Hi,
I am implementing my pipeline for opencv processing using libcamera's simple-cam as a base.
I use a imx296 GS Camera that is externally triggered via an arduino nano.
For now i can stream at full resolution at 60fps, which is expected.
But when i use an hardware ROI very much like GScrop does (from HermannSW), it seems that whatever i do, i can't get above 60fps.
Even funnier, i observed that any fps value above 60 will give me half of the target fps (61 gives 30.5, 100 gives 50), and above 120, the values do not have this "half" relation anymore. They drop to 40~ fps.
What bugs me, is that i was able to record videos using GScrop at 1455 * 112 size, at 60, 200, 380, 410 fps easly, using the nano external trigger.
So my question is : is there a major difference between how rpicam-vid sets the framerate and how libcamera sets the framerate ?
Thanks in advance !![;)]( "Wink")
I am implementing my pipeline for opencv processing using libcamera's simple-cam as a base.
I use a imx296 GS Camera that is externally triggered via an arduino nano.
For now i can stream at full resolution at 60fps, which is expected.
But when i use an hardware ROI very much like GScrop does (from HermannSW), it seems that whatever i do, i can't get above 60fps.
Even funnier, i observed that any fps value above 60 will give me half of the target fps (61 gives 30.5, 100 gives 50), and above 120, the values do not have this "half" relation anymore. They drop to 40~ fps.
What bugs me, is that i was able to record videos using GScrop at 1455 * 112 size, at 60, 200, 380, 410 fps easly, using the nano external trigger.
So my question is : is there a major difference between how rpicam-vid sets the framerate and how libcamera sets the framerate ?
Thanks in advance !
Statistics: Posted by ricaizen — Wed Dec 31, 2025 12:28 am