URL of experiment: https://run.pavlovia.org/sheremat/rsvp
Description of the problem: I have an experiment- an RSVP task presented with a movie of random dot motion. There are 6 movies about 400 KB each of motion. The experiment runs mostly without issue up to about 60 trials (about 6 minutes into the experiment) and then crashes, displaying only a blank screen. When this happens, the experiment is also very slow to accept keyboard input, but if I press the escape key for a long time will gradually quit out and download the output file. When I can get the developer tools there is no error that would indicate that there is something wrong with the code.
Something else that is odd that could be related is that I cannot define the size of the movie in pixels. I can choose pixels, but if I select 1024,768, for example, I get it scaled to be ~1024,768x the size of the screen.
I’m pretty certain this issue is not specific to any specific movie file, as it crashes on different movies as reported in the output file. Also, each movie is repeated 9 or 10 times before the crash.
Please could you share the URL to the gitlab page of your project?
from your experiment page: view code> settings on left hand bar > permissions [set visibility to public
then share the URL of the project with us.
Thanks the site is:
It should be st to public already.
I’ve been working on this and still have had no luck. Things I have tried:
I have stopped updating the letters in the RSVP stream. I am now convinced this is a memory issue with playing 6 relatively small movie files (300-400KB each). Safari is giving me a memory warning about 30 trials into the experiment and then crashing around 60 trials in.
I have added reset and play at the beginning of the routine and stop at the end of the routine.
Occasionally a movie plays very slowly. I have color coded the letters based upon which movie is playing and it appears random, though more often on the slightly larger movie files.
I have stopped logging.
Anybody run into issues playing a movie and resolved it a different way?
Hey @sheremat, thanks for flagging, just now I saw this sorry, sounds like a bug, please allow me some time to investigate further, x
Thanks @sotiri I appreciate it. I’ve worked it down to just the essential parts and it’s still crashing.
KO, I hear you, let’s see, more soon, x
Hi @sheremat, apologies for the late response, as far as I can tell making a few post PixiJS upgrade tweaks seems to have helped in that I was able to go through well over 60 trials without getting any warnings or crashes. Please let me know if that fails to solve the problem or if you need more help bringing those changes into your project, x
Thank you so much- it works great. I’ve added back the RSVP stream (I was running into the issue that the server was not running the updated code, so it took a little time to test the code).
OK great, thank you, yes unfortunately you would need to make sure to import PsychoJS components from that custom-lib folder (lines 5-11) instead of relying on default Builder output, but only temporarily until we get Pavlovia up to speed with the latest patches. Please post back should that approach fail to work. Many thanks for your patience while we work to get the problem fixed for good
Hi @sheremat, OK PsychoJS latest now features the PixiJS post upgrade tweaks required to make your project work and you no longer have to rely on importing from ‘custom-lib’. Please post back if you come across any more issues getting your experiment to run as intended, x
Thank you for all your help, this is great!