All the settings of faceshift studio are saved in a single file called “faceshift.ini” located in the faceshift studio workspace. In the application menu you have the possibility to export the current settings to a different ”.ini” file. This new file is basically a copy of “faceshift.ini” which always holds the current settings state. When importing existing ”.ini” files, the current settings will be replaced. The Reset option enables you to retrieve the default values of all the settings.
Note that filter and bounds presets are not completely replaced because the existing ones might be used by some clips and need to stay in the preset lists. The imported presets will simply be added to the current ones. The existing presets are only kept temporarily and will not be saved automatically.
General Preferences dialog
This tab shows you some general settings. Here you can:
- Display and change the current workspace used by faceshift studio. This is the directory where the application stores actor profiles and recorded clips, but also the faceshift studio license and log data.
- Individually enable/disable pop-up messages/warnings
- Change the style of the user interface
- Select whether the sensor should be turned off when faceshift studio is not using it. Note that the sensor will not be turned off immediately when you pause the tracking, but only after a moment of inactivity.
- Change the number of cores being used. By default, faceshift studio will use the maximum number of cores available on your system for better performance.
- Select which kind of updates you want to be notified about, and how frequently the auto-update tool should check for new updates.
These settings apply to both the live tracking and the offline retracking, achieved when the Retrack button is pressed (see section Retrack).
Tracking Preferences dialog
Check out how to influence of the various tracking settings on the performance:
Tutorial - Tracking Settings
- Frame Width/Height - Size of the image region cropped around the tracked face. Smaller values can slightly improve performance and produce smaller files when recording.
- Iterations - Set the number of optimization iterations performed during tracking. Values larger than 4 do not produce noticeable gain.
- Bounds - This is the same setting as in the Track panel of the main window. Use Bounded when retargeting blendshapes; use Unbounded tracking when exporting virtual markers.
- Regularization - Pull blendshape coefficients towards zero. An increased value creates sparser blendshape curves, but results in less accurate tracking.
- Geometry/Texture - Weight given to the depth/video data for tracking.
- Lip-Closing - Higher values prefer closed lips when detected in depth and video image.
- Brow/Mouth Features - Weight of the eyebrows/mouth feature detections from the video.
- Eye diameter - Used to compute the eye gaze direction given the tracked pupil position.
- Eye Following - Higher values activate the eye blendshapes more with respect to the tracked eye gaze.
These settings apply to the post processing of recorded clips, i.e., when the Refine button is clicked (see section Refine).
Post Processing Preferences dialog
- Tracking Iterations - Set the number of optimization iterations performed during tracking.
- Optimization Frames - Number of frames taken into account for optimization.
- Remove Minor Coefficients - Check to remove blendshape coefficients smaller than a threshold.
- Coefficients Threshold - Change the threshold for defining the coefficients to be removed.
These settings control the streaming output to a remote machine.
Streaming Preferences dialog
- Protocol, IP address, and port of the destination
- Head Pose - Select whether the head pose should be streamed in relative or absolute orientation. “Relative to neutral” streams it with respect to the neutral pose set by the Orient Head button, “Absolute” streams it with respect to the sensor, i.e., the sensor is at the origin.
- Markers - use the buttons to load and edit the set of virtual markers whose coordinates should be used for streaming. You can learn more about virtual markers in sections Marker Set Editing and Export of Virtual Markers.
- Reference - defines which shape will be used to compute the marker positions, the actor-independent Template model or the Actor model that is used for tracking.
These settings control the audio in- and output. They allow you to select the quality of the audio recording (number of channels and sample rate) and which devices to use for input/output.
Audio Preferences dialog