Yes, inter-frame GOP compression does work like that. You have a chain of one independent frame that encodes the entire picture, followed by partial frames that encode only the differences from the previous frame. This sequences repeats and defines the GOP length.
If the changes from one frame to the next are small, it is quite possible for the codec not to need the target bitrate -- all the cameras we are concerned with in this thread use variable bitrate codecs, and the target bitrate is a maximum.
Try encoding 30-seconds of some random video using variable bitrate with AVC/MPEG4/H.264, then repeat by using a color matte. It is an extreme example because the image is static, but it demonstrates that the complexity of the scene (or lack thereof) affects the actual bitrate.
Do you think doubling the target bitrate will give you a file twice the size? Do you think it will improve the quality?
If the changes from one frame to the next are small, it is quite possible for the codec not to need the target bitrate -- all the cameras we are concerned with in this thread use variable bitrate codecs, and the target bitrate is a maximum.
Try encoding 30-seconds of some random video using variable bitrate with AVC/MPEG4/H.264, then repeat by using a color matte. It is an extreme example because the image is static, but it demonstrates that the complexity of the scene (or lack thereof) affects the actual bitrate.
Do you think doubling the target bitrate will give you a file twice the size? Do you think it will improve the quality?