psychopy.org | Reference | Downloads | Github

Movies display as solid colors for some users

URL of experiment: https://pavlovia.org/hekragness/puppy-in-the-park

Description of the problem:

I have asked multiple pals to pilot test my project, and most seemed to have no issues. Even though everyone used Chrome, one had an issue where the videos did not load properly and displayed as solid colors on the screen.

I thought this might be due to loading issues related to file size. I used Handbrake to “web-optimize” one test video, the first one in the experiment (2_1.mp4), which made it must smaller (13.1 MB -> 1 MB). Now THAT video is improperly displaying as solid colors on my side (whereas I previously had no issue playing the larger file size).

I have followed the recommendations for video properties and encoding on the Wakefield crib sheet, and the videos start at time = 1 as recommended. All resources are loaded at the beginning of the experiment.

Are there further steps I can take to ensure the videos play properly?

As an update: the error on my end seems to have been an one-off, as it works now. For the user on the other end, the glitch is still present. Sometime it presents as a solid color on the screen, and sometimes as in the example below. Any ideas? This person is using Chrome browser on a Windows 10 computer.

pavlovia_screenglitch

Hi Haley,

Yeah, video encoding can be a bit tricky. I’d recommend trying other software to encode the MP4 to see if that helps. As a little check, does this experiment show the video correctly? https://run.pavlovia.org/tpronk/demo_video/html/

Best, Thomas

I also want to link another discussion here that could be helpful Movie stim is just a black box on Pavlovia

Thanks! She says that video DOES display correctly. I’m a little puzzled because it seems like she’s using the same browser as virtually everyone else, yet is the only one with this issue. It’s possibly notable that she reported the longest loading time out of everyone, as well. So maybe a slow connection? But if the vids all load in advance then I’m not sure why a slow connection would be problematic throughout the experiment.

Thanks! I was trying to access the custom-lib folder mentioned but it links to a 404!

What do you mean with the “custom-lib” folder?

About your video. I’m not sure how Handbrake works, but I can tell you that the video in the demo I shared is encoded via H.264; that codec tends to work on a lot of different browsers. Maybe encoding your videos using that codec could work too?

Ah, the custom-lib response was to Becca’s comment and link above. Sorry for the confusion!

Yes, according to Handbrake and VLC that is the codec fort the videos used… but I think you’re right that it might be worth trying some different software and just see what happens. Thanks!

1 Like

Hi @HEKragness, sorry to hear PsychoJS is giving you video playback problems :blush: Could you add me in as developer so I can send you a merge request for incorporating the same tweaks as Cailey’s, see if that helps any? Thanks, s.

Done!

OK great thanks, merge request issued :crossed_fingers: Cool study! s.

Thanks, I’ll let you know what happens! If it doesn’t end up working for some percentage of participants we’ll just have to live with it (such is online testing I guess) but very worth trying :slight_smile:

Alright, no problem, thank you :slightly_smiling_face:

As an update: after merging the tweak I had her try the experiment again. It didn’t seem to fix her issue, but it did introduce a new issue for all users such that if a video played once and was called again later, the “force end of routine” did not seem to work the second time - it just hung out at the end of the video forever. I didn’t do any testing on this to figure exactly what was going on since it didn’t fix our initial issue.

Thanks for trying, though!