To achieve good finger tracking, it’s important that you follow the correct practice, which enables the system to define your fingers with the most accuracy. Gaining this understanding will drastically improve the quality of your animations.
Check out this Move.ai Academy tutorial:
Camera settings and setup
When preparing for a shoot it’s recommended that the capture devices are set up optimally depending on the number of cameras. Click here for more information on capture volumes.
The most effective way to improve finger tracking is by using more capture devices. The optimal number of capture devices for one actor is 4 cameras. Walking closer to the front 2 cameras whilst still remaining in the frame on all 4 cameras will provide the system with more data.
By setting the cameras up in this layout, the system can utilise multiple angles to reinforce its prediction of where the hands and fingers move within the volume. Positioning the actor closer to the cameras means that the fingers have a larger pixel count which will make it easier for the system to track your fingers with increased precision.
Calibration
The calibration quality is essential when it comes to good finger tracking. This stage in the workflow holds such importance because, without a successful calibration, the system is unable to create animation data. The calibration is mapping the camera positions to then be able to track anyone inside the capture volume.
Click here to learn how to calibrate.
How to achieve optimal finger tracking:
Appropriate clothing
When using our system we recommend wearing lighter clothes so that the system can visualise your joints with greater accuracy. Not to mention that wearing baggy clothes can alter the performance of the system.
Wearing a long sleeve top which covers your wrist will affect the system’s ability to track your hands which will also negatively impact the finger tracking on your video.
Filming environment
We need to have optimal lighting conditions. This means the lighting can’t be too bright and overpowering as well as too dark.
The contrast between the actor and the background should be high in terms of colour. For example, it would be better if the actor stood out in the capture volume instead of mixing in with the background. This also links to what clothes the actor wears. For example, you don’t want the actor wearing a top that has the same colour as their hand.
Hand positioning
The fingers and wrist need to be clearly visible to as many cameras as possible when creating your takes. Moving closer to two of the cameras will achieve a better result as this will increase the pixel count on your hands.
Bad finger tracking
There are a number of different reasons why users may get bad finger tracking on their animation data.
Hands in pockets
Hands clenched together
Poor Calibration
Hands hidden by an object being held
Hands heavily occluded by objects in the scene
Fingers out of sight of cameras
Fingers too far away from cameras