Subscriber Discussion

What % Of Recorded Video Footage Is Never Used/ Looked Upon?

Avatar
Gert Molkens
Jan 10, 2018
IPVMU Certified

Does anyone have any statisctics on this or links to reports containing them?

JH
John Honovich
Jan 10, 2018
IPVM

See: Live Video Monitoring Usage Statistics, we estimate, based on those survey results, that less than 1% of recorded video surveillance is monitored live. And then only a small fraction is actually later watched.

May I ask what you are looking to determine based on that percentage? It could be reasonably argued that the goal is not to maximize how much surveillance video is watched but how effectively the video is used to deter, detect or provide evidence of issues. Unlike entertainment video (e.g., watching a movie), watching surveillance video is not an end in itself.

(2)
Avatar
Gert Molkens
Jan 11, 2018
IPVMU Certified

John,

 

We're gathering massive amounts of data with all camera's installed and what do we do with all that data? With the vast amount of that data we do absolutely nothing at all.

This way, these installations are mainly a cost and egenrating useless data. I'm looking into ways to use them more effectively so that they become an investiment that provide ROI and along the way also provide security.

Regards

 

(1)
JH
John Honovich
Jan 14, 2018
IPVM

This way, these installations are mainly a cost and egenrating useless data

The fundemental problem is that you cannot know what 1% of video will become important a week or a month later. The best you can do is not record on motion but that's as far as you can really go as no system is anywhere close to knowing exactly what later will lead to an investigation, crime, dispute, etc.

I do think you should look at other ways to make use of the video but I don't see how focusing on what percentage is not used is an effective way to do so. Yes/no?

(1)
Avatar
Gert Molkens
Jan 15, 2018
IPVMU Certified

The point of the % is to make customers aware of the issue that they're generating lots of useless data.

With the emergence of AI and the ever increasing number of detection rules and the reducing false alarm rate, you could just as well start using your camera's for building control. Why not dim the lighting in an area that's not occupied? If done the right way, you'll hardly notice it but the reduction in power consumption is real. The camera's are allready there, just need to add some AI rules and interface that all together with the building control. Same goes for heating, opening/closing of doors and barriers etc etc.

 

this way, the data can be put to practical use and thos systems can start generating ROI. The security side of the recorded video hen becomes a bonus :-)

 

(1)
U
Undisclosed #7
Jan 16, 2018

IoT and AI will bring the change you are looking for. As for now you will just have to wait or create these reality yourself.

U
Undisclosed #1
Jan 10, 2018
IPVMU Certified

Here’s a link to some speculation about a similar metric.

And I wouldn’t though I wouldn’t ascribe to much weight to it, Surveillance Cameraman’s antagonist claims 99% of all recorded video gets wiped every couple weeks or so without ever being viewed;)

(5)
CR
Chad Rohde
Jan 10, 2018

A lot of the free VMSs give you 5-7 days storage. Or about 1% of a year. I don't have the patience to sit and look at more than a few minutes at a time. Even without motion detection or other analytics, and full time record, normally you have an idea on time and date of event in question. The normal home owner or small business might go years without checking. 

But it's hard to define "looked upon". So I would bet over on the 1% line. Under 5%

UI
Undisclosed Integrator #2
Jan 10, 2018

The general consistence from the circles I frequent is less than 1% of the video footage is used in a standard surveillance model. However that number does tend to increase with event based monitoring, but I have no idea on what the actual numbers would look like. 

(1)
(1)
Avatar
Skip Cusack
Jan 10, 2018

It's also interesting to consider what percentage of live video is realistically observed.  It's well accepted that a wall of cameras paired with a catatonic guard does not make for an effective detection. Has anybody ever tried to quantify this? How many feed can one person observe, and for how, long before missed detections become problematic? 

(1)
UM
Undisclosed Manufacturer #3
Jan 10, 2018

Maybe not directly related to the question but interesting nonetheless when it comes to video usage.

I'll never forget Charlie Pierce who was an avid teacher/trainer of video technology in the 90's (and before and after I think). One of his great stories was stealing a bulldozer from a heavy equipment dealer. It was a staged event where the guards knew something was going to happen but not when. This was in the days of video matrices and video walls with forty 9" monitors all over a multi-bay steel console. 

Two hours after the shift change in the middle of the night, he drove a bulldozer at snails pace over a hundred yards through the view of numerous cameras and through the gate without being detected. Bye bye $250K (+/-)!

As I recall, the note was that the attention span of a human watching monitors drops 50% in 20 minutes then precipitously thereafter. 

Live viewing without exception based notification/alerts has a very low likelihood of preventing crime. 

Alternately, more to the question posted, post event investigations only view footage associated with the loss or infraction (generally short period of time over low percentage of cameras on site). Regardless of the great variations in the number of investigated events and the number of cameras in a system, the percentage of recorded video used must be extremely low, with rare exception. 

 

 

(1)
(2)
CR
Chad Rohde
Jan 10, 2018

As I recall, the note was that the attention span of a human watching monitors drops 50% in 20 minutes then precipitously thereafter.

For the majority of video feeds definitely, but I bet there are a few that get full attention at certain venues.

(1)
UM
Undisclosed Manufacturer #3
Jan 10, 2018

Totally agreed. But in those highly sensitive applications I'd bet they're not asking one individual to watch multitudes of live video feeds. It's all about the purpose and realistic expectations of the performance of a human being and his/her limited capacity to process a given amount of visual stimulus.

(1)
CR
Chad Rohde
Jan 10, 2018

I couldn't do it. Bet it is hard to find a really qualified person to handle a multiple live video feed scenario.

How many live feeds does one employee monitor in a busy major casino?

UM
Undisclosed Manufacturer #4
Jan 10, 2018

That's not how it works.  Casino Surveillance Operators are performing tasks and watching specific people most of the time.  They don't ask people to watch a certain number of cameras just in case something happens.  

UM
Undisclosed Manufacturer #3
Jan 10, 2018

Interesting side of discussion when looking at gaming surveillance. The big difference is in gaming, the live operator is looking for suspicious or illegal behavior rather than waiting for it to happen. In most facility based surveillance, there is nothing to look for till it happens (generally speaking) and one spends hours looking at nothing. That's where attention span is defeated by technology. 

(1)
UE
Undisclosed End User #5
Jan 12, 2018

Our operators (Medium Casino) monitor between 20 and 32 live feeds out of about 600 available, depending on user preference.  All have their preference, I find I don't think I am paying attention, but I notice things out of place, cash payout done wrong, dealer making wrong movements, etc..   That is not counting the video wall.  That is over 2-3 monitors at the work station.  I can keep going for 8 hours if I want to be proactive.  

 

Edit: If I am really watching someone or something, I have only 3-4 cameras up and all my attention is focused on the event.  But that is not 100% of the time, + doling a review to determine what occurred as it is unfolding. 

(1)
Avatar
Campbell Chang
Jan 10, 2018

1 hr a week is 0.6%

You can probably make a decent guess out of that.

Avatar
Jared Beagley
Jan 15, 2018
Seagate Technology

the optimal use case for our hard drives engineered for surveillance use, SkyHawk, is described (for prioritizing writing massive streams of highly detailed video recording without frame drops) as writing to the drive 90% of the time, and playback the other 10%, which, guesstimating by the other answers here, seems to be more playback than is usually felt necessary, however should help give somewhat of an idea here.

(1)
U
Undisclosed #6
Jan 15, 2018

Is it just me, or does the OP sound like a qualifier that some provider might attempt to use to try and quantify the value of their own product?

i.e. who cares what the % is except those attempting to show how their product is somehow able to 'use' all that 'unused' video?

Avatar
Gert Molkens
Jan 15, 2018
IPVMU Certified

That's exactly the point: why not use all that unused video? What's holding us back to think of smart applications where this data can be used for the benefit?

 

U
Undisclosed #6
Jan 15, 2018

...and there we have it.

I've got nothing against using available (yet unused) data for customer benefit.

I was just attempting to bring you to your ultimate point - which you left out of your OP.

so... mission accomplished.

Also, just because data points can be created does not mean that all of this 'unused' data has any operational significance (vs cost) to most customers.

i.e. 'big data' is mostly oversold in the surveillance industry (imo). (cough PSIM)

analytics solutions certainly have the potential to gather and analyze data that is already being created by video streams, yet is currently (mostly) not being used/analyzed.

But how much of that data is actionable beyond trending what is 'normal' or 'show me people wearing red shirts" applications?

Avatar
Gert Molkens
Jan 16, 2018
IPVMU Certified

I think one of the main/first things would be presence of people in an area. If you look at it at the moment, in many cases there are several devices allready doing/could be doing the same thing such as burglar alarm motion sensor, building automation presence sensor, camera+VCA, ..

U
Undisclosed #1
Jan 16, 2018
IPVMU Certified

Ok, I’ll bite.

What does it really matter what % is currently viewed?

Why is this not just “What new things can we do with video?”

Avatar
Gert Molkens
Jan 17, 2018
IPVMU Certified

That's true but i find that stating the low usage percentage resonates more with potential customers and gets them thinking faster

New discussion

Ask questions and get answers to your physical security questions from IPVM team members and fellow subscribers.

Newest discussions