The choice of the video resolution in the Control Center need to be balanced between detection area and CPU usage
The maximum detection distance increases with the resolution (you can calculate the ratio Detection Range / Resolution here).
However, a larger resolution demands more CPU power and, at one point, you may exceed VidiReports‘s capabilities - A higher resolution may also reduce the frame rate provided by the camera. You should use the highest resolution that leave the FPS consistently above 10.
Switching between YUY2 or MJPG if your camera support both format can also improve the FPS - Using YUY2 will avoid decompressing JPEG frames and will save CPU, But only for "smaller" resolutions - Since it's uncompressed, you are limited by the USB2 bandwidth (480 Mb/s). Don't hesitate to try both mode to see which ones works for your installation.
Also to take into account: the viewing angle (FOV). A larger viewing angle decreases the maximum detection distance.
Note:
- While cameras may provide a FPS as high as 60, VidiReports will limit the acquisition framerate to 15.
- 8 FPS is the minimum required for accurately tracking traffic - Under this threshold, you will get an alert and the detection will not be accurate.
- CPU resources will depend of the amount of concurrent face detection - If you are testing Vidireports in a place with low traffic (an office for example), keep in mind that FPS might fluctuate when it is deployed in a space with more traffic.
- This doesn't apply to Footfall - Processing is linear no matter how many body are being detected. Resolution also doesn't impact Footfall detection as much as face detection, bodies being much bigger object to detect. Detection range will not noticeably vary between 800*600 and 1920*1080.
You may prefer the 16/9 aspect ratio when available to the 4/3 one, since faces generally appear in a narrow stripe in the field of view. You can find a more detailed article on aspect ratio here