[open] Simple tracker experiment not working (Tobii drift check problem)
Hi,
We are using a Tobii X120 Eye Tracker and we want to use pyGaze to communicate with it.
However, when we executed the simple_tracker_experiment.py example that comes with pygaze, we have encounter problems. It runs the calibration and validation smoothly, but it freezes when it goes into the while loop that checks the drift correction.
No errors appear, so we do not know what may be the cause of this.
Any help will be much appreciated, thanks.
jl
Comments
Welcome to the forum!
Great to hear you're testing PyGaze on a Tobii! I've had another look at the example, and it seems the while loop is obsolete. Could you try replacing it with
tracker.drift_correction()?If this fails to run properly, could you answer the following questions?
1) What happens when you press the Space bar during the drift check?
2) What happens when you press the Q or the Escape key during the drift check?
3) Does everything work normally if you remove the drift check?
Thanks!
Edwin
Dear Edwin, thank you for your response, sorry I could not use the eyetracker until today.
If I replace the while loop with
checked = Falsetracker.drift_correction()the experiment runs correctly, but each time it gives me the feedback (any of "correct/incorrect") and I press the SPACE bar to go back to the cross, it sound a buzz, like an 'error' sound.
If I remove the
checked = Falseline, the experiment window does not respond and I have to close it manually. Any key I press, does not have any effect.>3) Does everything work normally if you remove the drift check?If the drift check is removed, the experiment runs normally, but it does not wait for me to press the SPACE bar after the feedback is given. That 'error' buzz does not sound either.
By the way, if in the calibration stage, no enough data was collected, the program window
also freezes and I have to abort the program. Is there a solution for this? Is it possible to repeat the calibration in a case like this?
Finally, the last output line is
WARNING! libtobii.TobiiController.flushData: no data to write to file.What does that mean? The file default_TOBII_output.tsv is not empty, it contains data for all the trials.
Thank you very much,
jl
Hi @jlda,
Thanks for checking all that!
The buzz sounds when a drift correction fails. That is: if the sample that was obtained before you pressed the space press was not reasonable close to the fixation, you are beeped. This behaviour sometimes occurs even if you are looking at the fixation dot, but have set the wrong
DISPSIZE(this should match your resolution). Could you check whether this is true for you? If not, please do report back!The fact that you do not have to press Space when you remove the drift check, is because pressing Space is part of the drift check. The same is true for the buzz sound.
I'm still figuring out what happens when there is too little data for the calibration. Will come back to that as soon as I find a testing device around here (when I switched universities, I lost access to a Tobii).
Finally, the warning is nothing to be concerned about. Sometimes the
flushDatamethod is called internally, sometimes even when there is nothing to flush (to be sure that data that does not need to be in the log will not end up there).Good luck!
Edwin
Hi Edwin!
Thank you very much for your response and sorry for taking so long in answering.
I have not had the chance of checking if with the right
DISPSIZE, the buzz does not sound. I will update as soon as I can.Edit: I had the wrong resolution. But I have change it and the buzz sound still persists after the 'correct/incorrect' feedback.
Thank you,
jl