Welcome!

Sign in with your CogSci, Facebook, Google, or Twitter account.

Or register to create a new account.

We'll use your information only for signing in to this forum.

Supported by

EyeTribe - Inconsistent Data Quality

tsummer2tsummer2 Posts: 20

I was wondering if anyone using the eyetribe ever got bad/noisy recordings occasionally, even when some other recordings are good? We have been having a heck of a time troubleshooting our occasional bad recordings. It seems like only 1 out of 4 recordings are actually good looking data... A lot of recordings are just coming out with values equal to zero for a lot of samples even though we are calibrating below .7 degrees error, and using a chin mount. Just curious if this is happening to other people. Thanks!

Comments

  • JoshJosh Posts: 360

    Hi,

    Values equal to zero are usually due to a loss of signal, e.g. due to blinking. Are your participants instructed to not blink during a trial? Aside from that, of course the EyeTribe is not the greatest eye-tracker, despite it being amazingly cheap for what it can do. It may be that things such as eyeliner (with female participants) distort the data to a stronger degree than would be the case with other eye-trackers.

    Cheers

    Josh

    Thanked by 1tsummer2
  • tsummer2tsummer2 Posts: 20

    Thanks Josh. Participants are allowed to blink naturally, however the signal we are receiving is clearly unnatural. Here is an example graphed with plotly: https://plot.ly/~tsummer2/57/raw-x/
    Each time series represents a separate recordings of the Raw X gaze position, and each time series is about 50 seconds long. You can see that some of the trials have a lot of unnatural signal dropping, while others only have occasional blinks.

  • sebastiaansebastiaan Posts: 2,594

    Basically what Josh says. But one thing to be aware of with the EyeTribe is that it's quite prone to losing the eye if you move the head a bit, and this an all-or-nothing deal. A good trick is to tell participants that they should be able to see the reflection of their own eyes in the EyeTribe. That's the optimal position, and gives the most room to move a bit.

    Also, replacing the 0's with nan values will do wonders for your signal!

    Thanked by 1tsummer2

    There's much bigger issues in the world, I know. But I first have to take care of the world I know.
    cogsci.nl/smathot

  • tsummer2tsummer2 Posts: 20

    The crazy thing is that the recordings I shared are all of me, doing the same thing without moving the tracker or head mount. I am still not sure what's going on, but I'll update here if I find a cause

Sign In or Register to comment.