1

First off, thanks for the great software! I have a question concerning how frames and timecodes are aligned in Kinovea. I'll try to illustrate the issue:

Take for example a video with a framerate of 50 Hz. That way each frame would cover a time slice of 20 ms.

Time    |1-------20-------40-------60-------80-------100------->
Frame  |<---1--><---2--><---3--> etc.

If I now annotate my video, this annotation happens necessarily on a frame by frame basis, since this is the upper limit of temporal resolution. When exporting my annotations (say x/y coordinates for each frame) I was wondering how the timestamp for each measurement is actually derived. Multipe possibilities come to mind:

Is it the starting time for the frame, e.g., 1 for frame 1?
Is it the ending time for the frame, e.g., 20 for frame 1?
Is it the timepoint in the middle of a frames time slice, e.g., 10 for frame 1?

Note though, that may illustration does not take into account the fact that a frame may not always cover the whole time segment, because it's exposure time could be much shorter (e.g., in bright conditions). Then of course, middle or end timepoint wouldn't make much sense.

Thanks for any hints to clear this up for me!

2

Yes it's a good observation. There is no "absolute" time reference so whether you assume the events in the images are happening at the beginning, middle or end of the frame interval doesn't really matter as long as you always assume the same alignment, addition/subtraction between time stamps will work the same. The timestamps are coming from the video file and ultimately from the camera where they could mean any alignment. Inside the playback screen as soon as the time interval starts it tries to display the corresponding frame.

To further complicate the issue many cameras have a rolling shutter, different *lines* are captured at different times within the frame interval.

3

Thanks a lot, that clears it up!