This is a highly debated topic and one that must be weighed carefully. There are many factors that have to be taken into account when choosing which technology to pursue. As such, we can examine the Pro’s and Con’s regarding this technology choice and the corresponding infrastructure that must support it. First, it’s not just about the camera selection; it is also about the supporting infrastructure. Agreeably when you consider multi-megapixel cameras, then the capabilities (and limitations) of the cameras themselves come into the discussion as significant factors, but infrastructure must be realistically examined. This paper emphasizes the impact on the infrastructure.
The other major important consideration is cost, of which there are two variables. Cost to implement and cost to maintain. The one aspect of both technologies is that the cost continues to go down as the value goes up. There is research and development going into both technology camps, although it could be argued a greater amount is going towards IP camera technology.
This paper is a high level overview of the two technologies, their infrastructure, and an overall architecture approach. Things to consider for both cameras are not only resolution, but clarity, especially in low lighting conditions. Relying on manufacture’s spec sheets, which skew the information to make their products look better, or by the consultant for the A&E firm that cut and pastes the spec sheet into the system spec is not the right answer. Low light performance is critical. One big difference between IP vs. NTSC is the image sensor. IP tends to be CMOS and NTSC is CCD; CCD classically performs better (less noisy) that CMOS when there is not a good amount of light hitting the image sensor. This is a hot topic within a hot topic.
Analog (NTSC) Cameras
Let us first look at analog cameras and corresponding infrastructure. The quality has continued to improve while the cost has continued to drop. However, there is a practical limit to the resolution of the technology within the National Television Systems Committee (NTSC) standard. The NTSC standard (and corresponding Phase Alternating Line (PAL) standard in Europe) was developed by the Television industry in the 1940s. The NTSC image standard is 480 Horizontal Lines of Resolution. When the image is encoded (digitized), this typically becomes D1 and most common resolution is 4CIF. The maximum number of pixels is limited to the number of TV lines available. The image quality is very good and this is the preferred technology for low light conditions. The cameras are becoming more feature-rich with DSP technology embedded in the cameras allowing more settings to be considered which affect lighting, alerting, tracking, etc. Basically, if it’s a NTSC camera, it works with virtually any back-end system as it is truly a standard. As such, interoperability is not an issue, although agreeably some of these camera- unique features may not be directly supported by the back-end system (but are still supported indirectly). Encoding can take place in an independent encoder (located adjacent to the camera or to the server), or via an encoder card in the Hybrid or Video Management System (VMS) server. The frame rate of the video image is real time, 30 fps and is only limited by the encoding process on the back end. Storage costs are traditionally lower than in an IP system. These cameras typically work or they do not work and have a five + year realistic life cycle if you purchase a reputable brand.
However, that is the limit of the NTSC camera. It is normally not remotely serviceable or maintainable, or even configurable. If there is failure, to troubleshoot typically requires a site visit. The only way to assess the health of the camera is to view the video. For post-event analytics, limited to CIF-4 image size, the camera image tends to become pixilated rapidly on a digital zoom function.
Digital (IP) Cameras
Now we can look at IP Cameras. Although the price of IP cameras has dropped significantly, they still are at least double that of their analog counterparts. This price gap increases as the image quality of the camera (in Megapixels) increases.
Interoperability is a major factor when dealing with IP cameras as each VMS vendor supports a limited (but growing) number of cameras and also a limited (but growing) number of features within the camera. Frame Rate diminishes as image resolution increases and storage costs increase significantly for this variable. IP cameras are typically poor performers at night but this is being improved. The lower the resolution, the better the performance is at night. IP Cameras, by nature, are computers and as such, traditionally can have more features over their analog counterparts. As with anything else, many times these features are ignored or unused and as you do use them, you take away something (typically processing power) from somewhere else.
There are two schools of thought regarding the intelligence of video systems. There are IP Camera manufactures that believe the intelligence should be pushed out to the edge – i.e., camera, and the camera should perform the image processing, management, etc. The opposite thought is to have a classical centralized management approach where the video is processed and managed by a VMS (or similar) system. There are Pro’s and Con’s to both however it is the belief of this author that the systems management of the architecture is more feasible and achievable in a centralized VMS type architecture. You have one place to manage and maintain the system vs. a more distributed model. This also allows for easier management of cameras of different manufacture and capability. One of the biggest operational advantages of IP cameras, specifically megapixel, is the post- event analytics. Because the resolution is so great, digital zoom after the fact is effective, whereas the NTSC counterpart (limited to VGA) creates a pixilated image that becomes rapidly unusable. For practical purposes, I have created two categories of IP cameras, megapixel and non-megapixel. Non-megapixel cameras tend to adhere to the Video Graphics Array (VGA) standard developed by IBM.
High-definition television HDTV Standards
HDTV provides up to five times higher resolution than standard analog TV. This standard is defined by the Society of Motion Picture and Television Engineers (SMPTE) and the most common formats or standards are SMPTE 296M and SMPTE 274M. SMTPE 296M, also known as HDTV 720P defines a resolution of 1280x720 pixels with high fidelity color and 16:9 format, using progressive scanning. SMTPE 274M, also known as HDTV 1080, defines a resolution of 1280x720 pixels with high fidelity color and a 16:9 format. Depending on the letter at the end of the standard, it uses either progressive scan or interlaced scanning.
One important characteristic is the standard is based on square pixels – the same as computer screens so this means it can be displayed on computer monitors or HDTV screens. This technology, as applied to digital video surveillance, is just beginning to make its mark. As somewhat of a pre-requisite, the compression standard H.264 was needed to develop and mature in order to make HDTV transmission feasible. This is an area to keep a watch on as the industry applies this technology to the surveillance world.
The supporting network, power, storage, and management is possibly more the driver to determine which architecture and camera type to choose. Many IP camera manufactures claim one large advantage is you can use existing networks and infrastructure. Although this may be the case in many instances, it typically is not. Most IT data networks are fairly well populated, and usually were built around 100 Mbps horizontal technology. Additionally, locations for data drops and IDF/MDFs typically don’t take into account security camera needs. Finally, the nature of the data traffic of video is very different that traditional interactive network traffic. Most organizations use their data networks to perform routine administrative and other mission- critical functions, such as email, web hosting, internet access, HR application systems, etc. The majority of these systems create bursty, interactive, unbalanced network traffic.
For example, a typical web query will have a small amount of data (query) initiated by a user resulting in a large response (query result); yhe time between the query and query response is typically governed by the web site serving the data – and the impact on the network depends on the content returned. However, most web serving delivers content in a uniform, managed, way. Small or normal sized frames of data are delivered onto the local area network but are usually limited by the wide area network bandwidth. The impact on the local area network is small. Even when this is multiplied by a larger number of users, the WAN limitations still define the bandwidth. Inter-LAN traffic, such as email, is also bursty and interactive in nature. Once a user hits ‘send’, they don’t have visibility into what goes on across the network. Their response is governed by the other user’s timeliness of response and for the most part, accepting of whatever it is. Database access is similar to the web query. Video, by nature is very different; different by how the user sees the video and by the impact to the infrastructure. Users notice choppy video, they notice low frame rate, they notice long seek times on archived video. This is similar to Voice Over IP (VOIP) where distortion, delay, and dropped conversation are very noticeable. Users don’t notice a network delay on a sent email as it is out of sight, out of mind. What this means is your existing network must be carefully evaluated to determine what the impact of putting video across the network would be. You must take into consideration the traffic from:
• Camera to Server
• Server to Server
• User to Server (viewing)
• User to Server (viewing archive)
• Systems Management usage
So what is being said here is unless you have a Gigabit modern network being under utilized with drops everywhere, then you should build a separate network for video. Agreeably things like VLANs can help isolate traffic along with Quality of Service (QoS) attributes assigned to application level traffic. But the reality is depending on the size of the video surveillance system you are putting in, you may have to consider (or should consider) a new network dedicated for the video. Infrastructure costs can be the drivers in decisions regarding IP vs. NTSC. Assuming you have to introduce a new data network for video, then there are typically more upfront costs for installing an IP video network over a NTSC network. However, depending on the size of the deployment, there is a point where that cost crosses over and favors the IP architecture. Once a scalable, modern data network is installed and operational, then expanding to new cameras can also be significantly cheaper. If the original network takes into account the geographic boundaries of the facility being protected, then once the initial costs into installing the baseline MDF/IDF infrastructure are made, then adding a camera or two to a nearby IDF becomes significantly less.
Operations and Maintenance. Too many times the costs and effort to design and install a system are focused on and the costs and effort required to operate and maintain a system are overlooked. This is possibly due to the nature of how digital video surveillance systems have evolved.
Figure 1: Video Surveillance Evolution
Digital video surveillance systems have evolved from two classically independent sources. The traditional source was an evolutionary outcome of analog CCTV that started back from tape- based time-lapse surveillance systems. The skill sets needed to deploy these systems were basically wiring. CCTV companies rose from the cable TV/Coax wire skilled people. There were no intelligence in the systems and not a lot asked of the installer. As technology progressed to digital, the infrastructure did not. No real additional demands on the installer were made. Systems had few options therefore setup was usually a simply 1-2-3 step process. As DVRs became slightly more sophisticated, then a handful of the installers needed some proprietary DVR training, but again, that was it. Due to the nature of these systems, the demand for professional installers for these systems is shrinking as they become more end user-installable.
The other source for this technology has come from the IT industry. Intelligent network cameras with sophisticated features, server-based VMS system hosted on different operating systems and most importantly, an IT infrastructure holding the whole thing together created the demand for skilled IT professionals with strong networking, server, and systems management backgrounds. With that said, systems management is a significant differentiator between the two competing technologies. In one respect, analog NTSC systems do not require any significant Systems Management as defined from the classic IT Systems Management. These systems are too basic and un-manageable. And for the most part, they are fairly reliable in that there is not too many things to go wrong. IP-based systems, however, follow classic IT Systems Management requirements in that there needs to be a fair amount of instrumentation, monitoring, and systems maintenance to properly maintain the system. This component of the architecture, and its cost, is usually overlooked, especially with designs from the classic CCTV side of the house. Change Management, Configuration Management, and Systems Management are just a few of the critical IT processes that need to be considered when designing and installing a system of this complexity. If you’re joining an existing IT infrastructure, even with a new dedicated network, then these processes should already be in place and just need to be applied to the new components. If not, then depending on the complexity of your system, they need to be factored in.
Using the System
Another often overlooked factor is by whom and how the system will be used. Passive Surveillance vs. Active Surveillance dictates two different levels of intelligence of the system and may also dictate other architectural components. Aside from Pan Tilt Zoom (PTZ) vs. fixed, they may dictate the nature of the storage infrastructure, the viewing/monitoring components, and other system elements. For example, alerting and alarming may become more important in passive surveillance than with active.
There are active surveillance applications where the user really does not interact with the system at all. Archive reviews are reserved for specialized personal after the fact and the majority of the system is for real-time safety monitoring (as in a prison). Conversely, passive surveillance may be more used for review –only, upon incident, without the need for robust real-time monitoring and interactivity (as in a car dealership). These differences in how the system is to be used may also determine the level of sophistication – and capability – that is needed in the system.
For those who jumped ahead to the Summary section for the answer, I apologize for the subtle disappointment. The answer is it depends on your circumstances. There is not a one size fits all when selecting the two competing technologies, or the sub-technologies within each architecture. It really does depend on the specific needs, and priority of the user. It also depends on the sophistication of the people who are going to use it. We recommend not limiting yourself to either technology. Digital video surveillance solutions should be designed to easily accommodate multi-format cameras, legacy components, and a varied means of storage. We believe in a loosely coupled architecture, where each architectural component is independent of the next.
Figure 2: Loosely Coupled Architecture
For NTSC Camera systems:
• For smaller systems, can be significantly less expensive to install and maintain
• Can be installed by less sophisticated installers
• Are limited in features
• NTSC Cameras outperform IP cameras in low light conditions
• Require less (if any) systems management
• Typically cost more to expand than IP, once initial costs are laid out.
• Tend to be closed, proprietary in nature
• Are very limited on post-even forensics
• There are few standards for infrastructure
For IP Camera systems:
• Can be more cost effective to install large scale systems
• Typically require more trained and sophisticated installers
• Are feature rich
• IP Cameras, especially megapixel, are poor performers in low light conditions
• May require ample Systems Management
• Once the infrastructure is in place, can be more cost effective to expand
• Are typically open systems in nature,
• Can have excellent post-even forensics
• Can have more intelligence (alerts and alarming) built into the system
• Are easier to remotely management and maintain
• Infrastructure follows classic IT standards
Finally, anticipating when your system actually gets deployed may make a difference in costs and available technology.
Since this is such a rapidly evolving technology, knowing where vendor’s products are headed is important. If you must first design and build a network infrastructure that may take several months, then holding off on camera selection may have it’s benefits.
©Copyright 2017 - MTS Intelligent Surveillance Solutions. All Rights Reserved.