Understanding the Three-Point System (TPS)

A practical introduction to camera geometry for stop-motion animation

How do you put a camera back exactly where it was?

It’s a deceptively difficult problem. Even small changes in camera position, orientation, or focal length can make two otherwise identical shots look different.

The Three-Point System (TPS) is a measurement-based technique for documenting and recovering camera setups.

Developed from practical stop-motion and visual effects production workflows, TPS uses three identifiable points in a scene to establish a geometric reference between the camera and the objects being photographed.

The goal is simple: Make camera positions measurable, repeatable, and easier to recover.

1. The Problem: Recovering a Camera Position

Imagine you’re animating a shot when something happens:

1. Your camera gets bumped.
2. You need to remove the camera to work on the set.
3. A shot must be recreated weeks later.
4. You’re rebuilding a set for additional photography.
5. You need to match a reference photograph.

You might have the original image, camera settings, and even measurements of the camera’s physical position.

But reproducing the original framing can still be difficult.

A camera has six degrees of freedom:

Translation

1. Left and right
2. Up and down
3. Forward and backward

Rotation

1. Pan
2. Tilt
3. Roll

Focal length and focus settings introduce additional variables that can affect the resulting image.

TPS helps organize the recovery process around measurable reference geometry rather than relying entirely on visual judgment.

2. The Sacred Triangle

At the heart of TPS is a simple idea:

Three identifiable points form a triangle.

We call these points:

1. A — First reference point
2. B — Second reference point
3. C — Third reference point

Together, they establish what we sometimes call the Sacred Triangle.

The points should be stationary, clearly identifiable, and visible in the reference image.

They can be physical markers, identifiable features of a set, or other stable reference locations.

Why Three Points?

A single point provides a position in the image but very little information about the camera’s orientation.

Two points establish a line, allowing you to compare position, direction, and apparent scale.

Three non-collinear points establish a triangle.

That triangle provides additional geometric information about the relationship between the camera and the reference scene.

Changes in camera position, orientation, or focal length can change the triangle’s appearance in the image.

By comparing the original triangle with the current camera view, you gain useful information for aligning the camera.

Important: Three image points alone do not uniquely determine every camera parameter. TPS combines image observations with physical measurements and camera information to constrain the problem.

3. Measuring the Triangle

The three reference points define three sides:

AB — Distance between A and B

BC — Distance between B and C

CA — Distance between C and A

These distances describe the physical dimensions of the reference triangle.

Unlike the apparent distances between points in a photograph, physical measurements remain the same when the camera moves.

For example, suppose your reference triangle has these dimensions:

1. AB = 300 mm
2. BC = 400 mm
3. CA = 500 mm

These measurements describe a valid right triangle.

The values are illustrative.

Your actual reference triangle does not need to be a right triangle.

Why Physical Measurements Matter

A photograph contains a two-dimensional projection of a three-dimensional scene.

Without additional information, apparent distances in an image do not directly reveal physical distances.

Recording the triangle’s actual dimensions provides a scale reference for the camera geometry calculations.

Accurate measurements are therefore an essential part of the TPS workflow.

4. The A-Side: Your Original Camera Setup

TPS distinguishes between the original camera setup and the setup you’re attempting to reproduce.

The original reference is called the A-Side.

Think of the A-Side as the camera position you want to preserve.

It contains the information needed to describe your original setup, including:

1. The reference image.
2. The image positions of points A, B, and C.
3. The physical measurements of the reference triangle.
4. Camera and lens information.
5. Relevant camera orientation and focus information.
6. Calculated geometric data.

LIMNTPS helps collect and organize this information into a portable camera setup record.

Why Save the A-Side?

Because you may not know when you’ll need it again.

A camera setup that seems permanent today might need to be recovered tomorrow.

Documenting the A-Side before anything moves provides a reference for future alignment.

5. The B-Side: Recovering the Camera

The B-Side represents the camera’s current position during a recovery or alignment process.

The objective is to make the B-Side image match the original A-Side reference.

This involves comparing the projected positions of the same three reference points.

The Alignment Process

A practical alignment workflow generally involves:

1. Restore the original reference triangle in the scene.
2. Position the camera approximately where it was originally.
3. Restore the original camera and lens settings.
4. Match the camera’s tilt and approximate orientation.
5. Compare the B-Side reference points with the A-Side.
6. Adjust camera position and orientation.
7. Repeat until the reference points align.
8. Verify the overall image composition.

When the same physical points project to the same image positions under matching camera conditions, you have strong evidence that the camera alignment is approaching the original setup.

However, matching three points is not an unconditional guarantee that every part of the image will match perfectly.

Lens distortion, calibration differences, measurement errors, and scene changes can still affect the result.

6. Understanding Camera Projection

A camera converts points in three-dimensional space into positions on a two-dimensional image.

This process is called projection.

The appearance of your reference triangle depends on several factors.

Camera Position

Moving the camera changes the apparent positions of the reference points.

Moving closer or farther away can also change their apparent separation.

Camera Orientation

Pan, tilt, and roll change where the reference points appear in the image.

Even a small angular adjustment can produce a noticeable difference.

Focal Length

Focal length affects the camera’s field of view.

Changing focal length changes the apparent size and arrangement of the reference triangle within the image.

Sensor Dimensions

The camera sensor’s dimensions help establish the relationship between focal length and field of view.

This is why accurate camera and sensor information matters.

Focus Distance

Focus settings can affect the effective optical behavior of some lenses.

LIMNTPS records focus-distance information as part of its camera geometry workflow.

7. Why Camera Tilt Matters

Camera tilt is an important part of the TPS calculation.

The reference triangle exists in physical space, but the camera observes it through a particular orientation.

Changing the camera’s tilt changes how that triangle projects onto the image.

Without sufficient orientation information, different camera configurations may produce similar projected geometry.

LIMNTPS therefore includes camera tilt as a required part of its setup information.

Camera roll may also be recorded when applicable.

The goal is not simply to match the size of the triangle. It is to reproduce its projected geometry under the correct camera conditions.

8. Choosing Good Reference Points

The quality of your reference points can significantly affect the usefulness of your TPS measurements.

Choose Stable Points

Reference points must remain stationary relative to one another.

Avoid anything likely to move during animation or set maintenance.

Use Clearly Identifiable Features

Choose points that can be located precisely in both the original and recovery images.

Ambiguous edges or soft features make accurate alignment more difficult.

Spread the Points Apart

A triangle covering a substantial portion of the image generally provides better visual leverage than three tightly clustered points.

Avoid Nearly Collinear Points

Three points arranged almost along a straight line form a poorly conditioned triangle.

A triangle with meaningful area is preferable.

Preserve the Reference Geometry

If the physical reference points move relative to one another, the original measurements may no longer describe the scene accurately.

Whenever possible, document and protect your reference points throughout production.

9. What LIMNTPS Calculates

LIMNTPS brings the Three-Point System into a practical mobile workflow.

The application combines:

1. Reference point locations in the image.
2. Physical triangle measurements.
3. Camera and sensor information.
4. Focus distance.
5. Camera orientation information.

These inputs support calculations used to document camera geometry and produce alignment references.

LIMNTPS can generate portable TPS data, visual overlays, and camera reports.

These outputs help preserve the original setup and support later recovery.

LIMNTPS does not automatically move or align the camera.

It provides the geometric information needed to make manual alignment more systematic.

10. Using TPS with Dragonframe

Dragonframe provides tools for viewing and aligning camera images during stop-motion production.

LIMNTPS can export reference overlays for use in compatible Dragonframe workflows.

A typical process involves:

1. Documenting the original setup in LIMNTPS.
2. Exporting the A-Side reference overlay.
3. Loading the overlay into the camera alignment workflow.
4. Comparing the live camera image with the reference.
5. Adjusting the camera until the reference points align.

The combination of physical measurements and visual overlays can make camera recovery more structured and repeatable.

LIMNTPS is an independent LIMNMEDIA application and is not affiliated with or endorsed by Dragonframe.

11. Limitations and Accuracy

TPS is a measurement-based technique, and its results depend on the quality of the information provided.

Potential sources of error include:

1. Inaccurate physical measurements.
2. Incorrect reference point placement.
3. Changes to the physical reference triangle.
4. Incorrect sensor dimensions.
5. Lens distortion.
6. Differences in focal length or focus settings.
7. Camera orientation errors.
8. Changes to the scene between setups.

TPS does not eliminate these factors.

Instead, it provides a structured way to document camera information and reduce uncertainty during alignment.

For critical production work, always verify the recovered camera position against the original image before resuming animation.

12. The Bigger Idea: Repeatable Camera Work

The Three-Point System is about more than recovering a bumped camera.

It is a way to make camera setups easier to document, communicate, and reproduce.

A properly documented setup can be useful when:

1. Returning to an interrupted shot.
2. Rebuilding a set.
3. Planning additional photography.
4. Sharing camera information with another artist.
5. Teaching camera geometry.
6. Integrating stop-motion photography with visual effects.

The underlying principle is simple:

If you can measure and document a camera setup, you have a much better foundation for reproducing it.

That’s the idea behind LIMNTPS.

Learn More

Next article: Choosing Reference Points for LIMNTPS

For additional assistance, visit the  ⁠LIMNMEDIA Support Portal.

Email: [email protected]

Community:  ⁠LIMNMEDIA Discord

Developed by LIMNMEDIA

Less guesswork. More animation.