1 Topic by james.stevenson

  • james.stevenson
  • -
  • Offline

Hi all,

I've been using Kinovea to run some lab experiments. We film an object at a high framerate, and get a motion profile of the object by analysing the video. Kinovea has been the only software that satisfies all our needs! Without spending bucketloads...

But I'm running into trouble, I have some interesting errors in the data. It could be due to time-jitter of the frames, but in our analysis we assume they were taken evenly in time. The camera is from Basler, and it should be possible to get hardware timestamps in the chunk data sent from the camera. I think this is nanosecond precision. The trouble is getting those timestamps out into a file that I can use in the analysis. I see timestamps in the mkv file, but they don't look like they have this kind of precision, so I think they are not from the hardward timestamps.

If anyone has any ideas on how to do this, that would be great!

2 Reply by impeccable (edited by impeccable Yesterday 03:09:53)

  • impeccable
  • -
  • Offline

Hi, interesting setup! From what you described, it does sound like the timestamps in the MKV container aren’t preserving the original hardware precision from the Basler camera. That’s pretty common since container timestamps are often limited by the encoding pipeline.

If you really need nanosecond-level accuracy, you might want to look into extracting the timestamps directly from the camera SDK (Basler Pylon) instead of relying on the video file. Some cameras can embed precise timestamps in metadata or chunk data, but you usually have to log that separately during acquisition. Another option could be saving the data in a format that preserves metadata more explicitly, or exporting timestamps alongside the video during capture. Kinovea is great, but it might not be designed for that level of timing precision. hollow knight

Curious to hear if you find a good workaround!

[url=https://hollowknight.io]hollow knight[/url]