Graph Plots reference

This section provides details in addition to those in the rest of the Graphing your data section, about options and modifiers found within the Graph Plots view.

For a description of all channel types and the channel sources in each type, see Channel types.

For information on the calculations used within Graph Plots, including modifiers, see the relevant topic:

• Derivatives
• Rotation unwrapping
• Magnitude

For a more general introduction, see Graphing your data.

Channel types

Analog Devices

• Device Output The output from an analog device
• Reprocessed Device Output The output from an analog device re-calculated from the raw analog source data. This is slightly more expensive to compute than the Device Output and should produce the same result when not in Review mode, but may have fewer gaps in Review mode.

Cameras

• Accelerometer The movement measured for the camera from the onboard accelerometer. This is given as a unitless vector scaled by acceleration due to gravity; the number varies, depending on the orientation of the camera. The magnitude of the accelerometer at rest is approximately 1, with some variation according to calibration.
• Centroid Count The number of centroids detected by the camera. (Note that in Review mode, this channel is only supported if the camera's Grayscale Mode is set to Auto or All.)

Objects

• Enabled Count The number of objects enabled in the Tracking panel

• Position The object position relative to the volume origin
• Quality The object quality. This is a measure of the fit of the object to the observed marker positions, and is influenced by the accuracy of the object model, and the camera coverage and accuracy of camera calibration at the current object position and rotation
• Rotation The object rotation relative to the global coordinate system axes
• Tracked Count The number of objects currently tracked

Relative Measurements

• Angle From 3 Positions The angle ABC defined by the positions A, B and C (objects, markers or reconstructions)

• Difference Between Positions The difference between two positions (objects, markers or reconstructions)
• Relative Position The position of object B relative to object A (in the frame of reference of object A)
• Relative Rotation The rotation of object B relative to object A

System Health Metrics

• Camera Calibration Health Assessment of the overall camera calibration health
  
  • Plots Image Error and Centroid Connectivity. For more info on either of these metrics, see Monitor system health.
• Datastream Output Latency Latency measured for the live low-latency datastream port

  • Delivery latency is the time difference between the start of data capture at the cameras and delivery of the data to the application. This covers internal camera processing, network transfer and host PC data receipt.

  • Processing latency is the time difference between the data delivery time and processed datastream output. This covers host PC processing to track the observed objects and output tracking data.

  • Total latency is the sum of the delivery and processing latency.

• System Accelerometers The movement measured over all cameras from the onboard accelerometer
  • Mean magnitude is the average acceleration vector magnitude of all cameras. This is given as a multiple of the acceleration due to gravity, with 1 indicating normal acceleration at rest.
  • Maximum magnitude is the highest acceleration vector magnitude of any camera. This is given as a multiple of the acceleration due to gravity, with 1 indicating normal acceleration at rest.
• System Centroid Count The total number of centroids detected in all cameras. (Note that in Review mode, this channel is only supported if the camera's Grayscale Mode was set to Auto or All.)
• System Temperature The temperature measured for all cameras
  • Current mean is the current average temperature of all cameras
  • Calibration mean is the average camera temperature measured at the end of the data collection stage of a Calibration Wave operation

Trajectories

• Labeled Marker Count The number of object markers currently tracked

• Marker Position The marker position relative to the volume origin
• Reconstruction Position The unlabeled reconstruction position relative to the volume origin
• Unlabeled Reconstruction Count The number of marker positions reconstructed from camera centroids but not labeled as part of an object

Derivatives

Default derivatives

For most channel types, the derivatives are computed using a central finite difference approximation.

First order (velocity):

x'(t + 1/2) = x(t + 1) - x(t)

Where x is the channel sample value and t is the channel sample number.

Note that first-order derivatives are offset by half a sample with respect to the input channel.

Second order (acceleration):

x''(t) = x(t + 1) - 2x(t) + x(t - 1)

Rotation derivatives

First order rotational derivatives (rotational velocity) are computed using the relative rotation between successive samples:

R'(t + 1/2) = R^-1(t)R(t+1)

Second order rotational derivatives are computed using the central finite difference of the first order rotation derivative:

R''(t + 1) = R'(t + 3/2) - R'(t + 1/2)

Rotation unwrapping

The angle range (-π, π) radians is sufficient to describe all orientations. However, as an object continues to rotate in the same direction and approaches these boundaries, a flip in orientation (and sign) can be observed to keep the angle within the (-π, π) radians range. Rotation unwrapping allows you to strategically add factors of 2π to the angle to generate equivalent representations of the same orientation. This helps to remove any discontinuities in the data.

For example, an object rotating through an angle of 4π might result in a trace like this:

Note that unwrapping doesn't necessarily tell you the trajectory of orientations the object went through to get from A to B, and the results are not easy to interpret if the axis of rotation changes. Gaps in the object tracking can also cause the rotation unwrapping to restart.

Rotation unwrapping is enabled by default, but can be disabled in user preferences (Settings > Preferences > Graph Plots section > Enable rotation unwrapping).

Magnitude

Magnitude can be used to compute the resultant of a vector quantity, or the absolute value of a scalar quantity. Some common use cases include:

• Magnitude of position = distance from origin
• Magnitude of rotation = angle rotated about the rotation axis
• Magnitude of difference between or relative position = distance between
• Magnitude of velocity = speed