Video "Failed to load because no supported source was found"

Online experiments
,
1

URL of experiment: Sign in · GitLab

Description of the problem: My experiment has videos in it using the movie component in builder. I previously had a version of the experiment where everything worked in Pavlovia, including all of the videos. However, something has changed recently (and for the life of me I can’t figure out what) which means that some of the videos no longer play, while some play fine. The experiment doesn’t crash, the video simply doesn’t show at all.

I am getting the following error when I look at developer tools:

image
image813×195 75.9 KB

Here’s what I have tried:

  • Checked the videos which don’t play exist in the relevant video&audio folder which is in the experiment folder (they are)
  • Checked video file names match the paths in the associated Excel spreadsheet (they do)
  • Checked whether the video file names contain spaces (this was previously an issue, but they don’t)
  • All videos are encoded the same way (h264 encoded .mp4 files, sampling rate of 44,100khz, AAC audio, 25 frames/second)
  • Checked the videos that don’t play in Pavlovia play locally in PsychoPy (they do)
  • Checked whether it’s the same videos not playing on each pilot (it is. If you want to test this, an example of a trial that doesn’t work is the word ‘feudal’; one that does work is ‘primal’)
  • I have made a bespoke videoTest experiment which has all other elements removed except for the routine that is meant to play the videos. This routine is directly copy and pasted from the main experiment so everything should be exactly the same. All the videos run in this! Link to this here: Sign in · GitLab
  • I have also tried adding all the videos as additional resources in the ‘Online’ tab of the experiment settings, to no avail

If anyone else has previously had this error and managed to solve it, please tell me what you did! Help and ideas for a solution very welcome - I could be missing something obvious.

Many thanks!

2

I had the same problem. I have 200 videos in my experiment. I had issue playing videos if I load all 200 videos but no problem at all if I shorten my experiment to only 40 videos.

It seemed that this was due to the recent update of Chrome which limited the number of web media players. Here is a relevant topic: Chromium, Chrome, Edge version 90 and more limits number of opened MediaPlayers
Also see the thread here: google chrome - Blocked attempt to create a WebMediaPlayer as there are too many WebMediaPlayers already in existence - Stack Overflow
In this thread, it was suggested that Chrome will be raising the limit to 1,000 today (3rd of August). I’ve just updated my Chrome and retried my experiment and it seems work.

1 Like
3

Hi @dwang, thanks so much for the reply. This is interesting. My laptop doesn’t seem to have the update available yet, but presumably it might appear over the next few days. The issue though is that my participants might not have the most recent browser version and I’m not sure how difficult it will be to ask participants to update their browser before participating (some of my participants will be older adults)… I’m wondering whether I need to split my experiment into smaller chunks and perhaps manually manage some of the counterbalancing - what a pain!

4

Hi,

I can confirm that for Chromium, Chrome, Edge version 90 and more limits number of opened MediaPlayers using the new version 92.0.4515.131 of chrome solved the issue.

regards,
JEM

2 Likes
5

I see. Maybe just wait for several days until the browser completes updating by itself. Good luck!

1 Like