Hello Sam,
I also had problem converting your videos with vlc. It only converted two of your example videos. To test whether your PsychoPy-setup copy the experiment to another computer. Given that your test-experiment ran on two different computers, there is a chance that something is wrong.
Best wishes Jens
Hi Jens,
Sadly, I’ve tried it on my laptop as opposed to desktop computer and the frame dropping is even worse.
This could be a very naive question, but do you reckon there is any way I could possibly use the coder or a code chunk to prevent PsychoPy from ending the experiment after the _updateFrameTexture Runtime Error (that I get with using opencv)? Instead, if it sees that error I want it to just move onto the next routine?
Hello Sam
sorry, I have no idea how to do that.
Best wishes Jens
No worries at all! One other thing I perhaps should have mentioned is that I’m getting this warning every time too:
WARNING User requested fullscreen with size [1280 800], but screen is actually [2560, 1600]. Using actual size
Which is often followed by:
5260.4998 INFO Loaded monitor calibration from [‘2021_07_07 10:58’]
5305.9259 WARNING Monitor specification not found. Creating a temporary one…
5306.1706 INFO Loaded monitor calibration from [‘2021_07_07 10:58’]
Could this be causing the problem? If so, how might I fix it? I have no idea how/why it is requesting a fullscreen with size [1280 800] when my resolution is 2560, 1600, or how to change it?
Hello Sam
that is not likely the cause of our problem. PsychoPy comes with a default size of a monitor (1200, 800) which you can check under Edit experiment setting → Screen. If you want to get rid of this warning, create a monitor-description under Monitor settings and calibration.
Best wishes Jens
OK, good news! I seem to have found a solution which gets my videos to play perfectly (but only using opencv rather than moviepy).
Amazingly, all I did was change the dropdown setting of “stop” on my targetVideo component from duration (s) to time (s). Now the videos run absolutely as intended! When I change it back to duration (s) the experiment starts crashing again.
Does anyone know why that might be the case? I’m absolutely puzzled after spending 2-3 days on this!!!
Hello Sam
good that you found a solution. Anyway, a word of caution. Image a movie starting at 0 with a duration of 3 seconds. If you set duration to 3, then you might loose some frames because the loading at routine start is not taken into consideration. You could leave duration undefined and check the **Force end of Routine ** to end the routine and register response following your routine.
Best wishes Jens
Hi Jens,
Thanks for your help on all this! I’ll definitely keep your warning in mind
Best wishes in return,
Sam
Hi Sam,
Could you describe your solution in detail?
Best wishes
Woody
Hello Woody
Best wishes Jens
For anyone joining the bottom of this old thread, I would strongly suggest using the ffpyplayer backend (it is faster to render frames than opencv, which was also faster than moviepy)