The second step to take is to select the "model" to use. Model is sort of a marker, attached to ones head, simplifying the head pose determination (unless you are using facetracking, then the "model" is your face).

There are several available model types:

Three point Cap - provides 6 degrees of freedom
Three point Clip - provides 6 degrees of freedom
Single point - provides only 2 degrees of freedom, but the highest precision
Face - provides 3 degrees of freedom

Model can be either active (using LEDs or other light emmiting components) or passive (using reflectors), although this differentiation is necessary only when you use the TrackIR or SamrtNav devices - depending on the model type the TrackIR/SmartNav turns its illumination IR LEDs on.

If you intend to build a model on your own, loook at the Freetrack project pages - it contains the "model" creation hints, tutorials and more.

There are several predefined models - "NP TrackClip", "NP TrackClip Pro", "Face", "Single point reflective" and "Single LED" - if you use one of those, just select the one of your choice in the Model Name combobox and you can continue to the next pane.

If none of the predefined models suits you, you can define your own.

Model Creation

Press the Create New button and the new model creation form will pop up.

Type your model's name to the Model Name box and then select the appropriate model type. Available model types are these:

1 point - single LED diode or reflective marker; simple but gives you only 2 degrees of freedom (pan verticaly and horizontaly)
3 point Cap - three LEDs or reflective markers, for example on the top of your cap; provides 6 degrees of freedom
3 point Clip - three LEDs or reflective markers, for example on the side of your headset; provides 6 degrees of freedom
Face - used for face tracking; provides 3 degrees of freedom (pan verticaly and horizontaly, move back and forth)

Now fill-in the model dimensions (where applicable) according to the diagram (all dimmensions are in milimeters). The last thing to do is to select, if the model uses LEDs (obviously with the exception of face model type) - if it does, check the Model with LEDs checkbox. After pressing the Create button, the new model should become available in the Model Name combobox for you to select it.

Model Tweaking

To minimize unwanted viewpoint movement while panning, you can tweak the 3-point Clip and 3-point Cap model. Note however, that the movement can't be fully suppressed, both due to the head movement model simplification (assuming that all panning is done using a single joint) and also due to the human's inability to stand absolutely still while turning the head. Note also that turning the head inherently contains some viewpoint movement - for example look down at your shoes and realize that your viewpoint shifted forward and down...

Model tweaking procedure

Both three point model types share the same system of tweaking - there are several sliders, that allow you to adjust the model's center-point and minimize the unwanted movement. To see the particular movement precisely, go to the Tracking setup tab, open the Detailed axes setup dialog and select the axis of your interest. Observe the horizontal movement of the cross in the axis visualization area, that represents computed head movement or angle.

For example, the first slider in the 3 point Clip model tweaking pannel allows you to minimize the back and forth head movement while looking left and right. To find the "sweet spot", employ the "halving of the interval" method, utilizing the fact, that at that point the movement is smallest and the movement direction reverses.

Move the slider to the left extreme and observe the Z axis (back/forth movement), while panning left and right - note the direction of the movement of the cross; for example, when looking left, the cross moves right and vice versa. Now move the slider fully to the right and see that the response is reversed (e.g. look left, move left).

The spot you are looking for is somewhere in between the two. Move the slider to the middle and see what the response is in there; for example the look left moves the cross to the right - which is the same direction as the leftmost slider extreme. That means, that the sweet spot lies in the second half of the interval (in this example). Now you can continue by repeating the halving it, or just try moving the slider around in small increments to minimize the movement.

Repeat the above steps for both (3-point Cap)/all three (3-point Clip) sliders to reduce the amount of the unwanted movements; if the result doesn't satisfy you, you can remove the rest of it by increasing the deadzone for the axis of interest.