ProductivityFree

fSpy

Free open-source camera matching tool for 3D artists. Load a photograph, set vanishing points, and export a perfectly matched camera to Blender for environment modeling, matte painting integration, or architectural visualization. Essential for recreating real-world perspectives in 3D scenes and matching concept art camera angles for game environment blockouts.

Open Website Official Docs

Best For

Matching a Blender camera precisely to a reference photograph using vanishing points, enabling accurate concept-to-3D blocking and matte painting projection setup for game environments.

Primary Use Cases

- Matching a Blender camera to a concept art piece to block out accurate 3D geometry that matches the intended perspective for an environment art portfolio shot
- Setting up a camera-matched matte painting scene in Blender where 3D game assets must align with a photographic background plate
- Establishing correct building proportions when modeling architecture from a photographic reference with an unknown camera lens
- Creating a camera-matched scene for a game environment screenshot composition where the hero prop must sit convincingly within a photographic context
- Teaching perspective theory by demonstrating how focal length, sensor size, and camera height interact using a real photograph as ground truth

Pros & Cons for Game Art

Pros

+ Free and open-source — delivers camera matching capability that costs thousands of dollars in professional matchmoving software for the specific use case of still-image matching
+ Simple, focused UI with only the controls needed for still-photo camera matching — productive in under 30 minutes for an artist who understands basic perspective
+ Direct Blender integration via addon makes the camera import a single-click operation with no manual matrix entry
+ Accurate enough for game art portfolio compositions and matte painting setups where sub-pixel tracking accuracy is not required
+ Teaches perspective fundamentals concretely — manipulating vanishing points to match a photograph builds stronger spatial reasoning than textbook exercises

Cons

− Works only on still photographs — no video tracking, no moving camera, no lens distortion correction for anamorphic or fisheye footage
− Requires the photograph to contain at least two sets of parallel lines; curved architecture, organic environments, or aerial photography are difficult or impossible to match
− Does not output lens distortion data — cameras with significant barrel distortion require correction in image editing software before fSpy will produce accurate results
− Single-purpose tool: the Blender addon is the only integration point, making it irrelevant in Maya, Houdini, or engine-native camera workflows

Suggested Learning Path

1.Load a photograph with clear parallel lines (building exterior, room interior) and place fSpy's vanishing point handles on two orthogonal sets of parallel edges
2.Set the camera reference height (a known real-world measurement in the photo — a door, a person) to calibrate the scene scale accurately
3.Export the fSpy camera data and import it into Blender using the fSpy Blender addon — verify the camera matrix matches the photograph by toggling camera view
4.Model a simple box representing a building footprint and adjust its scale until it aligns with the photograph's perspective lines from camera view
5.Practice a full matte painting setup: camera-match a landscape photograph, model rough terrain geometry, and project a Substance Painter texture onto the geometry to match the photo's surface detail

Quick Tip

After exporting the fSpy project to Blender, immediately add a reference image plane textured with the matched photograph at the camera's near clip plane — this gives you a persistent visual guide for mesh alignment as you model without switching back to fSpy.

Setup Notes

- Install the official fSpy Blender importer addon from the fSpy GitHub release page — without it, the exported .fspy file cannot be read by Blender directly
- Use photographs with clearly readable vanishing points — wide-angle shots of buildings or interior rooms work best; organic landscapes with no parallel lines cannot be matched
- After importing into Blender, lock the camera object's transform to prevent accidental movement that breaks the perspective alignment
- Set Blender's render resolution to match the original photograph's aspect ratio immediately after import — a mismatch causes the camera sensor model to be incorrect in the viewport

How N-hance Uses It

N-hance uses fSpy in the 3D Art for Modern Games course during the environment layout module, teaching students to camera-match concept art and photographic reference so that modeled environments match the intended perspective and scale before detailed modeling begins.

fSpy

Primary Use Cases

- Matching a Blender camera to a concept art piece to block out accurate 3D geometry that matches the intended perspective for an environment art portfolio shot

- Setting up a camera-matched matte painting scene in Blender where 3D game assets must align with a photographic background plate

- Establishing correct building proportions when modeling architecture from a photographic reference with an unknown camera lens

- Creating a camera-matched scene for a game environment screenshot composition where the hero prop must sit convincingly within a photographic context

- Teaching perspective theory by demonstrating how focal length, sensor size, and camera height interact using a real photograph as ground truth

Pros & Cons for Game Art

Pros

+ Free and open-source — delivers camera matching capability that costs thousands of dollars in professional matchmoving software for the specific use case of still-image matching
+ Simple, focused UI with only the controls needed for still-photo camera matching — productive in under 30 minutes for an artist who understands basic perspective
+ Direct Blender integration via addon makes the camera import a single-click operation with no manual matrix entry
+ Accurate enough for game art portfolio compositions and matte painting setups where sub-pixel tracking accuracy is not required
+ Teaches perspective fundamentals concretely — manipulating vanishing points to match a photograph builds stronger spatial reasoning than textbook exercises

Cons

− Works only on still photographs — no video tracking, no moving camera, no lens distortion correction for anamorphic or fisheye footage
− Requires the photograph to contain at least two sets of parallel lines; curved architecture, organic environments, or aerial photography are difficult or impossible to match
− Does not output lens distortion data — cameras with significant barrel distortion require correction in image editing software before fSpy will produce accurate results
− Single-purpose tool: the Blender addon is the only integration point, making it irrelevant in Maya, Houdini, or engine-native camera workflows

Suggested Learning Path

1.Load a photograph with clear parallel lines (building exterior, room interior) and place fSpy's vanishing point handles on two orthogonal sets of parallel edges

2.Set the camera reference height (a known real-world measurement in the photo — a door, a person) to calibrate the scene scale accurately

3.Export the fSpy camera data and import it into Blender using the fSpy Blender addon — verify the camera matrix matches the photograph by toggling camera view

4.Model a simple box representing a building footprint and adjust its scale until it aligns with the photograph's perspective lines from camera view

5.Practice a full matte painting setup: camera-match a landscape photograph, model rough terrain geometry, and project a Substance Painter texture onto the geometry to match the photo's surface detail

Setup Notes

- Install the official fSpy Blender importer addon from the fSpy GitHub release page — without it, the exported .fspy file cannot be read by Blender directly

- Use photographs with clearly readable vanishing points — wide-angle shots of buildings or interior rooms work best; organic landscapes with no parallel lines cannot be matched

- After importing into Blender, lock the camera object's transform to prevent accidental movement that breaks the perspective alignment

- Set Blender's render resolution to match the original photograph's aspect ratio immediately after import — a mismatch causes the camera sensor model to be incorrect in the viewport