Frame Rate

The frame rate of a video stream is the number of individual images per second that compose the video. When these images are shown in rapid succession, they appear like a smooth video. Each video stream has its own frame rate. Typical video frame rates are 15, 24 and 30, but there are many other options. Some cameras allow users to set the frame rate of the video stream that the camera produces. Some cameras even produce multiple streams with different frame rates.

Video from network cameras is almost always compressed to save network bandwidth. To process this video, EyesOnIt must decompress or decode every frame of the video. This decoding operation is quick for a CPU or GPU. Once a frame is decoded, EyesOnIt can choose whether to process the frame or not. By comparison to decoding a video frame, processing a frame can be a very intensive operation. Therefore, skipping frames that don’t need to be processed can generate significant savings in CPU and GPU usage.

In certain use cases, processing every frame is important. As an example, in a manufacturing process where EyesOnIt is analyzing video of a fast-moving assembly line, operators may need EyesOnIt to detect a condition that is only present for a small fraction of a second. In that case, processing every video frame may be essential.

In other use cases, skipping frames is entirely acceptable. In our documentation we have used the example of a vehicle on a driveway. Processing one frame per second is more than sufficient for this scenario and leaves the CPU or GPU available to process other video streams.

When you enter your settings for video processing, choose the lowest frame rate that is acceptable for your use case. Enter the frame rate in the Frame Rate box:

The minimum frame rate value is 1. If you set the Frame Rate higher than the number of frames per second in your video stream, EyesOnIt will process every frame from your stream.