If you read the specification sheet for any modern CCTV Digital Video Recorder (DVR) you’ll see one of the headline figures is the number of frames per second (FPS) the system is capable of recording. It’s not unusual to see 200FPS being stated. When you consider that standard television is broadcast at 25FPS then it really is impressive to see a such a high frame rate listed. Surely the images will be ‘super smooth’? The good news is that these systems can indeed record impressively high frame rates. The bad news is that there a lot of caveats to be factored into how this is achieved. In this article we will try to provide a plain English explanation of what the numbers mean and how they affect the quality of the CCTV recording.
Frames divided by cameras
DVRs can record CCTV footage from a number of different cameras. Typically up to 16 cameras per unit. If we took the example of a 200FPS capable DVR with 16 camera inputs (channels) then you have to divide the number of frames per second by the number of cameras connected to the system. 200fps divided by 16 cameras leaves us with 12.5FPS per camera. This is half the frame rate of standard television. Suddenly our smooth images are going to take on a stuttering appearance but that’s not all.
More frames less pixels
Once you start digging into the specifications of some DVRs you find that the 200FPS rate only applies at very small frame sizes.
This is what you might see listed for a 16 channel DVR:
D1(720×576 / 50FPS), Half D1(720×288 / 100FPS), CIF(320×288 / 200FPS)
Does that make sense? If it does then you shouldn’t need to read this blog post. If some of it resonates or you’re completely lost then I’ll break these figures down and explain the terminology.
D1 is a full frame of standard definition digital video. There are 720 horizontal pixels across the image and 576 vertical pixels from top to bottom. 50FPS means fifty frames per second. So at the highest frame size we get 50FPS. Dividing this by 16 channels gives us just over 3FPS per camera. At frame rates this low the images will look very jumpy indeed – pretty much like a series of still images (which is what they effectively are).
Half D1 has half the number of vertical pixels as D1. It still has 720 across but there are only 288 pixels from top to bottom. The frame rate the DVR is capable of recording at half D1 is 100FPS which works out at 6.25FPS per camera. It’s a lower resolution but a higher frame rate. Slightly smoother but still well short of the 25FPS for smooth(ish) looking video.
CIF means Common Intermediate Format but in simple terms it’s one quarter of D1 with 320 horizontal pixels and 288 vertical pixels. At this resolution we see the magic figure of 200FPS appearing. We still need to divide this by the number of cameras (200/16) which delivers a frame rate of 12.5FPS per camera.
Quality vs quantity
In the example above the 16 channel 200FPS DVR can either record around 3FPS in full frame standard definition or 12.5FPS quarter frame images per camera. When setting up the system there is a compromise between frame size and frame rate. This is NOT because of storage issues – hard drives are cheap and plenty fast enough to record multiple streams of compressed video. The system is only capable of processing a certain amount of data per second. This data (bandwidth) has to be shared between the number of inputs.
There are DVRs capable of recording much higher resolutions and at much higher frame rates. They can deliver excellent quality recordings with very high frame rates. As you might expect the more capable units are priced much higher than the type of system in this example.
Low cost DVRs are very popular in retail, industry and domestic premises. The evidence they gather can be used to help investigate crime and other matters such as accidents. It’s always an advantage to have the best possible quality video evidence. Any compromises can limit the usefulness of the exhibit.
Low resolution CCTV evidence contains less detail that higher resolution footage. This can make the images difficult to interpret and provide doubt about what has been recorded. An example of this might be that a person is said to be seen holding a knife. The defendant claims he was holding a mobile phone. The footage does show the person holding an bright object but the resolution is so low that it’s not sufficiently detailed enough to determine whether it’s a knife or a phone. A higher resolution image would show more detail and potentially show exactly what was being held.
Low frame rate CCTV evidence poses other issues even if the resolution is good. At three frames per second (often much less) the footage can look very jerky when replayed. A fast moving incident can be difficult to follow because of the amount of activity which has occurred between each frame. Even when the footage is slowed down the movement between frames is so great that it doesn’t help to determine who did what and when. A higher frame rate provides more fluid motion allowing the court to see what happened more clearly.
In a future article we will look in more detail about how frame rate and image quality compromises can affect video evidence.
We’ve tried to avoid jargon and have simplified the data presented in this article. Please see our video forensics page for more information about the evidence processing services we offer. We hope that this article has been helpful and welcome comments and feedback.