I have vocally resisted the IP (Network Camera) trend for ages.
Why I have never preferred IP Cameras
One of my major concerns has been that of latency.
The TCP/IP protocol that is used by network cameras is inherently designed to introduce noticeable delays in the video display. The strict security and data integrity requirements of the protocol – which require transmission notifications, manifest publication, receipt notifications, manifest confirmations, and rechecking fail safes, between devices, are what introduce the delays.
The impact of Latency on mission critical applications
What does this mean in the real world?
You may be observing armed robbers exiting their vehicle on the forecourt on your video display, at the exact moment that they actually burst through your office door.
This is not good for life critical or mission critical applications – where real time observation of fast movement is required.
There are so many things causing latency – it’s hard to manage.
You can find reams of white papers across the net, dedicated to the discussion of managing latency in IP video networks. There are a wide range of factors that contribute to latency in these networks – it’s a major project, and can be quite expensive to get everything fine tuned.
- Capture Rate Latency – Set a camera to 30fps capture rate and you have a 33.3ms capture latency,
- Image processing – a number of image processing tasks occur throughout the process – each introducing it’s own latency,
- Image Compression – introduces further latency,
- Image buffering,
- Audio,
- Network and Infrastructure capacity,
- Amount of video streamed,
There are many more points of latency in such a system. When added together, it isn’t hard to quickly get to 100ms latency.
Ask a casino table manager – to determine how significant a 100ms delay is.
Managing latency demands performance trade offs
If the surveillance objective involves a high risk process with high consequence – investing in such a system may be worthwhile. Latency however will hardly ever be completely eliminated in these systems.
Managing network video latency involves many trade-offs. Reduced resolution, Lower bitrate, Lower Compression Rate, Lower Frame Rate, less secure protocols like UDP and RTSP and more.
Why have sales people hyped up IP Camera technology
For years, salesman have hyped up the advantages of IP video surveillance over HD Analog surveillance.
70% of the claimed advantages don’t exist.
Why would they want to hype the technology?
Mostly because an IP camera system will cost three times as much as a comparable HD Analog System. There is more money to be made.
In some cases visionaries may have foreseen the eventual power of IP technology.
The advantages of IP camera networks.
There are of course advantages;
- Wireless systems can be deployed where it’s too expensive to run cable,
- IP Camera systems can be expanded, and scaled from a single camera to hundreds – using the same infrastructure.
Recently, one advantage of IP camera systems has emerged. This advantage is responsible for forcing me to consider IP camera systems more frequently than before.
No, the concerns about life safety and mission critical applications does not go away. We may just need to spend a lot more money, time and effort on getting the infrastructure as optimal as possible, and in some cases deploy a hybrid – IP /HD Analog System.
How Artificial Intelligence is driving IP Camera Adoption.
The major advantage which is causing this refocus – is that an IP camera is literally a mini computer or server. IP Cameras have CPU capability on-board. They are now outpacing HD Analog systems in the area of compute power.
Compute power is the essence of Artificial Intelligence.
Modern surveillance technologies place a heavy burden on compute power. Artificial Intelligence is also the juice that makes modern surveillance systems so effective and beneficial.
Now we are forced to consider IP Video
I have to acknowledge that the IP camera topology now seems the most preferred and convenient method for deployment of advanced surveillance systems.
It is possible to build powerful HD Analog systems using PC and Server hardware. I would prefer this method and will provide an article with more details -later on.
For now – If you are deploying a modern system that will need to accommodate future development sin AI – IP Network CCTV may be the way to go.
