[solved] trying on microsoft surface 3 at 1900x1200

Comments

• elaci0

  May 2015 edited 1:43AM

  little update: if I run my script with opensesamrun.exe in full screen mode seems working fine.
  
  is that adviceable or should I run it at a fixed resolution setting?
  

  thanks
• Josh

  May 2015 edited 1:43AM

  Hi,

  I'm not aware of any drawbacks occuring when you alter the resolution, as long as your computer can handle it. Of course when your experiment will start to become slower (with regard to displaying images etc) that's a clear signal you should go to a lower resolution (although sometimes you just need to write more efficient code!). You can play with the resolution settings when you click the "new experiment" tab in the overview window of your experiment.

  Cheers,

  Josh
• elaci0

  May 2015 edited 1:43AM

  Hi Josh,
  
  thanks for your kind answer. Because my experiment depends a lot on icon positioning I cannot play too much with the resolution, so my question was more about if runnig opensesamerun.exe has some side-effects or it's the same.
  
  thanks!
• elaci0

  May 2015 edited 1:43AM

  holy sh*t..
  
  I thought that running opensesamerun.exe within a batch file could have saved me from not recognizing the 1920x1200 resolution, but I was wrong
  
  from its GUI opensesamerun works perfectly
  
  from inside a batch it does not and switchs to 1200x800 (like openesesame.exe)
  
  how can I solve this? All the experiment is based for this higher resolution
  

  thanks
• eduard

  May 2015 edited 1:43AM

  Hi,

  If the problem doesn't occur in window mode, you could use this hack by Sebastiaan (see his last response).

  Good luck,

  Eduard

  
• elaci0

  May 2015 edited 1:43AM

  Hmmm.. I don't know if is the correct behavior but in this case, when I run the experiment windowed I see a window bigger twice my resolution (or at least it doesn't fit the screen)
  
  Following your link and the Sebastian's hack the experiment run at twice resolution too, in the meaning I need to scroll the window for reading other parts of it.

  Do you maybe know why running. Opensesamerun from a batch it doesn't run full screen correctly?
  

  Thanks
• eduard

  June 2015 edited 1:43AM

  Hi,

  Unfortunately, I have no idea, why Opensesamerun doesn't work.
  Sorry for that.

  Eduard

  
• sebastiaan

  June 2015 edited 1:43AM

  @elaci0 filed an issue for this on GitHub: #328. (I also have no idea what the problem might be.)

  
  
  Check out SigmundAI.eu for our OpenSesame AI assistant!
  
• elaci0

  June 2015 edited 1:43AM

  Little update:
  
  I'm trying to debug but I have no clue where to look exactly, so after running 'opensesamerun' from the sources with a debugger:
  

  • without any options the GUI pops up, I fill the blanks and everything works
  • passing ALL the required arguments before running (such as a dos batch) the GUI does NOT pop up and the screen is "enlarged"
    
  • passing SOME arguments the GUI pops up, I fill the field and everything is again OK
    
  • 'embedding' all the required options inside the script does not work either, the GUI does not pop up and the screen size is wrong
    
  • 'embedding' all the options AND modifying the script so that the GUI popups in any case and immediately close itself DOES WORK.
    
    
    
    my thoughts are that somehow the QT dialog box influence the rest of the script. How it's a mistery. I debugged "line-by-line" to check if the correct screen size is passed between all the python scripts and the answer is yes.
    
    One strange thing I noticed is that running a "not working" mode provide the usual wrong screen (somehow enlarged) but the "preview" icon in the taskbar (the one that shows the window's content) show my experiments correctly. But only the preview.
    
    
    
    In the meanwhile, because I am in a rush for this experiment, I will run this modified 'opensesamerun' script.
    

  thanks for your help
• elaci0

  June 2015 edited 1:43AM

  Ok, Sebastian seems to have found the solution:
  
  going to Control Panel--> display --> "let me choose ONE scaling level for all my displays (with Larger 125% -default)" seems to solve the problem: all the applications will run at the highest resolution (1920x1200) this means that also running the my experiments from Opensesame's GUI will work as expected.
  
  thanks!
  
• Daniel

  June 2015 edited June 2015

  Sounds like a display scaling problem again. A lot of funny things like these also occur on OSX with retina displays... I guess this has to be solved with newer QT libraries and more modern drawing libraries that have better support for scaling. Maybe QT5 will take scaling into account?