Python crashes

claudie

Hi everyone,
I am facing an issue while running my 2 block experiment on OpenSesame, it crashed in between the two blocks and I get the following message:

Starting experiment as ExperimentProcess-1
Expyriment 0.8.1.opensesame2 (Python 2.7.12) 
openexp.sampler._legacy.init_sound(): sampling freq = 48000, buffer size = 1024
openexp.sampler._legacy.init_sound(): mixer already initialized, closing
experiment.run(): experiment started at Fri Mar 17 10:53:25 2017
experiment.run(): disabling garbage collection

Python seems to have crashed. This should not happen. If Python crashes often, please report it on the OpenSesame forum.

item-stack:
time: Fri Mar 17 11:03:28 2017

I have tried running the second block first and vice versa, the experiment keeps crashing between the blocks regardless.

It worked fine before and I have to start running participants next week so any help would be greatly appreciated !

Many many thanks,
Claudie

sebastiaan

Hi Claudie,

The first thing to try (if you haven't already) is to restart your computer. If that doesn't fix it, could you please provide more information?

• What kind of operating system are you using?
• What version of OpenSesame?
• Does this always happen at the exact same moment?

Cheers,
Sebastiaan

claudie

Hi Sebastiaan,

Thanks for your reply.

I have indeed tried restarting my computer several times, and the problem is still there.

I am using a Dell Intel Core desktop and windows 7.
Version of OpenSesame is 3.1.4
And yes- it does always happen in the transition between my two experimental blocks (no matter which one comes first)

Thank you so much for your help,
Claudie

sebastiaan

Hi Claudie,

In that case, it would be best if you attach the script to this discussion, so I can see if there's an obvious explanation for why this might happen.

Cheers,
Sebastiaan

claudie

Apologies for the long post, I wasn't sure which script you'd want to look at so I thought I'd paste the entire general script.

Thanks for your precious help!

sebastiaan

Hi Claudie,

You can attach the experiment file using the attachment icon on this forum. Could you attach your experiment file that way? That's a bit easier (I deleted the posts with the script text.)

Cheers,
Sebastiaan

claudie

Oops. Sorry about that, please find the experiment file attached

sebastiaan

Hi Claudie,

This is really nothing more than a guess, but it might the videos. If you remove all media_player_mpy items from the experiment, is the crashing gone?

Cheers,
Sebastiaan

claudie

Hi Sebastiaan,

Yes , it works fine when all videos are removed. Problem is the videos are the core of my experiment.Do you think it has to do with the video format?

Thanks!

sebastiaan

My first thought is that your system may run out of memory. Under Windows you can monitor usage through the Task Manager. If you open the Task Manager while your experiment is running (with videos), do you see memory filling up (and if so, how much)?

claudie

Hi Sebastiaan,

I just tried running the experiment with the Task Manager opened and I the CPU usage goes up to about 25% whereas the Memory doesn't seem to me moving up much. However when I look at the physical memory box, I see that the free amount of MG goes from 489 to under 100 when the experiment is running. Is that the problem?

sebastiaan

I see that the free amount of MG goes from 489 to under 100 when the experiment is running. Is that the problem?

Possibly, yes. Does it go all the way to 0 Mb before crashing?

@Daniel, do you have some ideas? This experiment is quite video heavy, and I suspect that it crashes because it runs out of memory.

Daniel

I don't really have a clue. I did my best to let the player clean up when it can. Playing videos is a heavy operation in the plugins current form, as the frame loading is not hardware accelerated (rendering of the video frames is is). Players like vlc have a way to decode and render frames all in video memory, but currently media_player_mpy decodes the videos in normal RAM and then passes them to the video memory. I have searched endlessly, but I didn't find a way to also perform the frame decoding in video memory.

Considering this experiment heavily depends on videos according to Claudie, it may be that too many videos are loaded at the same time in the prepare phase? @claudie Maybe you can slowly increment the number of videos you have in your experiment and see after it crashes after a certain point? Then you know for sure the amount of videos is the problem.

claudie

Hi Daniel and Sebastiaan,

Thank you both for your help. I'll have a try at decreasing the number of videos, however the one thing I find a bit strange is that the experiment has worked just fine in the past with the same number of videos (and the videos were even longer than they are now), could it be that my desktop has now less free memory than before?

sebastiaan

however the one thing I find a bit strange is that the experiment has worked just fine in the past with the same number of videos (and the videos were even longer than they are now), could it be that my desktop has now less free memory than before?

Were these videos of a different format? Or perhaps the same format but a lower resolution?

claudie

No - my videos are all the same (created with Blender software)

Eris

Hi Caludie, I am having the same problem with presenting imagines.

Have you solved your problem?

Thanks