Adapting coroutines for OSWeb

Comments

• sebastiaan

  April 2023

  Hi @sflu ,

  Depending on how much flexibility you need, the solution may actually be very simple and not require coroutines at all.

  • Let's say that you create a single audio file that contains both the audio stimulus and (after 3000 ms) the timeout warning. You can create this file with an audio editor such as Audacity.
  • Then you play this sound file with a sampler (audio_stimulus) and set the duration to 400, which is the period during you (based on the screenshot) don't want to collect a key press yet.
  • Then insert a keyboard_response with a timeout of 3600, which together with the 400 ms duration of the audio_stimulus adds up to 4000.
  • Then, after the keyboard_response , stop the audio_stimulus from playing (if it was still playing) with a script like the one below. This will avoid the timeout from playing if the participant responded in time.

  // This should be in the run phase of an `inline_javascript` item
  let audio = exp.items._items['audio_stimulus']._sampler.sample
  if (!audio.paused) {
    audio.pause()
  }
  

  Do you see the logic? I attached an example experiment. (I made it in OpenSesame 4.0, but you should be able to run it also in 3.3.)

  sound-example.osexp

  — Sebastiaan

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

  April 2023

  Hi Sebastian,

  Thank you for your response! If I set the duration of the stimulus to 400 ms, the file will not finish playing if it's longer than 400 ms, though, right? We have a list of about 1000 nouns that participants will go through, and there is some variation in length.

  Also, when I tried to run your example I got this error: ModuleNotFoundError: No module named 'openexp._canvas.osweb' Any idea what I need to do?

  Thanks again!
• sebastiaan

  April 2023

  @sflu Right, it was set to use the OSWeb backend, which is new in 4.0. (Previously, OSWeb existed, but it integrated with the interface differently.) Here's the updated example. Does this run for you?

  sound-example.osexp

  You can also consider upgrading to OpenSesame 4.0. It's still in pre-release, but it's already quite stable at this point and has some nice new features.

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

  April 2023

  Hi! Thanks, I'll check out the latest version. I've actually gone into a different direction now, and just put a sketchpad telling participants that they've timed out (at 3000 ms) instead of playing an audio file. I've also added self-timed breaks after every 5 trials. The issue I'm having now is that the feedback_item doesn't seem to present accuracy and response times correctly. For example, even if you make a mistake, you might get 100 % accuracy. Or if you time out a few times, the average response time is 3000 ms, even if you've responded more quickly. Do you have any idea why that might be? I've attached a screenshot of the trial_sequence as well as of what I have in the feedback form. Thanks for all your help!

  

  

  
  
• Fab

  April 2023

  Hi @sflu,

  I just saw your post (when replying to a message, it is best to user handles: @sebastiaan so that your correpondent receives a notification).

  Yes, the sound will carry on beyond the 400ms of the sampler.

  I believe that the error you mentioned may possibly be due to using an version of Open Sesame older than that Sebastiaan used to create the example (Which is version 4, now available for a test drive on the website: https://osdoc.cogsci.nl/3.3/download/).

  Best,

  Fabrice.

  
• sflu

  April 2023

  Hi @Fab,

  Thanks for your message! I haven't upgraded to version 4 yet, but it seems the accuracy and response time problem only started to happen after I added a timeout warning and a self-timed break (as in the screenshot). Before that everything seemed to be fine. I didn't actually end up using Sebastiaan's example, instead of having an auditory timeout I'm presenting it on a sketchpad.

  But I'll download version 4 and run it to check if that might fix the error. Do you think I'll need to adapt anything in my experiment or should the parts that are working now continue to run ok?

  Thanks!
• Fab

  April 2023 edited April 2023

  Hi @sflu,

  From your screen shot it appears that you're using a Python inline_script, but Python is not supported in OSWeb. You need to use Javascript instead.

  Youo can modify your program without upgrading to version yet.

  Fabrice.

  
• sflu

  April 2023

  Hi @Fab,

  Thanks, I got rid of the Python script and changed the Run if statement instead. I also unchecked the Reset feedback variables box for the per trial feedback items in the sequence. It seems to be logging accuracy and average rt correctly now.

  Thanks!