To influence the facetracker while 'in-game', shortkeys can be defined to:
The user can define the curve, by which actual head-movement (vertical axis) is translated into in-game movement (horizontal axis). The curve can be changed by 'grabbing' a circular handle and dragging it (within the light-blue area). There are two tabs: one for rotations and one for translations.
The curves can be changed while the tracker is running.
FaceTrackNoIR currently only supports one face-tracker (faceAPI from SeeingMachines). Since we are prepared for the future, this tracker is selected using the Tracker Source listbox. When faceAPI is running, some of it's details can be viewed using the Settings button.
The FTNoIR server was introduced to make it possible to run FaceTrackNoIR on a different PC, so the CPU-load of the 'gaming-PC' can be reduced. To receive headpose-data from the 'face-tracking-PC', the FTNoIR Server must be selected as Tracker Source. When the tracker is not running, the port-number for the PC can be edited via the Settings Dialog.
The game-protocol can be selected with the Game-protocol checkbox. Check out the Compatible games section, if you don't know which protocol to use. Some protocols have their own settings. If so, the button will appear (with the protocol icon).
The FTNoIR server was introduced to make it possible to run FaceTrackNoIR on a different PC, so the CPU-load of the 'gaming-PC' can be reduced. To send headpose-data to the 'gaming-PC', the FTNoIR Client must be selected as Protocol. When the tracker is not running, the IP-address and port-number for the remote PC can be edited via the Settings Dialog.
It is very likely that it is necessary to change Firewall settings in Windows to make LAN-communication possible for FaceTrackNoIR.
When the tracker is started, or on program startup, the name of the first usable video-input-device is shown by FaceTrackNoIR (Deskshare IP Camera Source in the example below). Unfortunately the non-commercial version of faceAPI (which is used, because it's free...) only uses the first camera it finds.
To process the raw headpose-data, FaceTrackNoIR has two features: smoothing and EWMA-filtering. Both features are meant to stabilize the raw headpose-data and remove spikes and jitter.
The smoothing setting works for all 6 DOF's at once. It adds the values of the last x raw values and divides them by x, thus creating a 'moving average'. Try to keep this value low, because smoothing will logically lead to a slower response.
From version 1.5 on, FaceTrackNoIR has a new EWMA-filter. It was designed to respond quicker to movement and still provide a smooth image, when not moving.
Finally, it is possible to invert an axis, by checking it's Invert checkbox. Some games interpret the -/+ directions different than faceAPI does...