Frequently Asked Questions

Frequently Asked Questions

Troubleshooting

Should I create LookZones before or after recording data?

The short answer is that you can do it either way; your results will be the same. The long answer is two-fold. First, your analysis is always performed on-demand after recording has taken place. So whether you create LookZones, record data then analyze or record data, create LookZones then analyze, your LookZones will still be used in your analysis. Second, various studies have different requirements. In some cases, you may know exactly what your areas of interest are on your stimuli, so creating the LookZones first makes sense. In other cases, you may not know exactly what the areas of interest are until after you’ve had a chance to see some of the recorded data from your subjects. In the latter cases, using some of GazeTracker’s visualizations (Contour, Spotlight, Heatmap) can help you decide where to place the LookZones.
return to top

 

 

When should I use Shift Data to correct my data?

If you can identify a calibration problem with your eye-tracker that caused your GazePoints to be generated in a way that all the points are shifted in roughly the same direction by a similar amount, you might be able to correct the data by using GazeTracker’s Shift Data function. However, even if you can shift the data to cause it to “look” correct, that doesn’t necessarily mean that it is correct. Ultimately you will still need to correct the calibration problems with your eye-tracking system to ensure the correctness of your results.
return to top

 

 

When should I use algorithmic data shift?

The Algorithmic Data Shift function should only be used to attempt to correct calibration errors with data collected using an ERICA eye-tracker.
return to top

 

 

When should I use filtering in GazeTracker?

The filtering features in GazeTracker should really only be used when working with the ERICA eye-tracker or with other eye-tracking systems that do not perform filtering and smoothing before sending the data to GazeTracker. Many of the eye-trackers that GazeTracker supports already do some level of filtering and smoothing before they send their data to GazeTracker. Because of this, using filtering in GazeTracker could be adding a second layer of filtering to your data and might introduce inaccuracies.
return to top

 

 

Why do I have gaps in my data?

The first thing to realize when trying to find the answer to this question is that GazeTracker simply records the eye-tracking data being fed to it from the eye-tracker. It is not generating that data itself.

So the first thing to check is always to make sure your eye-tracking system is operating properly and that the subject has been calibrated properly. If you’re having problems with this step, please consult the manual for your eye-tracker or contact the manufacturer.

The second thing to check is to make sure that your eye-tracker and GazeTracker are communicating properly. On many versions of GazeTracker, there is an eye-tracker specific tab with a Test Connection command on it. That command will let you make sure that GazeTracker can receive the eye-tracking data appropriately. If this isn’t working, please consult the eye-tracking specific appendix of your GazeTracker manual to ensure you have the two systems setup properly. If you’re still having problems, please contact Eyetellect support for assistance.
Lastly, you may wish to check the Gap Interval setting in GazeTracker and ensure it is at an appropriate value for your eye-tracker. See the question below for more information concerning the Gap Interval.
return to top

 

 

What is the Gap Interval setting and when should I change it?

The Gap Interval setting tells GazeTracker how large or small of an interval between GazePoints should be considered a gap in the data. When GazeTracker detects a gap, it does not connect the GazeTrail between the successive GazePoints, and it resets the fixation calculation algorithm, meaning that the end time for a fixation gets set to when the gap starts if a fixation occurs at the gap. If you’re not familiar with how GazeTracker records and calculates eye-tracking data from an eye-tracker, please see the Fundamentals of Eye-Tracking page.

If your eye-tracker is running at 60Hz, then you are optimally getting a GazePoint roughly every 16.6 milliseconds. Another way to say that would be that, under optimal working conditions, there is a gap of roughly 16.6 milliseconds between each GazePoint that GazeTracker receives from a 60Hz eye-tracker. Since that is normal operating conditions, you would not want GazeTracker to think there was a gap in the data just because it didn’t receive a GazePoint for 16.6 milliseconds. However, if it did not see a GazePoint for more than about 33.3 milliseconds (the time during which you would normally receive two GazePoints), then that should probably be considered a gap in the data. In that case, you’d want to set the Gap Interval setting to somewhere between 16.6 milliseconds and 33.3 milliseconds.

Since 16.6 milliseconds would be the gap under “optimal” conditions, you probably don’t want to set the Gap Interval too close to that. It is quite likely that you’ll have gaps between GazePoints of 17 milliseconds.

GazeTracker defaults to a Gap Interval of 35 milliseconds that is useful for 30Hz eye-trackers and which won’t be unnecessarily burdensome on faster systems. A 30Hz eye-tracker would have an optimal gap of roughly 33.3 milliseconds. You should adjust the Gap Interval to an appropriate value for the speed and accuracy of your eye-tracker. Some examples might be:

Eye-Tracker Speed Gap Interval Setting
30Hz 0.035 seconds (35 milliseconds)
60Hz 0.018 seconds (18 milliseconds)
120Hz 0.010 seconds (10 milliseconds)

return to top

 

 

Why would I use Load Layout for LookZones?

Any time you need to have multiple LookZone Layouts stored for the same stimulus (but not loaded at the same time), you’ll need to use the Load Layout command to load the various layouts as you need them.  Additionally, in Application Analysis, if you do not have a LookZone Layout set as the default layout for the software you are studying, you will need to manually load the LookZones.

For example, you may be experimenting with one LookZone Layout that has LookZones focusing on the minute details of a stimulus and another LookZone Layout that has LookZones focusing on the larger parts of the stimulus. You might want to get statistics about each of those LookZone Layouts separately in order to complete your study. In this case, you would create each LookZone Layout independently and then use the Load Layout command to load each one as necessary for analysis.
return to top

 

 

What types of images can GazeTracker display as stimuli?

GazeTracker supports the following image file types:

  • JPEG (.jpg and .jpeg)
  • GIF
  • PNG (GazeTracker 8.0 and above)
  • Bitmap (.bmp)
  • Windows Meta File (.wmf)

return to top

 

 

What types of videos can GazeTracker display as stimuli?

GazeTracker supports AVI and WMV video files. However, this question can be tricky because AVI files can be created with numerous different CODECs. The computer running GazeTracker must have the appropriate CODEC installed in order for GazeTracker to display the video. It is usually the case, though not always, that if Windows Media Player on the computer running GazeTracker is able to open and play an AVI, then GazeTracker will be able to as well.
return to top

 

 

What is the difference between a GazePoint, Fixation and GazeTrail?

A GazePoint is a measurement of the point of gaze of a subject as measured by an eye-tracker at a specific point in time. This is the data that is sent from the eye-tracker to GazeTracker and is used to calculate Fixations.

A fixation is a calculation of where and how long a subject focuses on a particular point. This calculation is performed by GazeTracker with the GazePoints it receives from the eye-tracker as well as the parameters that a researcher can set on the Options tab.

The GazeTrail is simply a visualization of the path that the subject’s eyes took as they moved around the stimuli. It can be configured to connect between GazePoints or Fixations to provide the researcher with the visualization that best suits their needs.
return to top

 

 

Do I need to analyze my data at the same screen resolution that it was originally recorded?

In general, yes. If you open a recording that was done at a resolution other than the one currently in use on the computer, GazeTracker will warn you about the situation. When you continue, GazeTracker will attempt to scale the stimulus, gaze data and LookZones to match your current resolution. This may not always work correctly, however, and you may see inaccuracies. It is always a good idea to perform your analysis at the same screen resolution that it was originally recorded.
return to top

 

 

How do I view a .out file or a .ord file?

There are a couple of ways to do this. The files are simply text files with different extensions, so they can be opened in any normal text editor (Notepad, WordPad, Microsoft Word, etc).

The first way is to simply rename the file. Instead of the file being GazeTrackerOutput.out, rename it to GazeTrackerOutput.txt. Now, when you double-click the file it will automatically open in your default text editor (most likely Notepad).

The second way is to tell Windows how to open the file. When you double-click a .ord file, you’ll get a dialog like this:
This images shows the Select a program from a list choosen on the Windows Open Dialog

Choose Select a program from a list of installed programs and then click OK. You’ll then see a dialog that lists the programs you can use to open the file:

Open With Dialog with notepad chosen

Choose Notepad from the list of files and click OK. If you always want to use Notepad to open these files, make sure to check Always use the selected program to open this kind of file before clicking OK.
return to top

 

 

How do I access additional options in GazeTracker?

GazeTracker attempts to put the most commonly modified options directly into the user interface so that you can find them easily. In some cases, however, there are additional options available in other options dialogs. Those options can be accessed by clicking the dialog box launcher buttons that you can see highlighted here:

Highlight of the Advanced Options button on the ribbon

Wherever you see one of those dialog box launcher buttons, there are more options and settings available to be configured.
return to top

 

 

Why is my GazeTrail so long on video files?

By default GazeTracker enables Trail Truncation for video slides.However, if the user disables it GazeTracker will remember that change and not automatically re-enable it. To enable Trail Truncation, go to the Options tab and click on Enable Trail Truncation. The default setting for Start delta is 2 seconds and the default setting for End delta is 0.02 seconds. This will show 2 seconds of gaze data before the current video time (into the past) and 0.02 seconds of time beyond the current video time (into the future).
This image shows an untruncated gazepath over the first frame of a video return to top

 

 

Why is GazeTracker reporting inaccurate data?

There are a few settings to check in GazeTracker if your data is inaccurate:

  • Verify the screen resolution on the analysis computer is the same as it was on the recording computer. GazeTracker will warn you if you try to view data that was recorded at a different screen resolution.
  • Make sure the Gap Interval is correct. This option is listed in the GazeTrail Options and is accessible by clicking the dialog box launcher (This button will take you to the advanced options) in the GazeTrail section of the Options tab. The default of .035 should be relevant for 30 and 60 Hz eye trackers. If your eye tracker records at a different speed, or for more information, see the “What is the Gap Interval setting and when should I change it?” FAQ.
  • Is Filtering enabled in GazeTracker? Filtering in GazeTracker changes the data and should only be used under certain circumstances. We recommend leaving it off unless you have been directed otherwise by technical support. Filtering can be found under the Analysis tab on the ribbon. For more information on Filtering, see the FAQ “When should I use filtering in GazeTracker?”On Off Data Filtering from the ribbon
  • Are there any data shifts? You can check this under the Other group on the Options tab and clicking the Shift data button from the ribbon button. This will show the Data Shifts dialog; if the shifts list does not say “No Data Shifts!”, then the data has been shifted through GazeTracker. These shifts can be removed by selecting the shift and clicking Remove.
  • Verify, using the Test Connection mechanism on the eye-tracker tab, that communication has been properly established between the eye-tracking system and GazeTracker. If the Test Connection dialog reports connection failures, refer to the eye-tracker-specific appendix of the GazeTracker manual to verify the setup procedures.

If you have checked these settings in the GazeTracker software and your data is still inaccurate then the most likely problem is that the data that is coming out of your eye tracking hardware is inaccurate. You may need to perform a new calibration or other diagnostic procedures. Refer to your eye-tracker’s manual or contact the manufacturer for more assistance.
return to top

 

 

Why is GazeTracker not showing my data at the advertised Hz of my eye tracker?

The advertised speed of an eye tracker is the maximum possible speed for the eye tracker, and GazeTracker only records valid data from your eye-tracker.  A number of factors including identifying the eye and the computer’s processing power can affect the actual speed of the eye tracker. GazeTracker does not adjust the data that it is receiving, so if you are not receiving the required speed out of the eye tracker, the first step should be working with the eye tracker manufacturer to ensuring that the eye tracker is really sending data at the desired speed.
return to top

 

 

Change what LookZone Collections are active by default?

Setting which LookZones are active at the time a stimulus is shown is very simple. Click on the button in the Visibility group on the LookZones tab. This will open the Choose Active Collections dialog.  LookZone Collections that are highlighted in blue will be active as soon as the slide is shown.  Any other LookZones will need to be made active during the slide through the conditional LookZone options.
return to top

 

 

What is the difference between Play and Play Single OR Rewind and Rewind Single?

The Play and Rewind buttons operate across the entire experiment. So, for instance, when you press the Play button, GazeTracker will play back the entire experiment and the recorded data for every stimulus in the experiment. The Play Single and Rewind Single buttons operate just on the currently selected stimulus in the experiment.
return to top

 

 

How do I see the rest of my data in the timeline?

By default the Timeline only shows 10 seconds worth of data. You can scroll which 10 seconds are shown by clicking one of the two arrows.
An example of the timeline with back and forward buttons highlighted

You can also change what is shown at one time in the Timeline Options. You can access the timeline options by clicking the highlighted timeline buttonbutton in the Other group on the Options tab.
return to top

 

 

Why is the webpage different during analysis than during recording?

During a web recording session GazeTracker logs the URL for the individual webpages that the subject visits. When analyzing a recording GazeTraker pulls the current website from that URL and lays the gaze data over top of it. This can cause consistency problems if the website changes between when you did the original recording and when you go back to analyze the data.

If this is a problem for your experiment we recommend one of two approaches. You can save the website you wish to analyze to the local machine (most major web browsers support this capability), and then point the web browser at that local copy during the experiment. In this case GazeTracker will record the URL for the local copy and use that during analysis, ensuring that it won’t change. For this to work, you would need to save each page you want to analyze locally and adjust the internal linking of the page. If saving the page(s) to the local machine is not realistic, you can also turn screen recording on for the experiment. This will create a video of what the user saw during recording and allow you to use GazeTracker’s video analysis tools on the screen recording to analyze the data.
return to top

 

 

Why is the ordering section of the Per-Slide Wizard disabled?

The Per-Slide Wizard is accessible by clicking the Customize… button on the Per-Slide Display Options pane of the Slide Show Wizard dialog. Ordering on the Per-Slide Options is used to guarantee that a particular slide follows, or comes before, another slide. By default the ordering is specified directly on the Image Slides (or Video Slides) pane, which means that this ordering is only used if you check the Randomize slide presentation order on that pane. The Per-Slide Ordering is then only used for guaranteeing sub ordering. For instance, you may always want an instructions slide to precede the true stimulus slide, but you want the stimulus slides to appear in a random order. To do this, you would check Randomize slide presentation order on the Image Slides pane and then set the instructions as the Slide to show directly before the stimulus slide on the Ordering pane of the Per-Slide Wizard.
return to top

 

 

When should I create a new database?

GazeTracker uses Microsoft Access to store recordings. Access begins to run into scaling problems that can cause errors during recording at around 1GB of data. Due to the way Access handles caching and a variety of other things, there is no true number of experiments, recordings or GazePoints that equate to a 1 GB file, but a good rule of thumb is that you should create a new database at around 5 million gaze points. Using the Hz or frame rate of your eye tracker, you can calculate how long you can record before you should create a new database. For instance, a 60Hz eye tracker creates 60 gaze point rows for every second of recording. This means that you can record for ~1389 minutes before you hit the 5 million records mark and need to set up a new database. We recommend looking at how much data you are going to be recording and creating a new database on a set schedule. For instance, create a new database at the end of every week or at the end of every month depending on how often you are doing recordings and how long those recordings are. GazeTracker allows analysis across multiple databases, so having multiple databases will not normally cause a problem.

Creating new database files on a regular basis before they get too large should preempt any database errors, but if you do run into database errors when GazeTracker is saving a recording you should immediately create a new database file for future recordings.
return to top

 

 

When should I use the Compact & Repair feature?

Access does a large amount of caching of data which can cause the database to grow and possibly cause errors. Compact and Repair cleans the data and clears out any cached data. If you are getting errors that do not tell you that the database is too large, we recommend first trying Compact and Repair. If you continue to get errors, we recommend that you then create a new database.

Deleting experiments and/or recordings creates an especially large amount of cached data. Because of this, we also recommend using the Compact and Repair feature after you have deleted a large amount of data.
return to top

 

 

Why/when should I split a database?

Splitting a database is completely a data management feature. It is used to split certain experiments and recordings into their own database while optionally leaving them in the original database. This command is primarily used to share a subset of your data with a third party, be it a fellow researcher or GazeTracker support.
return to top

 

 

How do I retrieve the order in which a randomized experiment was presented?

When you have an experiment set up to present the stimulus slide in a random order you will need a way to determine the order in which the slides were presented. To do this, first load the recording you need the order from and then click the Save Order… button at the bottom of the Properties pane. This will save an .ORD file with the order in which the stimuli were displayed.
return to top

 

 

How do LookZones work in Application Analysis?

LookZones are added to applications using the same techniques as in Image or Video Analysis. They are tied to their location in the application on which they were drawn and will move with that application. They will not scale, and with the notable exception of Internet Explorer, they will not move if the subject scrolls or otherwise changes the data under the LookZone.
return to top

 

 

What are intro and buffer slides used for?

Intro and Buffer Slides are defined on the Intro/Buffer Slides step on the Slide Show Wizard, which walks you through creating an image experiment. There are three primary reasons a researcher would setup intro or buffer slides. The first is when the research is focused on the pupil size as an important metric. Setting a black and then white slide at the beginning of the experiment is often used to determine the natural range over which a subject’s pupil size will vary. Setting a solid color slide in between each stimulus will allow the subjects pupil to normalize in between each stimulus. The second reason a researcher might use buffer slides is to normalize the area at which the subject is looking between slides. For instance if there is an area on one slide that draws the majority of users attention and the experiment advances to the next slide without a buffer, the majority of subjects will still be looking in that area when the next slide appears, which may skew the data for the next slide. Inserting a buffer gives the subject a moment to look away from the area of interest on the last slide, and you will not have the majority of subjects looking in the same area at the beginning of the next slide. The final use is to simply give instructions between each slide. This is only useful if you want to give the same instructions between each and every slide. Otherwise, the instructions will need to be a slide themselves.
return to top

 

 

What are conditional LookZones used for

Conditional LookZones are used to activate and deactivate LookZone Collections. Conditional LookZones are used primarily in two scenarios.

The first and most common is that you are eye tracking over a piece of software that changes its content while not changing the parent window. GazeTracker only creates a new stimulus slide in application analysis when it seems a new parent window, or in the case of Internet Explorer, when the browser loads a new web page. Using conditional LookZones, you can place a LookZone over a menu button that changes the windows content, and set it up so that if the user clicks on that button and therefor the LookZone, it will change the active LookZone collection to match the new content of the window.

The second scenario would be in a case where the researcher does not care about certain LookZone statistics until the participant looks at or clicks something in particular. A classic example of this would be that the researcher doesn’t want to start collecting LookZone statistics until the participant has read the instructions. In this case, the researcher can setup a LookZone around the instructions that only activates another LookZone collection once the participant has looked at the instructions for a set amount of time.

See the “Utilize LookZone Collections” tutorial for a step by step tutorial on setting up a conditional LookZone.
return to top

 

 

 

 

Troubleshooting

“MS Access Error: Database has reached maximum size” error message

GazeTracker uses Microsoft Access to store all experiment and gaze information. Microsoft Access has a file size limit of 2 gigabytes. If your database reaches this limit you will get this error. If you see this error, we recommend that you use the database utilities to create a new database. To help prevent this error, we recommend creating a new database when your current database reaches 1 GB in size.  It is a good idea to set schedule for when to create a new database depending on the amount of data you are recording.
return to top

 

 

ADODB or other database errors

The first thing to do when you encounter a database error is to try to Compact and Repair your database. This will often solve any problems you are running into. If you continue to encounter database errors you should create a new database.
return to top

 

 

Cannot display data error message when trying to load a recording in Application/Web Experiments

This is generally caused because Internet Explorer or other applications that GazeTracker is attempting to interact with are not responding. We recommend closing those programs and clicking the Refresh button from the ribbon button in the Playback And Recording group on the Home tab of the ribbon.
return to top

 

 

Cannot display image/video error message when trying to load a recording

This is generally caused because the stimulus file has been moved. To correct this, right click on the stimulus slide in the Properties pane and choose Redirect; then point to the correct file.
return to top

 

 

Why are the LookZones in an incorrect location when I load an experiment

This is generally caused when the LookZones were added at a different screen resolution than the one at which they are currently being viewed. By default GazeTracker warns you when you load a recording at a different resolution, but this warning can be disabled.
return to top

 

 

Excel Graph is blank

On occasion the Excel graph automation does not complete, and the researcher is left with a blank graph. Correcting this is easy to do.

1. Right click on the graph and choose Excel Select Data button

This will bring up the Select Data Source dialog that allows you to define the features of the graph.

2. Click on the Chart data range: button

This will bring up the Select Data Source dialog that allows you to select the table to turn into a graph.
Excel Dialog for selecting the data source for a graph

3. Make sure you are on Sheet1

4. Highlight the table, if it is already selected, remove the selection and reapply it

5. Click the button on the Select Data Source dialog

You should now see the correct data in the Select Data Source Dialog. If not, repeat steps 3 through 6

6. Click OK
return to top

 

 

Invalid CD Key or Out of Licenses Error when activating GazeTracker

GazeTracker is activated to the computer it is installed on and needs to be reactivated if moved to a new computer. We recommend that, before you move GazeTracker or upgrade the computer it is on, you contact your GazeTracker distributor to ensure you will be able to reactivate GazeTracker. If you run into this message during activation, please contact your distributor for a resolution.
return to top

 

 

‘Unspecified Error’ when attempting to activate GazeTracker

The activation server was changed in January 2010. If you purchased a copy of GazeTracker before that time, and are trying to activate it, GazeTracker will be activating against the old server. Eyetellect has released a patch that solves this problem. Download the patch (GazeTracker.exe.config), from here, and place it in the same folder on your computer as the GazeTracker executable. Once you have placed the patch in the correct location and restarted GazeTracker, the software will use the correct activation server.
return to top

 

 

Your eye tracker is not working or will not communication with GazeTracker

Each eye tracker has its own way of communicating with GazeTracker. The first thing to do is check the settings under the eye tracker tab on the ribbon. Make sure the communication is set up correctly. If that does not solve the problem, the next step would be to call the eye tracker manufacturer for additional assistance.
return to top