The biggest savings on a VFX-heavy shoot don’t come from cutting shots. They come from capturing extra reference footage that takes minutes on set and prevents days of post-production reverse-engineering. None of these references are exotic — they’re the standard practices that have been written down by VFX supervisors for decades. The challenge is that “standard practice” doesn’t mean “actually done on every shoot.” Many productions skip these references because nobody insisted on them, and then post-production spends time and money trying to recover what could have been captured for free.
This post is a working list of eight things that a VFX-heavy production should capture on set. Each one takes 1–10 minutes of crew time. Each one saves substantial post work. The total cost in production is small; the total savings in post is large.
1. Clean Plates
A clean plate is the same shot as the take, but without the foreground subject (the talent, the prop, the moving element that has to be removed or modified in post). Same camera position, same lens, same lighting, same framing — just without the subject.
Why it matters: clean plates are the easiest reference for any cleanup work. Wire removal, rig removal, talent removal, marker cleanup — all of these are dramatically easier when the post team has a clean version of the background to project against. Without a clean plate, the post team has to reconstruct the background from other frames in the take, or from neighboring shots, or from scratch.
How to capture: at the end of each take that needs background reference (or before the talent enters the scene), let the camera roll for 5–10 seconds with the same framing and no subject. Most shoots find these takes happen naturally during the final reset — the talent is moved into position, the camera is rolling, and there’s a few seconds of empty frame before action. Capturing those seconds deliberately is free.
When to insist: any shot with rig removal, wire removal, or significant cleanup. Any shot with greenscreen if a non-greenscreen background reference would help. Any shot where the talent moves through a complex environment that has to be reconstructed.
2. Chrome and Gray Ball Reference
Chrome and gray balls are spheres held in the scene during a brief setup capture. The chrome ball reflects the surrounding environment — providing reference for what was lighting the scene from every direction. The gray ball provides a neutral reference for color and exposure.
Why it matters: when CG elements have to be lit to match the plate, the chrome ball tells the 3D CGI team what light sources existed in the scene and where they were. The gray ball tells them how those lights affected a neutral, mid-gray surface. Together, they’re the most reliable on-set lighting reference for CG integration.
How to capture: at the start of each lighting setup, hold the chrome and gray balls in the scene for a few seconds while the camera rolls. The position should be where the talent will be, at roughly the same height. A second pass with the balls held at different positions in the scene (closer to camera, further from camera) is worth the additional 30 seconds.
When to insist: any shot with CG elements that need to be lit to match the plate. Any shot with significant beauty work where lighting reference matters. Any shot where the existing lighting will be modified in post and the post team needs to know the original.
3. Lens Distortion Grid
A lens distortion grid is a checkered chart photographed through the lens. The known geometry of the chart, captured through the lens, reveals exactly how the lens distorts the image — barrel distortion, pincushion, vignetting, chromatic aberration.
Why it matters: every lens distorts. CG elements have to be rendered with matching distortion to integrate cleanly. Without a distortion grid, the post team has to estimate distortion from the plate, which is approximate and often wrong on the edges of the frame.
How to capture: shoot the chart at the focal length being used, at the focus distance being used, with the same camera and lens. A 5-second shot is enough. If the lens is a zoom and multiple focal lengths will be used, shoot the grid at each focal length. The crew can do this during a reset or while the next setup is being lit.
When to insist: any shot with CG elements that need to integrate at the edges of the frame. Any shot with significant set extension or sky replacement. Any shot at extreme focal lengths (very wide or very long) where distortion is more pronounced.
4. HDRI Capture (Or Replacements If HDRI Wasn’t Possible)
A spherical HDRI (high dynamic range image) is captured by photographing the scene from a single point with a camera that captures the full range of light directions and intensities. The result is a panoramic image that can be used as the lighting environment for CG renders.
Why it matters: HDRI is the gold standard for image-based lighting. When CG elements are lit using an HDRI captured on set, they’re lit by the actual scene’s lighting environment — and the integration is dramatically more convincing than CG lit by approximations.
How to capture: dedicated HDRI capture rigs (panoramic camera setups, motorized turrets) are the most thorough method. A faster and cheaper alternative is bracket-photographing a chrome ball with a still camera from multiple angles and combining the results. The fastest is using a smartphone’s spherical panorama feature, which is approximate but usable.
When HDRI isn’t possible: chrome and gray ball reference (item 2 above) is the fallback. A still photo of the scene from the camera’s position is a weaker fallback. None of these are as good as HDRI, but each is better than nothing.
When to insist: any shot with a hero CG element that needs to be lit to match the plate. Any shot with significant CG environment work where the lighting has to feel correct. Any shot where the plate’s lighting is unusual or hard to recreate procedurally.
5. Witness Cameras
A witness camera is a separate camera (typically smaller, separate from the main capture) that records the scene from a different angle during the same take. Witness camera footage isn’t used in the final cut — it’s used as reference for the post team to understand the scene’s geometry and the talent’s performance from a second viewpoint.
Why it matters: when CG elements have to interact with talent or props in 3D space, knowing where things are in the scene from multiple angles helps the matchmove team reconstruct the geometry. Witness camera footage also provides reference for talent’s body position, eye line, and gesture timing — useful for rotomation work.
How to capture: any consumer or prosumer camera works. The witness camera should be positioned to see the full scene from a different angle than the main camera. Time-coded sync is ideal but not required — frame-accurate alignment can usually be done in post.
When to insist: any shot with talent interacting with CG elements that don’t exist on set. Any shot with complex talent movement that needs to be reconstructed in 3D. Any stunt or action sequence where the geometry has to be precisely understood in post.
6. Set Survey or Lidar
A set survey is a measurement of the physical environment: the dimensions of rooms, the positions of major elements, the relationships between objects. Lidar (light detection and ranging) is the modern, fast version — a scanner that captures the entire scene as a 3D point cloud in minutes.
Why it matters: when CG elements have to be placed accurately in the scene, the post team needs to know the scene’s geometry with precision. Estimated geometry from the plate is approximate; surveyed or lidared geometry is millimeter-accurate.
How to capture: a tape measure and camera report works for simple cases. Lidar scanners are the standard for any production with significant CG-to-environment integration. Lidar capture takes 15–30 minutes per location for a typical setup.
When to insist: any shot with set extensions that need to align with real geometry. Any shot with CG architecture that has to match real architecture. Any shot where the camera moves through space and CG elements need to maintain accurate position.
7. Tracking Markers (Where They’re Genuinely Needed)
Tracking markers are visible reference points placed in the scene to help post-production reconstruct camera movement. They’re not always needed — see this post on when to use them — but when they are, placing them deliberately on set is dramatically cheaper than reverse-engineering the track in post.
Why it matters: a matchmove on a low-feature plate (greenscreen, smooth surfaces, snow, sand) without markers takes 5–10x as long as the same shot with proper markers. Markers in the right places save hours per shot.
How to capture: physical markers (small stickers, pieces of tape, painted dots) placed in the scene at distributed positions across the frame. The exact placement should be discussed with the VFX supervisor in pre-production, not improvised on the day.
When to insist: greenscreen plates always. Low-feature environments (snow, sand, water, smooth walls). Long camera moves where parallax is the only reference for tracking. Set extensions where geometry has to align precisely.
8. Color Reference Charts
Color reference charts (such as a Macbeth ColorChecker) provide a known reference for color and exposure in the scene’s lighting. Photographed through the lens at the start of each lighting setup, the chart anchors the post team’s color pipeline.
Why it matters: when post has to match colors across multiple shots, multiple cameras, or multiple lenses, the color chart provides the reference. It’s also useful for color-correction work where the original lighting needs to be modified — the chart tells the colorist what neutral looked like under the original lighting.
How to capture: hold the chart in the scene at the talent’s position, lit by the same lights, and photograph for 5 seconds. Repeat at the start of each lighting change.
When to insist: any production with multiple cameras or multiple shooting days. Any production with color treatment work in post. Any production where exact color matching across shots matters.
What This List Does Together
Each of these references is small. Together, they form the on-set foundation that lets post-production work efficiently. Productions that capture all eight tend to have post pipelines that run smoothly. Productions that capture few or none tend to have post pipelines that spend a significant fraction of their time reverse-engineering missing information.
The total time cost on set, for all eight items, is typically 30–60 minutes per location. The post savings, on a VFX-heavy production, are usually a meaningful fraction of the post budget — far larger than the time invested on set.
How FXiation Digitals Asks for Plate Photography
In pre-production we’ll send a checklist of what we’d like captured per shot, calibrated to the work you’re scoping. Some shots need all eight items. Some need three or four. Some need only one or two. The checklist isn’t a template — it’s a per-project document that our VFX supervisor builds against your specific shot list.
We’ll also flag what we can work without if the production schedule doesn’t allow for full capture. Some references are nice-to-have; some are critical. We’ll be specific about which is which.
If you’re planning a VFX-heavy shoot and want a per-shot capture list, send us the breakdown. The list takes us a day to build and consistently saves productions multiples of that day in post-production. The producers who run smoothly are the ones who treated plate photography as a deliberate part of the shoot, not as an afterthought.
Common Questions