[solved] ERP-experiment slows down in fullscreen mode

Comments

• Josh

  September 2015 edited 12:12PM

  Hi Johanna,

  There's always the possibility that your experiment could be programmed to run faster. Did you create your stimulus displays beforehand? (so that you only have to show them during the trial with canvas.show()). In general experiments should run perfectly fast as long as they're programmed efficiently. If you use inline_scripts, try to do as much as possible in the prepare phase, and keep the run phase minimalistic. More details are required before I'd recommend you to buy a new computer, for sure.

  Cheers,

  Josh
• Johanna

  September 2015 edited 12:12PM

  Hi Josh,

  Buying a new computer indeed seems a bit radical! :-)

  I think it would be a problem to create the stimulus displays beforehand in this experiment, because stimulus presentation is (partly) dependent on the participant's response. Therefore stimuli are presented in a feedback item, although I know this is not really optimal.

  In each trial, participants are presented two visual stimuli (circles) simultaneously (created in one .png image); one left, one right. Both circles are divided in eight equal slices. Participants then choose one of these stimuli (by pressing a key), after which a rotating 'pizzaslice' is superimposed on the chosen stimulus. For each trial, a random number generated in that trial indicates the number of circle-slices the pizzaslice will cross (which is the length of the rotation).

  It is this rotation of the pizzaslice that is running too slow. It is run by an eight-step sequence, in which I set the duration of each step on 75 ms, but in fullscreen it seems to be limited on 150 ms; it will not get faster.

  I wouldn't know how to prepare my stimuli in advance, as so much is dependent on what happens within each trial. Maybe there are other options to try out..?

  Johanna
• Josh

  September 2015 edited 12:12PM

  Hi Johanna,

  How do you create (draw) the superimposed pizza slice? Do you have 8 sketchpads in a sequence item, or did you with 'sequence' refer to e.g. a for-loop?

  As far as I can see, you would need to create 16 canvases beforehand (inline_script); 8 containing slices superimposed on the left circle, and 8 containing slices superimposed on the right one. After the keyboard_response item you can just place an inline_script with a for-loop that displays 8 of these canvases depending on what participants responded. It could look something like this:

     import random
  
     if keyboard_response == 'z':
     octant = 1 # will indicate which canvas needs to be displayed
            for i in range (0, randint(0,20)): #here 20 is the max amount of steps
                   if octant == 1: 
                         exp.canvas1.show()
                         self.sleep(75)
                         octant += 1
                   elif octant == 2: 
                         exp.canvas2.show()
                         self.sleep(75)
                         octant += 1
                   elif octant == 3: 
                         exp.canvas3.show()
                         self.sleep(75)
                         octant += 1
                   elif octant == 4: 
                         exp.canvas4.show()
                         self.sleep(75)
                         octant += 1
                   elif octant == 5: 
                         exp.canvas5.show()
                         self.sleep(75)
                         octant += 1
                   elif octant == 6: 
                         exp.canvas6.show()
                         self.sleep(75)
                         octant += 1
                   elif octant == 7: 
                         exp.canvas7.show()
                         self.sleep(75)
                         octant += 1
                   elif octant == 8: 
                         exp.canvas1.show()
                         self.sleep(75)
                         octant = 1         # important, octant becomes 1 again!
  

  And a similar for-loop for 'if keyboard_response == '/'' (or whatever buttons you use).

  Good luck!

  Josh
• Johanna

  September 2015 edited 12:12PM

  Hi Josh,

  I was just about to write a reply with extra info and therefore checking the experiment again, and mysteriously it suddenly works well (at least on my desktop computer). If the problem arises again, I will get back with more detailed info!

  Thanks for your help anyway!
  Johanna
• Josh

  September 2015 edited 12:12PM

  Ok! And yes - let me know if any problem occurs again.

  Josh