Tracking is a technique that involves analyzing the motion of an image over time. Images can be tracked using an automatic image tracking system called planar tracking using either Silhouette’s Planar Tracker or the Mocha Planar Tracker. In addition, you can use a Point Tracker which utilizes either one, two or four track points. One point tracking obtains horizontal and vertical movement. Two point tracking analyzes scaling and rotation in addition to horizontal and vertical movement. Lastly, four point tracking evaluates and applies perspective movement using a corner-pin transformation.

However the tracking data is obtained, it can then be applied in one of two ways: match moving or stabilization. Match moving applies the motion of the trackers to an object so that it follows the motion of the clip. Stabilization inverts the tracking data so that the clip appears to be stable.

Transform.

The key to getting the most out of the Planar and Mocha trackers is to learn to find planes of movement in your shot which coincide with the object that you want to track. Sometimes it will be obvious--other times you may have to break your object into different planes of movement. For instance, if you were tracking a tabletop, you would want to draw the spline to avoid a flower arrangement in the center of the table, since it is not on the same plane and will make your track less accurate.

Both the Planar and Mocha trackers require a selected layer with at least one shape in it to drive the tracker.

The shape should surround the object you are tracking, but it does not have to be exact and it is best to leave a little extra room around it. In addition, you can use more than one shape as long as they are on the same geometric plane.

In some cases, there are parts of an image that can interfere with the effectiveness of the Planar and Mocha trackers. To handle this, you can create an exclusion zone in the area you are tracking. For instance, a computer screen with strong reflections could cause the track to jump. To exclude an area while tracking, create a subtractive shape above the primary tracking shape. Then, select both the shapes prior to tracking.

Go to the Planar Trackers tutorial to see how it works.

As you track, the tracker’s transform data is automatically applied to the selected layer.

Determines which image channel the tracker uses to analyze the image.

The tracker analyzes the luminance or brightness of the image.

The tracker analyzes the image’s red channel.

The tracker analyzes the image’s green channel.

The tracker analyzes the image’s blue channel.

Tracks corners, edges and ridges and automatically selects the best result.

Tracks corners, crosses or points.

Tracks edges of a solid object.

Tracks lines appearing as a strips or bands.

Tracks perspective changes. Good for large shapes.

Tracks translation, rotation, scale, and skew. Good for medium shapes.

Tracks the XY offset. Good for small shapes.

The minimum number of trackers to generate.

The maximum number of trackers to generate.

The minimum distance between trackers.

The maximum number of frames a specific tracker will live until it is removed and a new tracker is generated.

The number next to Valid Tracks is a running total of the number of trackers that are viable as the tracker progresses.

Displays the reference frame number. The reference frame is the frame you first started tracking on and it locks in the picked set of track points at that frame.

The Mocha Planar Tracker provides 2D transformation data by tracking planes rather than points.

As with Silhouette’s Planar Tracker, the Mocha Planar Tracker requires that you create a layer in the Object List and have at least one shape within it to drive the tracker. The shape should surround the object you are tracking, but it does not have to be exact and it is best to leave a little extra room around it. In addition, you can use more than one shape as long as they are on the same geometric plane.

As you track, the tracker’s transform data is automatically applied to the selected layer.

Determines which image channel the tracker uses to analyze the image.

The tracker analyzes the luminance or brightness of the image.

The tracker analyzes the image’s red channel.

The tracker analyzes the image’s green channel.

The tracker analyzes the image’s blue channel.

Min % is one of the most important parameters to look at for tracking. By default, the minimum percentage of pixels used is dynamic. When you draw a shape, the Mocha Planar Tracker tries to determine the optimal amount of pixels to look for in order to speed up tracking. If you draw a very large shape, the percentage will be low. If you draw a small shape, the percentage will be high.

In many cases, the cause of a drifting or slipping track is a low percentage of pixels. If you want a more solid and accurate track, try setting the Min % Pixels Used value to a higher amount. Keep in mind, however, that a larger percentage of pixels can mean a slower track.

Min % Pixels Used is set to Auto by default.

These Motion parameters control what motion you are looking for when you track.

The position of the object.

Whether the object gets larger or smaller.

The angle of rotation of the object.

How the object is skewing relative to the camera.

How the object is moving in perspective relative to the camera.

The main difference between shear and perspective is the relative motion. Shear is defined as the object warping in only two corners, whereas perspective is most often needed where the object is rotating away from the viewer significantly in space.

As an example, if someone is walking towards you, their torso would be showing shear as it rotates slightly back and forth from your point of view.

The front of a truck turning a corner in front of you would be showing significant perspective change.

This is the default. It searches for motion and optimizes the track as it goes. Small Motion is also applied when you choose Large Motion.

Small Motion only optimizes. You would use Small Motion if there were very subtle changes in the movement of the object you are tracking.

The Search Area defines ranges for the tracker to search within.

Determines the distance in pixels within the footage to search for the next object position. This is set to Auto by default.

If you have a fast rotating object, like a wheel, you can set an angle of rotation to help the tracker to lock onto the detail correctly. The tracker will handle a small amount of rotation, less than 10º per frame, with Angle set to zero.

If you have a fast zoom, you can add a percentage value here to help the tracker. Again, the tracker will still handle a small amount of zoom with this set to zero.

The inner box is the Match Area. It defines a pattern that will be searched for from frame to frame. It's a good idea to choose a region with good contrast and detail. Corners with sharp contrast are usually good areas to track as movement can easily be detected in any direction. The Match Area can be scaled to the desired size.

The outer, dashed box is the Search Region, which should be the maximum amount your tracking point will move between frames. The larger the size, the slower the processing. The trick is to make the Search Region as small as possible, yet still large enough to cover the area the tracker will move from frame to frame. The Search Region can be sized in the same fashion as the Match Area.

The center cross is the track point. It represents the position of the motion track. Normally, the Track Point is at the center of the tracker, but can be offset if the Match Area becomes obscured.

At each frame, the Match Area is compared to the first frame at which you started tracking. If you stop tracking halfway through your clip, and start up again at a later frame, the later frame will be used as the reference.

The Match Area is compared to the previous frame. This method will cause inherent drift in your track, but is useful for clips that have drastic changes in perspective and scale.

For difficult shots, it is helpful to pre-set keyframes for the tracker at various points in the clip. The Match Area will use these keyframes as a reference while tracking.

Describes the level of accuracy between the Match Area that the Tracker is searching for and the area it actually finds when searching from frame to frame. A prefect match would be a value of 100. If the error that is calculated is below the Tolerance value, no keyframe will be set on that particular frame. The Tolerance is preset to a value of 75.

When OK is clicked, the tracking data is transferred to the Layer’s > Transform > Matrix parameter.

Go to the Applying Trackers tutorial to see how it works.

Applies X/Y position data to the layer.

Applies X/Y scaling data to the layer. Requires two trackers.

Applies rotation data to the layer. Requires two trackers.

To average trackers, select more than one tracker and click the Average button. A new averaged tracker is created.

Go to the Averaging Trackers tutorial to see how it works.

Ideally, the trackers should overlap in time by at least one keyframe.

Go to the Merging Trackers tutorial to see how it works.

Most tracker parameters are interactively set while moving the tracker in the Viewer. When a point tracker is selected, its parameters are displayed in the Object Parameters.

A magnified version of the Match Area is displayed to aid in the placement of a tracker. It is activated when you click the Create button or when you move an existing tracker. The magnified area is displayed in the color space selected from the Mode pop-up menu.

Sets the position of the Track Point. This parameter is automatically set when using the on-screen tracker controls.

Trackers can be nudged using the Arrow keys. One press of the Arrow key moves the tracker 1 pixel. Using the Shift key in conjunction with the Arrow keys moves the tracker 10 pixels. Ctrl plus the Arrow keys moves the tracker one tenth of a pixel. These nudge values can be changed in the Preferences menu located in File > Preferences > Nudging on Windows and Linux or Silhouette > Preferences > Nudging on Mac. Holding the Arrow keys down slides the tracker.

Determines which image channel the tracker uses to analyze the image.

The tracker analyzes the red, green and blue channels of the image.

The tracker analyzes the luminance or brightness of the image.

The tracker analyzes the image’s red channel.

The tracker analyzes the image’s green channel.

The tracker analyzes the image’s blue channel.

Offsets the Match Area from the tracker. This parameter is automatically set when using the on-screen tracker controls.

Offsets the Search Region from the tracker. This parameter is automatically set when using the on-screen tracker controls.

Sets the size of the Match Area. This parameter is automatically set when using the on-screen tracker controls.

Sets the size of the Search Region. This parameter is automatically set when using the on-screen tracker controls.

Trackers can be created based on a shape’s points, the center point of a shape or the shape’s bounding box. Trackers created in this manner are a concatenation of the shape and layer transformation data. So, why do this? Because, it is a handy way of creating trackers with all of this combined data for use in other programs.

Creates four trackers around the four corners of the shape bounding box.

Creates one tracker in the center of the shape.

Creates trackers from selected points.

Planar Tracker track points can be converted to point trackers which can then be treated like any other point tracker; they can be merged, smoothed, averaged, and applied to layers. Right-click on a planar track point, choose Create Point Tracker and a new point tracker is created.

Images that contain film grain or video noise, lack contrast or sharpness as well as images that have flicker can be tracked more accurately by applying certain filters during the tracking process. Filters such as Blur, Sharpen, Contrast, Gamma, De-Noise and Remove Flicker will increase tracker accuracy for problematic images.

Applies a blur to the image features.

Sharpens image features.

Increases contrast.

Increases brightness by leaving the white and black points the same and only modifies the values in-between.

Applies a de-noise algorithm to smooth out noise while retaining detail.

Corrects image flicker or brightness variation over time, which has a habit of interfering with tracker accuracy.

Shows the effect of the Pre-Processing parameters on the entire image in the Viewer.

To smooth, select a point tracker or planar tracked layer and click the Smooth button. Adjust the slider when it pops-up and adjust it to the desired level of smoothing.

Go to the Smoothing Point Trackers or the Smoothing Planar Tracked Layers tutorials to see how it works.

Importing and exporting tracking data takes place in the I/0 section of the Tracker tab.

After Effects Corner-Pin, Nuke, Shake or Simple Format formatted trackers can be imported into Silhouette. If you would like to import trackers from non-supported applications, simply export a Silhouette tracker in Simple Format to see how it should be formatted.

Go to the Importing Tracking Data tutorial to see how it works.

Trackers can be exported to the After Effects, Autodesk, Nuke and Shake tracker formats. In addition, the tracking data applied to a layer, either by the Point Tracker or Planar Tracker, can be exported as a four point corner-pin track.

To create trackers for export that are a concatenation of the shape and layer transformation data, see the Tracker Pop-Up Menu section for more information.

Go to the Exporting Tracking Data tutorial to see how it works.

Shapes, points or tracker vectors can be toggled on or off.

Toggles the display of shapes in the Viewer.

Toggles the display of Planar Tracker track points in the Viewer.

Toggles the display of tracking vectors in the Viewer. This option is not available for use with the Planar Tracker.

See the Tracker Preferences for information on their use.

Go to the Match Move or Stabilize Images tutorial to see how it works