction. No extra logging, no manual copy-paste rituals, and far fewer opportunities to lose something important between shoot and post.

Why It Is Different from Unreal

Live FX is fundamentally different from running an instance of Unreal Engine. Unreal is a full game engine that requires significant system resources, whereas Live FX is designed as a creative tool for cinematographers, DITs, DPs and VFX artists. It emphasises rapid real-time operation with a user interface tailored to film workflows rather than code.

Live FX supports workflows such as 360° plates and 2.5D compositing with ease. It can also import 3D environments like Notch Blocks or USD scenes. Notch Blocks are prebuilt, real-time graphics or effects authored in the Notch motion graphics tool and used here to add dynamic or procedural visual elements directly to LED walls without taxing the system as heavily as a full engine build. (If you have a look around the Notch site, you’ll spot Live FX in many of their demonstrations)

At its core, Live FX is one machine that can drive LED walls, control lights, composite live plates, and manage colour, all in a single, unified software environment.

The Software Side

The Live FX interface is built on immediate visual feedback. It is a GPU-powered environment that can load, layer, grade, and project footage onto LED walls, green screens, or 360° domes. It supports the projection of 2D, 2.5D, 180° and 360° content beyond 16K resolution in a variety of modes, including frustum, planar, spherical, dome or cylindrical output.

For 3D content, Live FX can import USD scenes and Notch Blocks. It also supports Netflix OpenVPCal, an open-source LED wall calibration tool developed for in-camera visual effects workflows. OpenVPCal creates colour space corrections tailored to the camera and LED pipeline to improve image accuracy across walls.

Live FX does not limit scene size or plate resolution; the hardware instead determines those constraints. Modern GPUs commonly support textures up to 32K by 32K, and Live FX can theoretically handle playback up to 65K by 65K at over 1,000 frames per second. However, real-world limits are set by machine capability.

Compositing

Real-time compositing includes built-in qualifiers:Live FX ships with RGB, Chroma, Luma, HSV and Vector keyers that can be combined in any way thinkable. It supports OFX plug-ins and Matchbox Shaders, the latter being small GPU-based shading programs developed initially for Autodesk Flame, adapted here for custom, real-time visual effects.

Inline colour grading tools include wheels, curves, masks, and LUT management, enabling DITs and Creative Artists to see accurate grades on set. Camera and metadata integration allows Live FX to read detailed information, such as lens, focus, and exposure, from SDI feeds from manufacturers including ARRI, Sony, Canon, and Panasonic. It supports all major camera trakcing systems, such as Mo-Sys, Zeiss, Vicon, OptiTrack and others, but also generic protocols like FreeD, Open Track IO or Open Sound Control protocols.

Inside Igelkott Studios: Pioneering Projection-Based ICVFX with Live FX

Playback in Live FX supports nearly any frame rate or aspect ratio the GPU can manage. It can record raw feeds and metadata while outputting a live composite and automatically rebuild the final online version after the shoot. Storage requirements vary by format; high-resolution codecs such as NotchLC or OpenEXR sequences demand high-speed, high-capacity storage, while lighter formats accommodate portable SSDs. Live FX can also capture and output via NDI for streaming to platforms such as YouTube through tools like OBS.

Live FX does not natively run interactive simulations or complex particle systems, but it integrates with engines such as Unreal, Blender or Unity. In these hybrid workflows, Live FX passes timecode and tracking data to the engine and receives the rendered output back, making it suitable as a composer in mixed real-time pipelines.

Scaling the Stage

A single Live FX Studio system can control up to seven LED processors from two GPUs, or up to eight LED processors through two 8K-capable SDI cards – provided no other hardware bottleneck limits the playback. For larger volumes, a multi-node configuration becomes necessary. In such setups, multiple Live FX systems split the rendering and control responsibilities across the volume, with each system driving part of the wall.

Multi-node workflows also help clarify roles on set. Typically the guy taking care of the lights, is a different guy from the one taking care of the LED volume and you don’t want the two to fight over who gets to use mouse and keyboard. So setting up a separate client Live FX system to take care of lighting might be a good choice. 

Hardware and Platform Requirements

One of Live FX’s core strengths is that it does not require datacentre level hardware. A modern workstation or laptop with a contemporary GPU is sufficient for smaller and medium setups. For larger stages, a Windows workstation with a high-end Nvidia Quadro card is advisable. You can run Live FX on really any machine, as long as the operating system is Windows or macOS. You can even reuse existing mediaserver hardware. The main question is, whether that hardware is up to the task its supposed to fulfill – the software does not impose any limitations.

CPU requirements are modest; any processor from the last five years that handles video rendering comfortably will work. Minimum RAM is 8GB, recommendations start at 32 GB with 64 GB or more for heavier workloads. If Unreal runs on the same machine, memory usage scales with that workload. Supported operating systems are Windows 10 or later and macOS 10.15 or later. Optional SDI I/O can be added via cards from AJA, Blackmagic Design or Bluefish444.

Existing media server hardware can often be repurposed for Live FX if it meets these requirements. The software itself does not impose restrictions; performance depends on hardware suitability. Even a compact Mac Mini can handle basic LED playback for smaller LED stages .

Connectivity: Walls and Lights

Live FX interfaces directly with LED processors, such as those from Brompton and Megapixel VR, to manage genlock, timecode, and perspective correction. It outputs Art-Net, sACN or USB DMX to lighting fixtures and is the only tool right now that supports video-based fixtures like Kino Flo Mimiks, wihch do not use DMX. (A short glossary of those formats at the bottom)

In image-based lighting mode, each fixture can respond to what is playing on the screen – or content that is supposed to appear off the screen in front, above, or next to the talent on stage (think 360° content, where only a portion of the entire 360 degrees is actually shown on the LED wall and the other content can be used for image-based lighting to simulate the environment!). If the plate shows an intense directional light from the left, the physical lights respond accordingly, making stage illumination react to LED content in real time.

Live FX integrates with lighting consoles, enabling gaffers to control physical fixtures or their virtual equivalents. Because Live FX handles pixel mapping internally, lighting desks only need to control up to 51 DMX channels per fixture, regardless of how many channels the physical fixture uses. This helps reduce network complexity and preserve universes on the lighting console.

A webinterface?

A built-in web interface allows remote control of fixtures from a tablet or phone, which is convenient for on-set adjustments. Using real stage lighting to complement LED walls provides brightness and full-spectrum illumination that LED panels alone often cannot deliver. Many LED walls remain based on RGB emitters that exhibit spectral gaps, particularly problematic for accurate skin tones. Stage lights with full-spectrum sources help close those gaps, producing more realistic imagery. Although newer LED tiles that incorporate warm and cool white emitters improve colour response, adoption in studios remains limited, due to cost.

Setup and Onboarding

It is claimed that a Live FX setup can be completed before your coffee gets cold. Installation, licensing and launch are straightforward. Connect the LED processors and lighting interfaces, load the background plates or environments, import LUTs or colour profiles such as ACES, Rec.2020 or HDR10, and begin playback. Lighting synchronises automatically.

Learning resources include the Assimilate YouTube channel,
demo scenes for green screen and LED wall workflows from sim-plates.com,
the Live FX User Guide,
and the Assimilate Discord server.

To learn Live FX, Assimilate offers dedicated training, which should be booked for at least 2 days. Alternatively, their free Assimilate Academy Masterclass might come to a town near you. But even without training, Live FX can be learnt fast and efficiently via the YouTube channel. Tutorials and dedicated playlist for almost any subject exist and allow new users to get up & running really quickly. While it is preferred that users get acquainted with the software beforehand, the reality is that Live FX is often deployed a day before the shoot, but even in those scenarios, new users can get a usable image onto the wall within minutes after installing the software.

But does it work?

How Live FX Is Changing The Face Of Virtual Production

Although a very young tool, Live FX has been used on major Netflix shows and feature films. Unfortunately, a lot of the details cannot yet be disclosed, but here is a very select list of the latest case studies:

• LiveFX on “Secrets we Keep” https://digitalmedianet.com/behind-the-scenes-of-the-netflix-hit-secrets-we-keep-a-virtual-production-deep-dive/
• Live FX for “Biffy Clyro” https://digitalmedianet.com/interview-with-russ-shaw-daniel-walker-on-the-making-of-biffy-clyros-goodbye-video/
• Live FX on “Ballad for a Small Player”: https://digitalmedianet.com/behind-the-scenes-visual-effects-for-ballad-of-a-small-player/

What’s new in the latest release and what’s next?

Just before Christmas, Assimilate released Live FX 9.9, which introduced true multi-node workflows, along with multi-GPU-output options, an updated Stage Manager, new Unreal nDisplay workflow, and various smaller updates to the IBL toolset, projection modes and format support. But Assimilate is already working on the next batch of features: REST API and native ST2110 support are coming up later this year.

Live FX 9.9: More Walls, More Nodes, More Fun

Compatibility, Technical Specs and Pricing

Live FX supports Windows 10 or newer and macOS 10.15 or newer. It scales from laptops to full LED control rigs. Typical configurations can hold multiple 4K feeds or a single 16K background plate. Supported file types include EXR, ProRes, NotchLC, HAP, RAW, H.264, H.265, LUTs (.cube and .3dl), CDL, Notch Blocks and USD. External I/O includes SDI, NDI, Spout, Syphon, OSC, DMX, Art-Net and sACN.

Camera metadata and colour data flow directly through the Assimilate ecosystem. Metadata including user annotations, scene and take information and all camera tracking and SDI-metadata can be recorded and passed along to postproduction. Integration with Unreal, Blender and Unity workflows can happen via GPU texture sharing or NDI/SDI. Licensing is available in multiple options: monthly ($695), 3-month ($1,595), annual ($4,995) and permanent ($7,495, including one year of support and updates). Annual support for permanent licences can be renewed at any time. And before everything, get the free trial / learning edition here.

Why It Matters

Live FX functions as the set’s nervous system, coordinating LED walls, lighting, compositing and colour with post-grade precision and real-time responsiveness. It brings film production language and workflows onto the stage, reducing guesswork and providing artistic control at the source. After twenty years of Scratch being called the Swiss army knife of postproduction, Live FX may take that title for virtual production.

As always, professionals should test thoroughly before deploying on live shoots, and you can do this easily with the free personal learning edition here. Register and drop us a line that Digital Production sent you. Assimilate will set you up :).

Appendix: Protocols, Extensions, Standards and Definitions

FreeD A tracking data protocol developed for camera tracking systems to transmit pose, lens and other metadata in real time.

Open Track IO An open and free metadata protocol designed by the SMPTE RiS-OSVP group for virtual production tracking interoperability. ris-pub.smpte.org

Open Sound Control (OSC) A network protocol for communication among computers and multimedia devices, widely used in creative and show control environments. Wikipedia

Art-Net An Ethernet protocol specification for transporting DMX universes over UDP/IP, used for lighting control in entertainment networks. Beckhoff Automation

sACN (Streaming Architecture for Control Networks) – A protocol (ANSI E1.31) for sending DMX data over UDP/IP networks, designed for multicasting lighting control. Beckhoff Automation

OpenVPCal An open-source calibration tool for LED wall and camera pipelines developed by Netflix, used to generate colour space corrections. GitHub

Notch Blocks Prebuilt, real-time content building blocks created in the Notch motion graphics tool for procedural visuals. Notch Blocks

Matchbox Shaders Small, GPU-based shader programs originally developed for real-time effects in Autodesk Flame, used here for Live FX real-time effects.

USD (Universal Scene Description) A file format and interchange framework for 3D scene descriptions, widely adopted for complex pipeline interoperability. But as Digital Production readers you should know about all this. Why are you reading this? Get the LiveFX-Trial!

RTMP (Real Time Messaging Protocol) A protocol for streaming audio, video and data over the internet.

P.S.: Yes, the headline is a wink at one of the funniest moments in Mel Brooks’ High Anxiety. We’ll just leave this here.

The post Assimilate Live FX: The VP-Stages’ very, VERY nervous system. first appeared on DIGITAL PRODUCTION and was written by Mazze Aderhold.

The Mac mini has become a surprisingly capable little workhorse. With Apple’s M-and M-Pro chips, it easily handles cutting, grading, and playback duties once reserved for towers. But anyone who has actually used one in a studio knows what follows: a nest of cables snaking from Thunderbolt hubs, audio interfaces, SSDs, and PCIe expansion boxes. For facilities that rackmount their systems in machine rooms, the mini’s minimalist form factor quickly becomes a liability.

Enter Sonnet Technologies with its xMac mini (2024+) family: modular 2U rackmount enclosures designed to house one or two Mac minis while adding full Thunderbolt-to-PCIe expansion. The result: less cable chaos, better airflow, and no more artists tripping over power bricks on the way to the coffee machine. And a power button you can reach without shoving your hand into the gordian knot of cables behind/below the machine. Yes, it is nicely at the front.

Modular Rackmount for Modular Macs

The xMac mini (2024+) comes in several configurations, all based on a half-width 2U chassis. The core xMac mini module mounts up to two Mac minis side by side, providing front-panel power buttons and access to all I/O ports.

Studios starting from scratch can opt for the xMac mini (2024+) Base System, which includes the module plus an empty half-width bay ready for a PCIe expansion of choice. For heavier workflows, Sonnet bundles two preconfigured systems with its Echo Thunderbolt-to-PCIe technology.

PCIe Expansion via Thunderbolt

The xMac mini (2024+)/Echo II DV pairs the Mac mini module with a half-width Echo II DV expansion chassis. It provides two full-height, full-depth PCIe slots, each connected via a dedicated Thunderbolt link. This layout allows two separate Thunderbolt paths from one Mac mini (or from two individual minis) ideal for cards with sustained high bandwidth such as AJA or Blackmagic capture, 25GbE network, or RED Rocket acceleration.

The higher-tier xMac mini (2024+)/Echo III module offers three PCIe slots through a single Thunderbolt port, targeting audio professionals running multiple Avid Pro Tools | HDX cards. Both Echo modules rely on quiet Noctua fans, which keeps noise low enough for control-room placement if required.

Quiet, Compact, and Very Specific

Sonnet’s enclosure design stays true to the company’s utilitarian/professional aesthetic: steel construction, quiet thermals, no unnecessary LEDs. The half-width modules allow two full systems to occupy one 2U space, which makes sense for audio and broadcast racks where space is at a premium. The modular approach also allows users to mix and match components, for example, a single Mac mini alongside a PCIe expansion, or two minis for redundancy.

Admittedly, it’s a niche solution. Few studios are clamouring to rackmount Mac minis, but those who rely on macOS for editing or ingest tasks and want their systems housed neatly in a server room will find the xMac mini a practical, production-friendly setup.

And if your goal is to have a color monitoring workstation: We got you covered!

Color Monitoring for Cheapskates?Part 1

Pricing and Availability

All xMac mini (2024+) systems are available now from Sonnet Technologies and its authorised resellers.

• xMac mini (2024+) Module (XMM-MOD): $479.99
• xMac mini (2024+) Base System (XMM-SYS-0): $699.99
• xMac mini (2024+)/Echo II DV (XMM-SYS-IIDV): $1,799.99
• xMac mini (2024+)/Echo III (XMM-SYS-III): $1,799.99

All variants are compatible with Thunderbolt-equipped Mac minis, including the current M4 and M4 Pro generation.

For the Cable-Averse

In essence, Sonnet’s xMac mini (2024+) family provides what many studios quietly wish Apple would: a professional, rackmount-ready version of the Mac mini with PCIe flexibility. It’s not for everyone, and it’s certainly a specialised island solution but for facilities that want to stay within macOS without inviting the cable octopus, Sonnet’s design delivers a tidy, scalable path.

As always, professionals should test the system thoroughly in their workflow before deployment in production environments.

The post Rack ’n’ Mini: Sonnet Turns Mac mini into Studio Hardware first appeared on DIGITAL PRODUCTION and was written by Uli Plank.

Laetoli Production, together with illustrator and animator Samba Soussoko, has launched a free anatomy-reference app focused entirely on the arthropod kingdom. The Arthropods Anatomy Reference App (sometimes called Arthropodes) offers 3D scans of real insects, spiders, crabs and other exoskeletal oddities collected from European museum archives. The idea is simple but delightfully unsettling: instead of staring at flat reference photos of beetles and centipedes, artists can rotate, measure and dissect these tiny nightmares directly in 3D.

Small zoo, big ambitions

At launch, the app features twelve scanned specimens, an admittedly modest start, but one that should already cover the basics for anyone designing mandibles, antennae or far too many legs. More scans are promised “in the coming months”. Each model can be freely explored in a browser or, for those who like their bugs a bit more permanent, via a free Windows-only desktop version.

The app lets users hide or isolate specific body parts, allowing a clean view of wings, carapaces or leg joints without the clutter. You can even measure distances or angles between parts, handy for anyone trying to make sure their digital spider isn’t accidentally violating the laws of insect biomechanics. For the detail-obsessed, two species can be displayed side by side for direct comparison, say, a crab next to a centipede, if you need inspiration for your next fantasy hybrid.

Why anyone sane would want this

Creature designers, modellers and riggers often need grounded anatomical data, even for fantasy work. Having museum-accurate arthropods at your fingertips saves endless time hunting for blurry reference photos online. It also ensures that even the most monstrous alien bug in your next VFX shot has credible proportions. Still, one look at some of the included specimens makes it clear why we chose the tamest beetle possible for the image. Some of these creatures are perfect for Halloween mood-boarding and possibly for testing one’s courage.

A tool to bug your pipeline (gently)

Currently, the app runs in a web browser or via a Windows executable, with no macOS or Linux versions announced. There’s no detailed data on scan resolution, polygon counts or export options, so artists should test it carefully before integrating it into a professional workflow. Despite its early-release limitations, it’s a solid niche reference tool for designers who prefer to measure a millipede rather than invent one from scratch.

For now, the Arthropods app feels like a digital cabinet of curiosities: part science, part art, and part “please keep that thing away from me”. It’s free, it’s surprisingly informative, and it might even help your creature rigs look a little more biologically plausible. If your next project involves creepy-crawlies, carapaces or too many eyes, this app is worth a spin, literally. Just maybe don’t open it right before bed.

The post Creepy-Crawlies in 3D: Free “Arthropods” Anatomy App for Creature Designers first appeared on DIGITAL PRODUCTION and was written by Bela Beier.

We know Scratch, Scratch VR, Scratch Web and now we can also remember Scratch Play Pro. The family is getting bigger.

And tomorrow, world domination…

One Player to Rule Them All – That’s what it says on the ASSIMILATE website, which mainly refers to the universal format player that can swallow 16 different camera-specific formats without falling to its knees (hehe). But the other features are also impressive.
Let’s make it easy for ourselves and put them in the table:

• Universal Player: Play Pro can read almost any format – camera format, RAW or not, even h.264 and h.265 (HEVC), even frame formats.
• Go-to immersive player: 360…360° everywhere, 180° VR, stereo, mono, side-by-side or over/under, embedded ambisonic audio and real-time mesh de-warping of 180° VR media. Playback works on the screen or directly via headsets such as the Oculus Rift or HTC Vive. The whole thing can then also publish your 360/180 degree experiences on Facebook 360 or YouTube 360.
• Look-Creator: Play Pro supports 1D and 3D LUT formats in any size for import and export, as well as the import/export of CDLs in both common formats: .cdl and .cc.
• Media Transcoder: Each camera dealer transcodes in their own format… Play Pro reads almost every common format and can transcode it to ProRes, h.264, and h.265. For ProRes, this is even certified by Aplple – to our knowledge the first not to use ffmpeg. For VR delivery, it supports h.264 rendering up to 8K, including the required metadata for various online portals such as Facebook or YouTube. Custom metadata can also be added.
• QC tool: Quality check for post-production and directly on set. It supports SDI output, split screen, A-B overlay and audio monitoring & routing.
• Price: the trend is towards subscriptions: the Pro scratcher is available for 19 US dollars per month (199 per year).

You can find more detailed and further information HERE.

The post Scratch… Scratch Play Pro first appeared on DIGITAL PRODUCTION and was written by Alexandra Kaschny.