Trial processing often fails and shows red dots

[Christina Frese]

Dear OpenCap-Team,

I ran this weekend in some issues with opencap. We did measurements in the same laboratory was the weekend before, when most of the videos could be processed. This weekend I had big difficulties. 7-8 videos from 10 failes to synchronise even though it were fast movements (spike jump). Can you please look into the sessions and see what the problem was? I do not understand because before it worked really well the other weekend. Is it possible to fix the trial in a post processing? I would be really glad about some help and if some tirals can be fixed.

My name is knieschmerzen_im_volleyball. The sessions were many. From 09.09.2023 and 10.09.2023. Session 306c2449-ab09-4dc6-9b44-4ff313ba3843 up until session: cadf40d7-cf06-4e74-a4ce-1bc837a32aaa. About 20 sessions all in all. We recalibrated in between as well. Maybe the athlete was to small in some in the begging? You said that that could be a problem and you have a specific pipeline for that? But as I said. The weekend before we did not have those problems.

Best regards,
Christina

[Antoine Falisse]

Hey, we will check. Could you please list all the sessions? Thanks.

[Christina Frese]

Of course. Here are all the sessions. As I said. We measured all day long

cadf40d7-cf06-4e74-a4ce-1bc837a32aaa

27a9972c-5705-4306-b89b-720a87acd0e7

1e3c7b5c-be0e-4c73-89b2-7f541d4703b6

0d8e9f5a-4a32-4a99-8cce-2c6b7e32e4f1

e2b74976-6957-44d5-a5f1-e93bac92523f

6ce194a1-d2c7-4721-8543-554fbb05fc67

0ac9fbaa-8dcd-46e7-b7d7-f386334ab511

e846a5dd-ef50-4d57-a377-7781abd873bf

2d7fdfa3-64ce-4202-8347-25d5c0c982b3

4363db7d-e3c9-4362-969a-dd1f906f3491

0f1d17c9-0ebe-43ca-a424-57e989f76a79

79e6aa49-f79b-4d81-b1be-8622a214e61c

c6b10cb4-7bad-47ae-96eb-b1e0d1ce8a3a

33e08bf7-32bd-449f-8685-dab8cbe4d09d

57ba75f1-a612-4fb4-8686-b4f2224e8c9f

162fa02f-54de-47e8-84cc-998ed1da76da

5036bc7e-c5a9-4f5e-a5cd-4456cc18e366

9af3483e-18a5-4f26-b670-cb2556ef704c

226f90d2-2f08-4929-9a9c-5ea0258d389c

4a1c8d8d-917f-4e05-aa22-7186d415ca56

72394052-fdf8-4fef-a42f-59ef094df038

93ddf934-8fab-4d91-bd39-6f37adc61cdd

306c2449-ab09-4dc6-9b44-4ff313ba3843

98ed1d48-0870-42ed-bf7d-c03d811ee2e0

[Scott Uhlrich]

Hi Christina,

We identified a few problems:
1) we may have had a failing server over the weekend
2) make sure that the checkerboard is mounted to something very flat, and that there aren't people or posters that look like people in the field of view. This can cause intermittent problems with both calibration and neutral trials. Make sure the neutral skeleton looks reasonable before continuing to collect data - this can be an indication that there was a problem in calibration or the neutral trial.

We reprocessed all sessions except 27a9972c-5705-4306-b89b-720a87acd0e7. This one is trickier to reprocess. Check back in 1 day to see if the data from the other sessions looks better.

[Christina Frese]

Hello Scott,

thank you so much. I also got once the error that the server was down. It was also a bit slower than usually. But I could measure even still after the error message. I thought about the board. We had it on a chair first. That worked well for the first weekend. I also had the feeling that the paper bended a bit, so I put it on a box. Okay. I will be more careful next time. I was wondering why the calibration of the subjects were so bad in the last two sessions. But I see that in those session the camera looked on the posters and before not that much. But one at least looks really good now. The next weekend we will cover the poster or put the camera in such a way that it does not include it.

I already saw now that more data is green and looks good. A big relieve for me. I will check tomorrow. Thank you for your help.

Best regards,
Christina