What is ‘I Frame Interval’ and what should it be set to please?
- Thread starter virdi
- Start date
- Messages
- 4,691
- Points
- 113
Hi @virdi
Here is a simplified explanation.
Most network surveillance camera streams are made up of I-frames and P-frames, I-frames are full image captures that record everything the camera is seeing and P-frames are the frames between the I-frames where only the elements of the image that are changing are refreshed.
To explain this with an example, if you had an empty corridor and someone walked down it the I-frame would capture the entire image and then the following P-frames would only capture the elements of the image that are changing, which in this case would be the person walking.
The idea behind this is that it is another way to compress the bitrate of the camera stream because instead of capturing 25 full-resolution images every second you are capturing one full image every 50 frames (e.g. 2 seconds) and then all the frames in between only capture what changes in the image and thus are much smaller files/use less bitrate.
Reducing the I-frame interval (25 frames, 10 frames, etc...) will obviously increase the bitrate/bandwidth used by the camera and is only really recommended if you are installing the camera in a very busy scene where having the full image refresh more regularly can improve image detail. Increasing the I-frame interval (75 frames, 100 frames, etc...) will decrease the bandwidth used and is only really recommended when covering scenes with little to no activity as you will see a drop in image quality and maybe even distortion to any movement that happens between the I-frames.
Here is a simplified explanation.
Most network surveillance camera streams are made up of I-frames and P-frames, I-frames are full image captures that record everything the camera is seeing and P-frames are the frames between the I-frames where only the elements of the image that are changing are refreshed.
To explain this with an example, if you had an empty corridor and someone walked down it the I-frame would capture the entire image and then the following P-frames would only capture the elements of the image that are changing, which in this case would be the person walking.
The idea behind this is that it is another way to compress the bitrate of the camera stream because instead of capturing 25 full-resolution images every second you are capturing one full image every 50 frames (e.g. 2 seconds) and then all the frames in between only capture what changes in the image and thus are much smaller files/use less bitrate.
Reducing the I-frame interval (25 frames, 10 frames, etc...) will obviously increase the bitrate/bandwidth used by the camera and is only really recommended if you are installing the camera in a very busy scene where having the full image refresh more regularly can improve image detail. Increasing the I-frame interval (75 frames, 100 frames, etc...) will decrease the bandwidth used and is only really recommended when covering scenes with little to no activity as you will see a drop in image quality and maybe even distortion to any movement that happens between the I-frames.
- Messages
- 4,691
- Points
- 113
The default I-frame should be 50 and in most cases, we wouldn't recommend changing from this, but if you are not having any issues with the cameras set to 20/25 you can leave them at those intervals. (although this will mean each camera uses more data and more storage)
- Messages
- 4,691
- Points
- 113
The image quality is really dependent on how busy the scene is, but for a pretty standard residential scene with minimal activity you probably aren't going to notice a massive difference when just viewing the live view, but when looking at downloaded footage and/or screenshots side-by-side you may notice some smaller details are clearer (especially if you blow the images up).
