You are right - it's a complete pain in the .....
The logging and searching of video data could be split into three seperate incidents.
- Realtime using operator input or automated analytics
- Post event using the original system either on-site or remotely
- After export, back in the Police Dept. or Video Lab
If an integrated system has the ability to log events in realtime such as movements, direction of travel, vehicle colours etc then this is great but, this has to be exported with the video data, and the ability to use the analytical data has to be possible away from the scene/system.
A lot of Systems / DVR's have analytical services included but its sometimes not possible to operate it and use the system on-site.
In the last point you have two common types.
The first is that the footage has to be reviewed to search for an event linked to an incident. That could be a person running from a scene or a vehicle driving past.
The period of time to search and the actual event captured is also commonly not known. It is always a manual task to identify the search parameters and then review that camera, logging events of note until that viewer hits the jackpot.
The most difficult ones are when there are unclear directions of travel and perhaps a large timescale. When that viewer finds the target vehicle / person captured by a stores front door camera at 0400hrs - it can be a real breakthrough and that one image can dictate the entire direction of an investigation.
The second is when there are many events. A common one is thefts by an employee etc. These can be very time consuming and are used to show a pattern of intent.
Other times are when someones actions and behavior changes. A person enters a store every morning at 0900hrs. Then after an incident, they dont show up. It all builds the picture but finding it all takes time.
Effective searching comes down to intelligence led parameters. These parameters are then changeable as further intel / evidence comes to light. Its common knowledge that in the first few hours after the London Bombings, the location parameters were the entire UK!
So, at the start of an investigation we may grab as much as possible but as certain incidnets are discovered, decisions are made that negate further work.
After saying all that though I have been involved in many cases, where weeks worth of footage has been painstakingly viewed but the 3 frames of video found have made it all worthwhile.
There are a few tools now available to ingest video, set certain parameters and then have a large amount automated. For specific sites and footage these are useful but for the majority they can be more of a hindrence. As the technology improves and those functions are built in to systems the situation will improve.
Hope that helps a little?