Knowledge Base

How FaceReader Data Are Imported Into The Observer XT

See Visualize FaceReader data in The Observer XT for a full overview of how FaceReader data appear in The Observer XT.

Event Data

FaceReader event data are imported into The Observer XT as separate event logs within the observations. This is the case for the following FaceReader data:

• State Expression Values
• Facial States
• Stimuli
• Event markers
• Action Units (with the Action Unit Module, if you exported them to an odx file)
• Consumption Behavior events (Intake event, Chewing and Chew motion, with the Consumption Behavior Module)
• Gaze direction

Continuous Data

Continuous data are imported into The Observer XT at a sample rate of 5 frames per second. Your license for The Observer XT must include the External Data Module. The following data can be imported as continuous data:

• Facial expression intensity values.
• Valence.
• Arousal.
• Head orientation (Pitch, Yaw, Roll).
• Head position (Horizontal, Vertical and Depth position).
• Gaze angle (Vertical and Horizontal gaze angle).
• Action Units (with the Action Unit Module, if you selected 'Export action units as continuous values' in the Settings and exported them in a Detailed log file).
• Heart rate, heart rate variability, and respiratory (breathing) rate (if your FaceReader license includes the Vital Signs Module).
• Custom Expressions (if your FaceReader license includes the Action Unit Module).
• Expressions from voice (including Valence from voice, Arousal from voice, Loudness and Speech rate, with the Voice Analysis Module).

Note: See Set up FaceReader to work with The Observer XT how to select the data to export from FaceReader and to import in The Observer XT.

Independent Variables

The FaceReader independent variables Age, Gender and ParticipantID are imported into The Observer XT together with the FaceReader data. The values for Age and Gender are estimated by FaceReader and the ParticipantID is the participant number automatically assigned by FaceReader. If the independent variables are not present yet in The Observer XT, they will be added to the Independent Variables list. If you created them as User defined variables in the Independent Variables list of The Observer XT (with the exact variable names as mentioned above) and no values are filled in yet before observation start, the FaceReader values are imported into the list after the end of the observation. If you already entered values in The Observer XT before observation start, these values are maintained. FaceReader does not overwrite the Independent Variable values in The Observer XT.

To enter the independent variable values in The Observer XT instead of importing the FaceReader values, create user defined independent variables Age, Gender and ParticipantID in The Observer XT. Make sure the names match exactly. Choose to enter the independent variables before starting the observation (Choose Setup > Project Settings > Observation Settings and choose Before Observation from the Edit Independent Variables list).

To import FaceReader independent variable values in the Independent Variables list of The Observer XT, the variable name must exactly match the FaceReader variable name. FaceReader contains by default the independent variables Age, Gender and ParticipantID. If you, for example, have an Independent Variable age in The Observer XT (lower case a instead of upper case), or Participant ID (with a space), a new Independent Variable Age or ParticipantID will be added to the Independent Variables list when you import the FaceReader data.

Notes

Duplicate Elements

IMPORTANT Elements that are imported as events into The Observer should not already exist in the Coding Scheme of The Observer XT with the exact same name. As an example, you may have a mutually exclusive behavior group that contains the behavior Neutral, or Unknown. If you import FaceReader emotional expressions into The Observer XT, the expressions Neutral, or Unknown will then be scored for the behavior group that was already present in the Coding Scheme and will not be present in the imported event log State Expression Values. Therefore both the event log with the manual annotations and the event log State Expression Values will be incorrect.

Sample Rate

FaceReader data are imported into The Observer XT with a sample rate of 5 frames per second. This is to limit the amount of data that is transferred to The Observer XT. FaceReader itself samples with the camera frame rate and converts the data to a sample rate of 15 frames per second after the analysis is finished. Therefore, the data in The Observer XT may differ slightly from the analysis results in FaceReader.

See Camera frame rate and samples in FaceReader in Set up your project

Vital Signs

If your FaceReader license includes the Vital Signs Module, FaceReader can estimate the test participants' heart rate, heart rate variability, and respiratory (breathing) rate. These data can be imported real-time in The Observer XT.

In FaceReader, the first 15 seconds of the analysis is used for calibration and the heart rate and breathing rate data are post-processed after the analysis is finished. The Observer XT imports the heart rate and breathing rate data real-time. Therefore, the heart rate and breathing rate data of the first ten seconds that are imported into The Observer XT contain data that are not yet calibrated. The data that are imported are not post-processed and will differ from the FaceReader data. Since post-processing is used to re-calculate the data, based on the calibration, the data in FaceReader itself are more reliable.

Whereas for heart rate the first 8.5 seconds of an analysis are post-processed after the analysis is finished, for breathing rate this is 15 seconds and for heart rate variability this is the first 5 minutes (or even longer).

See Vital Signs Module

Event Log Duration

• There will be a slight delay between the moment you start an observation in The Observer XT and the moment FaceReader starts analyzing. Since the end of all event logs is determined by the end of the observation, the FaceReader event log will have a shorter duration than The Observer XT event log in the analysis. This is caused by the offset between the event logs.
• The observation in The Observer XT ends when you click the stop observation button. However, the FaceReader analysis will stop when you click the stop observation button.

Source: FaceReader 10 Reference Manual (Help), Noldus Information Technology