[open] Error when trying to run exp. using images

Comments

• elizabeth

  May 2013 edited 11:45AM

  It is still happening on the usual PC. I copied and pasted the script into a new blank experiment and then added the pictures to the file pool and I am still getting the same error (although with a different directory path on this PC).

  This is really bad timing as the scanner is scheduled today. Any help would be great.
• sebastiaan

  May 2013 edited 11:45AM

  Hi Elizabeth,

  It looks as though an empty string is specified for a filename. As you may notice /var/folders/yx/glwd9x2j51n04czh33_230y00000gn/T/tmpyyNLLj.opensesame_pool/ is just the filepool folder itself (which is always in some temporary folder), and not a file in the filepool.

  You can get this error when you specify an image like so:

  draw image 512 384 "[my_var]" scale=1.0 center=1 show_if="always"
  

  and then have my_var be an empty string, for example by setting it like this:

  exp.set('my_var', '')
  

  Could something along those lines be going on?

  Cheers,
  Sebastiaan

  
  
  Check out SigmundAI.eu for our OpenSesame AI assistant!
  
• elizabeth

  May 2013 edited May 2013

  Hi Sebastiaan,

  Yes, this could be it. There are trials where no images are shown, although the sequence for those trials doesn't include an image. Perhaps it is easier if I e-mail the experiment, but here are screen shots of the set-up. Basically, there is one loop (objects_saccades) that loops through trials showing either a series of images (uses the obj_seq sequence), a series of fixation targets (saccade_seq) or a rest period (ISI).

  Here are the two screen shots, but let me know if something else would be more helpful.

  http://img.cogsci.nl/?q=5190b9fd87a69
  http://img.cogsci.nl/?q=5190ba1296bf9

  This may not be the best set-up anyway. I ran the pilot, although without this part of this task, and had some timing issues where the experiment took longer than it should have and so the scanning finished and the task was still going, which means the onsets and duration of events are not in the expected places...I will post a separate post about this to keep things organized, but maybe this issue will be solved by organizing things differently for timing accuracy anyway.

  Thank you so much!
• sebastiaan

  May 2013 edited 11:45AM

  Hi Elizabeth,

  I think the problem is indeed that on some trials the image files are left blank (obj1, obj2, and obj3). Even though this may not seem like a problem, because no image will be shown on those trials anyway, it actually is: When an item is skipped through a 'Run if' statement, it is still prepared, and the prepare phase requires a valid image file. Does that make sense?

  For more information, see here:

  • http://osdoc.cogsci.nl/usage/prepare-run/#conditional-statements

  The easiest solution is to use some dummy image (dummy.BMP, say) where you now leave the image blank.

  Hope this helps!

  Cheers,
  Sebastiaan

  
  
  Check out SigmundAI.eu for our OpenSesame AI assistant!