Comparing Video Quality of different videos from same source

[LMyerson]

I’m dealing with two surveillance videos taken from a security system about 10 years old. The software was GeoVision Multicam. These videos span 12:00 to 1:00 AM and 1:00 to 2:00 AM respectively. The earlier one has the critical view, but it seems to be of lower quality than the later one. Both are incorrectly listed by various software as being about 90 minutes long, although the timestamp all through the video definitely shows 1 hour in length. These are AVI files, and both have metadata saying “lossy compression”. Both files lack GeoVision’s watermark. My question is this, could the earlier video have been processed multiple times to deliberately make it lose resolution? If so, how could I tell? I’ve looked at video quality measurement software, but these packages are all for comparing versions of the same video. These two videos do have all of the stationary objects in common, but otherwise they are not the same. Any suggestions would be greatly appreciated.

[poisondeathray]

video quality measurements (SSIM, PSNR) wont help you with this, as you need an "original" to compare it with

One way is to look at the file creation dates or modified dates e.g. in windows explorer. If the "earlier" video time/date precedes the later one, that suggests the "earlier" one has been processed later. (It's not foolproof , because it's possible to alter those dates/times as well)

You can examine the compression characteristics, distribution of frame types, GOP characteristics as well in various steam analyzing software. e.g. if this system always uses a GOP length of say 15, but the earlier video uses a different value, you know it has been altered. Similarly if original only uses certain types of frames (e.g. maybe it uses no B-frames), but you see B-frames in the "earlier" video, that means it's been processed farther

BTW, it's entirely possible that this is normal. The reason is when there is no motion, no activity, it's very easy on temporal compression. As soon as there is movement, differences in frames, there is more difficulty in compression because modern codecs store the differences between frames; bigger differences means more bitrate required to maintain a certain level of quality. If the recorder uses a fixed bitrate, this will manifest in lower quality frames for that section of activity - unfortunately exactly what you DONT want in security video type scenarios