Fixation duration in Tobii Glasses Analyzer

Home Forums Legacy SDKs Fixation duration in Tobii Glasses Analyzer

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #3776
    Sebastian Kraus
    Participant

    Hello,

    I have one issue with the calculation of the fixation- and saccade-duration. What if there is only one unique saccade/fixation in the dataset? (e.g. fixation (timestamp 0), fixation (timestamp 1), saccade (timestamp 2), fixation (timestamp 3), …) The duration for the first fixation is timestamp(1) – timestamp(0). But for the saccade it calculates duration of 0ms.

    Thank you for your advice.

    Greetings from Germany.

    #3788
    Jenny [Tobii]
    Participant

    Hi Sebastian,

    Do you mean that you have a problem with how the Fixation data stream in EyeX works, or that you have a problem calculating fixations for a specific corner case in an algorithm you have written yourself?

    #3789
    Sebastian Kraus
    Participant

    Hi Jenny,

    I’m working with the Tobii Pro Glasses Analyzer export function with the I-VT fixation filter. After exporting the tsv I calculate the fixation- (and saccade) duration by grouping the eyeMovementType with eyeMovementTypeIndex and apply “timestampEnd – timestampStart”. I wondered because the mean fixation duration for our participants are between 90 and 160ms. Is this realistic?

    Thank you for your feedback.
    Sebastian

    #3790
    Jenny [Tobii]
    Participant

    Hi Sebastian,

    I extracted your question to a separate topic since it is not related to Tobii EyeX software.
    I’ll notify my colleagues at Tobii Pro to take a look at this issue.

    #3794
    Sebastian Kraus
    Participant

    Thank you Jenny! I’m looking forward to hear from Tobii Pro Team 🙂

    #3795
    Jonas Högström
    Participant

    Hi

    If a saccade or fixation is only one gaze sample long, the duration will be equal to the interval between two samples (16.666ms on a 60hz tracker). It will start half a sample before the first sample that is included and end half a sample after the last sample.

    So in your original question, the first fixation will be two intervals long (not timestamp2-timestamp1, but rather (timestamp2-timestamp1)*2) and the saccade will be one interval long. The saccade will start half way between timestamp2 and timestamp3 and will end between timestamp3 and timestamp4. These are the times used when calculating statistics/metrics (fixation durations, time to first fixation etc.). They are also visible in the Replay module when hovering the mouse over a fixation or in the Visualization module when looking at a gaze plot.

    The start and end-times of fixations/saccades are not available in data export.

    /Jonas, software architect, Tobii Pro.

    #3801
    Sebastian Kraus
    Participant

    Thank you Jonas,

    that is exacly the information I needed 😉

    Greetings from Germany!

Viewing 7 posts - 1 through 7 (of 7 total)
  • You must be logged in to reply to this topic.