Issue playing media with PsychoPy backend
Hi there,
We are trying to play a video in our experiment using the opensesame-plugin-media_player_mpy plugin (software: Mac OS Catalina version 10.15, latest Opensesame version 3.2). Although the video plays fine with a Legacy (pygame) backend, we can't seem to get it working with the PsychoPy backend. We would love to just stick with Legacy, but unfortunately there were issues with getting that working with our Tobii eyetracker, so we had to switch. When we run the experiment, a grey box appears where the video should be placed, but with no video or sound.
We can see that the PsychoPy handler does most of the legwork in the process_frame() function in media_player_mpy/handlers/psychopy_handler.py, with a note to say that this is required to overcome some weird flicker that would occur if the structure were the same as that in the handlers for the other backends. Since the psychopy handler imports the OpenGL renderer, it seems like the issues might be stemming from there, e.g. perhaps in the OpenGL process_frame() or draw_frame() functions.
Have you ever had any issues with playing media with a PsychoPy backend on a MacOS before? Could you give some advice on how we might get this working?
Best wishes,
Bronagh
@Daniel
Comments
Hi Bronagh,
I must say that I have grossly neglected maintaining the media_player_mpy and underlying python-mediadecoder packages. There are some other open issues at https://github.com/dschreij/mediadecoder that are still on my radar, and I will add this one.
I haven't experienced the problem as you describe it, but then again I haven't used the mediaplayer myself in while, nor have I upgraded to MacOS Catalina. I do know that Apple is dropping support for OpenGL in favor of its own system 'Metal', and that may have something to do with it, but if other parts of psychopy do function correctly, I doubt this is the reason.
In any case, I need to take a raincheck on this one. I simply haven't found the time to look at media_player_mpy and python-mediadecoder and will in addition need some time to get myself familiarized with my own codebase again, since it has been a while since I worked on them. That said, the intention really is there, I just don't know when I'll get to it, but hopefully soon in the more quiet period that is about to arrive for me.
Sorry for not being of more help.
Hi Daniel,
Thanks for getting back on this and giving a realistic timeline of when you can look into it. We will check out the other routes then of either trying to run the experiment with different Mac/Windows versions, or perhaps even seeing if we can use a different eyetracker that is compatible with the Legacy backend.
Cheers,
Bronagh