Photoscan using Metashape | VFX fundamentals


Photoscan have in recent years become an important ingredient in the VFX toolbox. It can be an inexpensive way to capture a model for retopology or a way to capture a set for easier match move. Nowadays scanning companies such as clear angle also use photogrammetry techniques but in a customized way where they connect a large array of cameras into something we in Visual effects call a photo booth. One-click and all cameras go off at the same time and once processed you get a very detailed version of the subject. Most Digi doubles in movies are captured in this was today. But today’s VFX fundamentals is a bit more modest I only use one camera on a static object. In the studio, I have a Micheline man statue that to my dog’s delight I brought to my garden patio and blasted about 150 images from around the statue. All of this to demonstrate the basic principles of photoscan using agisoft meta shape. 

Capturing for photoscan

There are a few principles when capturing photos for photoscan. First off you don’t want to capture a reflective object as reflections will fool the scan process. It’s usually good to capture in overcast or diffuse light but I have had good success in daylight scenarios as well. When capturing an object you want to be strategic and shoot around the object in a circle and from different heights and vantage points trying to cover all angles from high to low. In my case, as it was a small object in hindsight it could have been beneficial to raise the object up from the ground but the shoot worked out pretty ok.

Let’s take a look at all of this in action


It’s pretty easy to get a decent object from a number of photos and I bet your photo would even produce a decent result so you don’t have to use the best camera and lenses. In this example, I used a commercial program but I will in a later episode come back and use free software as well, Stay tuned for this episode!