I updated GScrop tool to work with Pi5 as well, and it uses timelapse.py for frame delta analysis on Pi5:
https://gist.github.com/Hermann-SW/e604 ... /revisions
For a .mp4 video without frameskips the plot provides details about the frame delta distribution.
This time I clicked on matplotlib "Save the figure" button instead of capturing screenshot as in previous posting.
I only captured 0.2s of normal rpicam-vid video at 60fps in order to see the point to point lines in the plot:
For the 10s video with few frameskips from previous posting these narrow frame delta details are invisible on x-axis:
I forked rpicam-apps repo and added "--narrow" option to timelapse.py:
https://github.com/Hermann-SW/rpicam-ap ... 78091dd23d
Now "--narrow" option ...
... keeps the frameskips visible, but avoids them dominate y-axis resolution.
The same frame delta distribution details as in first plot without frameskips are visible now.
The view without "--narrow" is the first plot I look into to see whether multiframeskips are present.
That view is part of GScrop tool:
https://gist.github.com/Hermann-SW/e604 ... op-L13-L16Then I use "--narrow" plot for details on how narrow the frame deltas are.
P.S:
I just created pull request for "--narrow" option of timestamp.py:
https://github.com/raspberrypi/rpicam-apps/pull/624
https://gist.github.com/Hermann-SW/e604 ... /revisions
For a .mp4 video without frameskips the plot provides details about the frame delta distribution.
This time I clicked on matplotlib "Save the figure" button instead of capturing screenshot as in previous posting.
I only captured 0.2s of normal rpicam-vid video at 60fps in order to see the point to point lines in the plot:
For the 10s video with few frameskips from previous posting these narrow frame delta details are invisible on x-axis:
I forked rpicam-apps repo and added "--narrow" option to timelapse.py:
https://github.com/Hermann-SW/rpicam-ap ... 78091dd23d
Code:
usage: timestamp.py [-h] [--plot] [--narrow] filenamerpicam-apps timestamp analysis toolpositional arguments: filename PTS file generated from rpicam-vid (with a .txt or .pts extension) or an avi/mkv/mp4 container fileoptions: -h, --help show this help message and exit --plot Plot timestamp graph --narrow Frameskips outside viewNow "--narrow" option ...
Code:
pi@raspberrypi5:~ $ ~/venv/bin/python ~/rpicam-apps/utils/timestamp.py --plot --narrow test.mp4Total: 3988 frames (3987 samples)Average: 2.499 ms / 400.180 fpsMinimum: 2.471 ms at frame 1030Maximum: 12.444 ms at frame 953Outliers: 4 (100%) 9 (10.0%) 10 (1.0%) 3986 (0.1%)... keeps the frameskips visible, but avoids them dominate y-axis resolution.
The same frame delta distribution details as in first plot without frameskips are visible now.
The view without "--narrow" is the first plot I look into to see whether multiframeskips are present.
That view is part of GScrop tool:
https://gist.github.com/Hermann-SW/e604 ... op-L13-L16
Code:
then rpicam-vid "$workaround" --width "$1" --height "$2" --denoise cdn_off --framerate "$3" -t "$4" "$SHTR" "$5" -o /dev/shm/tst.mp4 -n ;echo ~/venv/bin/python rpicam-apps/utils/timestamp.py --plot /dev/shm/tst.mp4elseP.S:
I just created pull request for "--narrow" option of timestamp.py:
https://github.com/raspberrypi/rpicam-apps/pull/624
Statistics: Posted by HermannSW — Fri Dec 22, 2023 7:12 am