Audio static then crashing

lmgetzphd

OpenSesame version

3.3.14

Operating system

Windows 10 Enterprise

Backend

Pygame (legacy)

Expected behavior

Trial sequence: fixation, audio prime, fixation, audio target, response screen until button box response (or moves on automatically if no reply in 3000 ms)

Four blocks, with inline_script to rest every 14 trials

Actual behavior (what goes wrong)

On random trials, audio files get static-y and then stops playing sounds. Trial sequence gets stuck and no buttons will move the trial alone. Even using esc to stop the run will not fix the issue on run restart. Exiting open sesame and restarting also does not work. Only restarting whole computer lets experiment run begin again, but then same issues happens with different sound file. I don't have an Open Sesame or Python error message because it just says "experiment aborted" or "open sesame crashed unexpectedly" depending on how I exit the run after the issues occurs.

What did you try already?

It doesn't seems to be an issue with the file pool, but I trashed and reinstalled all sound files. It is not the same sound every time in crashes, and not the same trial number.

Any ideas of what might be happening would be appreciated. It have been running this experiment for 3 semesters without issue and now it will not get through the four blocks of the experiment. The inline_script no longer works with OS v 4.0.0 so I am trying to resolve the issue without updating Open Sesame.

cvanbuck

Hi @lmgetzphd

A couple of things you could try:

• First, double-check that all your audio files are in the exact same format (sampling rate, bit depth, and number of channels; e.g., with Audacity). Even small inconsistencies can sometimes cause the sampler to crash.
• Crashes like the one you describe can also come from bugs in the underlying libraries. If possible, you could try switching to the PsychoPy backend to see if that resolves the issue.

Let us know if that helps,

Claire

P.S. About your inline script: depending on what it does/how it is used, it should be possible to adapt it for OpenSesame 4 (which is generally recommended). If you’d like, you can share your experiment file so that we can take a closer look at it.

lmgetzphd

Thanks for the suggestions @cvanbuck !

We have been running the same experiment for a while, and the problem just started occuring this week, so I don't think it is the audio files, but I still double checked and all seems well.

I changed the PsychoPy backend and the same issue remains.

One issue I noticed that I'm not sure if it's related is when I open the program, it says "could not parse retrieved OSF settings" "Using cached OSF settings instead" (see screenshot image).

It just doesn't make sense that this issue would start now when I have been running participants with this same setup for months.

I'm happy to share my experiment file, but my inline script is using item.set() which has been removed in Open Sesame 4 so I would need to determine how to count trials and insert a break screen after every 14 trials with the new version of Open Sesame. I have not searched for help doing so yet as I was hoping to run all participants using the same version of Open Sesame.

cvanbuck

Hi @lmgetzphd

Thanks for the extra details.

Since the experiment and audio files haven’t changed + changing the backend didn't help + consistent audiofile formats + the issue only started recently, it does sound like it could be related to your device system (e.g. drivers, Windows update, or audio output). But I am not sure, so I'll do my best to help.

A few things you could try next to continue troubleshooting:

• Check whether your system or audio drivers were recently updated. If that's the case, and if possible, roll back to a previous version.
• If you’re currently using speakers, try switching to headphones (or vice versa) to see if the issue is tied to the audio output of your device.
• You could also try disabling the sampler items in your experiment, to check if the issue is tied to audio playback; if it's still crashing, then it might be something else.
• Try running the experiment on a different computer with the same Opensesame version, or, indeed, share your experiment file here so I (or anyone else) can test if the crash is reproducible.

Let us know how it goes!

Claire