trial fail

[Mina Salehi]

Hello OpenCap team,

I faced an issue today that I have not experienced in my previous sessions. Many of trials failed and turned red. I recalibrated cameras and changed the background color but it did not help. Since some of the trials successfully loaded and turned green, I don't think the problem is related to camera calibration. Do you have any idea what may have caused this issue? Is there anyway to recover these trials?

Session ids:
0dc0cf7f-8309-4fa8-93ce-bbe38b24c506
840a7bd6-7d20-42e6-902d-05d17420df1d

Thank you

[Lisa MacFadden]

This happened to us as well today during a few sessions with research participants. We have troubleshooted many different ways including swapping out video capture devices (iPads), recalibrating, logging out and back into OpenCap, and had no success. There were a handful of trials that worked and processed, but many that remained red. Our iPad OS did update yesterday, wondering if that might be part of it. We have more subjects scheduled for tomorrow so hoping that we can figure out the issue before tomorrow.

It is worth noting that there were two subjects collected today that we had no issues with; the issues started after our first two participants, so the update may not be relevant. There were also no issues with the neutral pose. There were also times that the trials went red, then back to yellow, then red for good.

Additionally in the past when I have gotten errors in processing (red dots) there has been a detailed description of the error

Usual Red Dot Error.png (136.46 KiB) Viewed 892 times

and today we are just getting a red box with the number 2 inside it.

July 17 Red Dot Error.png (13.31 KiB) Viewed 892 times

We appreciate any insight from the development team or users!

Adding Session IDs
6b83d0a3-6e89-48d9-bf09-b439792d68ad (Failed troubleshooting)
ae9a3b0f-ae9a-43fe-8184-8302b47b2fa4
42e8748e-8d8b-4522-a074-f2a7e7afa1bd
1273334c-66c5-44d6-9bcb-440e620b2aaa
a529f321-649a-4f6c-a56e-ec25d5be471a

[Lisa MacFadden]

Adding more information

I downloaded the data to look at the video files on my computer. It looks like they all recorded (except for our troubleshooting trial in which one camera did not record/upload) and the ones that did not process had issues synchronizing.

We are going to try again tomorrow with the fail-safe sync method and have participants quickly raise one hand above their shoulders and bring it back down before completing the rest of the data collection. Will provide an update.

[Mina Salehi]

Hi Lisa,

Thank you for your comment. That red "2 close" error was exactly the error that we got.
I will also try the fail-safe sync method tomorrow. I hope it works!

Mina

[Lisa MacFadden]

We did some captures this morning to try out things on ourselves before we worked with participants. All four trials worked. I have described them below.

First trial - used the hand trick then proceeded with squats (focus of our testing) - this trial synced and processed well.
Second trial - did not use the hand trick, just proceeded with squats - this trial synced and processed well.
Third trial - did not use the hand trick, did random movements - this trial synced and processed well.
Fourth Trial - did not use the hand trick, paused before proceeding with squats - this trial synced and processed well.

[Lisa MacFadden]

So far, after our troubleshooting trial success (not really "troubleshooting" because everything was working, and we didn't need to adjust anything), our research trials this morning are also working and processing as expected.

We downloaded the data from yesterday's failed processing and confirmed we had the videos from all cameras and plan to try to reprocess the data ourselves.

[Antoine Falisse]

Hi Lisa,

I sent back the trials with errors for reprocessing. Please take a look in a couple hours and let us know whether things are better. Some machines may have failed us. Side note, you might want to ask your participant to briefly punch the air (fail-safe synchronization mode); I see that camera synchronization clearly failed in some cases.

Best,
Antoine