In video surveillance cameras, what could be said is the most likely % of data space efficiency when using H264 over MJPEG?

In video surveillance cameras, what could be said is the most likely % of data space efficiency when using H264 over MJPEG?

• H264 uses around 10% of data space compard to MJPEG.
• H264 can use 10% to 90% of data space compared to MJPEG.
• H264 uses around 90% of data space compared to MJPEG.
• They would both use around the same amount of data space. 

EXPLANATION

It is more logical to say H264 has a range of efficiency over MJPEG than a set % of data space used (bandwidth and storage) because the efficiency of H264 over MJPEG depends on a lot of variables. MJPEG takes a full picture for every frame. H264 uses temporal compression, where 1 full frame is taken at intervals ("I frame"), with in between frames broken up in grid like sectors ("P frames" and "B frames"), and only those sectors that have activity in them are recorded. So if the camera is pointed at an all white wall with no motion going on, then there is very little data being recorded in the P and B frames because there is very little change in the scene. But if the scene is, for example, outdoors looking at a busy intersection with tress constantly swaying in the wind, each frame will have lots of differences so there will be much more sectors recorded in the B and P frames. That's why estimating bandwidth and storage for H264 recorded video is a "best guess" that can vary a lot depending on scene being recorded.
Along with the citation, also see....
IPVM: Bandwidth Basics for IP Cameras
https://www.youtube.com/watch?v=8ATxwkwjJho
SIW: Choosing the right video surveillance compression
http://www.securityinfowatch.com/article/10514588/choosing-the-right-video-surveillance-compression

SOURCE

https://en.wikipedia.org/wiki/H.264/MPEG-4_AVC#Design