The discussion among security and surveillance video manufacturers, systems integrators, and end-users about the relative advantages and indications for different kinds of video installations is easy to oversimplify. Simplification aids end-user decision-making and furthers the business interests of equipment manufacturers and systems integrators. Industry writers also have good incentive to simplify information. The many motives include making a larger point, supporting a specific-case argument, and getting and keeping readers' attention.

Simplification most often leads to statements like "IP video is more expensive than analog video," which are true with so many exceptions that they are not actually true at all - studies sponsored by interested parties have shown the opposite to be true, and editors and bloggers have covered the studies' findings widely. In fact, oversimplification often leads to an assertion and its opposite both being true.

Another manifestation of oversimplification is the argument that each installation is so unique that no useful rules-of-thumb can be developed.

In this paper we provide as simple as possible an explanation of the factors which indicate analog, IP, and hybrid IP/analog video systems respectively. This paper seeks to provide a reference for editors, end-users, and integrators who may need to evaluate a specific case or understand the general principles.

Indications for choosing a pure IP camera system

Ability to use an existing IT network - In some cases, digital video video systems with IP cameras can be plugged into the existing IP infrastructure. And other times the cost for the upgrade to the network to make it viable for video is easily managed. A user with an existing Ethernet network which is able to handle large amounts of data, who just needs a few cameras or plans to record low frame rates or low resolution, is a good case for a pure IP system. Recording at the "edge" of the network with only occasional requests for video over the main network may also provide a way to implement IP video on existing infrastructure.

Quick and easy data protection - In many mission-critical recording environments, lost or missing data is not acceptable. IP systems can shorten response times and speed up DVR reassignments. In cases where a DVR goes offline, the user can use software to reassign the camera to a different DVR without making any wiring changes. Response times recording problems are also fast, since the recording systems are monitored at the IT network management level alongside all the other servers, routers, switches, and network applications.

Ability to move and add cameras easily - Users with fluctuating camera counts and locations can avoid the need to power down recording servers to add, move, or remove cameras. This means seamless, pain-free recording from existing cameras, rather than scheduled downtime with alternate-server recording or missed recording.

Ability to collect megapixel images - Users who need really high-resolution imagery want the ability to selectively deploy network cameras providing images with resolutions at least four times higher than analog images - which means much more detail than an analog camera (which is limited to standard image dimensions that do not exceed 704x576 pixels (PAL) or 720 x 480 pixels (NTSC) image after the signal has been digitized in a DVR or a video server). Megapixel cameras can collect so much information that you can zoom in to catch the smallest, subtlest detail of a video frame. Megapixel IP cameras can provide superior, unambiguous images of point-of-sales transactions and other events that can be used for business intelligence, loss prevention, and security.

Ability to use facial recognition analytics - Facial recognition software depends on high-resolution images to be effective. Any camera being filtered for facial recognition should have more than the 704x576 pixels (PAL) or 720 x 480 pixels (NTSC) provided by an analog camera.

Minimum disruption and installation expense - Even in cases where a new Ethernet network needs to be installed to handle video traffic, pure IP surveillance systems are less disruptive to install than their pure analog counterparts. Unlike IP systems, analog systems require the installation of a direct coaxial, Fiber, or UTP cable running from every camera to a DVR, as well as additional encoding hardware to be installed on the DVR itself. IP systems can also distribute the power and HVAC loads to help users avoid expensive and disruptive site modifications to HVAC and other site features.

Need for video transmission over wide geographical range - Putting video on the IT network makes it possible to use switches, hubs, and routers to expand the network to a broader range. Analog cameras have significant transmission limits over wide surveillance areas, and they are not appropriate for some wide-area installations due to their need to be physically cabled to a DVR.

Need for advanced features like digital zoom, which are not available in analog cameras - Many new IP cameras have on-board encoding and analytics as well as sought-after features that certain users need to successfully implement their surveillance plan.

Need for camera-level redundant recording - Some IP cameras can provide redundancy by recording onto built-in memory cards.

Indications for choosing a pure analog system

Analog can meet user's recording needs for a lower cost - There are many lower-end, low resolution IP cameras that are low in cost. But high-end IP and megapixel cameras are very expensive, and the disk capacity required to store the higher volume of video data is a very significant increased expense. Across a network array of hundreds or thousands of cameras this cost can be prohibitive. A large network of IP cameras will usually require the installation of a separate network so traffic doesn't exceed bandwidth. An installation of just 40 cameras of 1000 Kbps-1 Mbps each will overtax many existing corporate networks. Many high-end analog cameras, though limited in resolution to 704x576 pixels (PAL) or 720 x 480 pixels (NTSC), use image processing, automatic back focus and imager sensitivity to produce images that are superior in quality and clarity to a similarly or higher-priced IP camera.

The cost advantage of analog recording is especially true for users who already have a legacy coaxial or UTP wiring in place and do not need to install a large number of cameras (analog cameras require a lot of wiring, which can be expensive and disruptive in terms of site modifications).

Network traffic exceeds the user's existing capacity - IP camera recording and viewing will increase network traffic, especially with lots of IP cameras or with or megapixel cameras. If recording will exceed the existing network capacity, a user will need to add the costs of installing an additional network for the video to the operational and equipment costs of pure IP when comparing it to analog and analog-IP hybrid options.

Ability to avoid hardware upgrade expenses - IP cameras tend to rely on the processing power of the CPU. Many analog systems use additional processors to share the video processing with the CPU. But an IP camera stream comes into the network port and requires the CPU to be recorded and viewed. This limits the number of IP cameras that can be added the load the server systems are able to process. These factors depend on bitrate and video encoding format (MJPG, H.264, MPEG). And most IP cameras send large files (MJPEG) to the server to process and store. These files provide a good image but are very large and consume large amounts of storage very quickly. Newer compression formats on the horizon for IP cameras will help address this issue.

Need for minimum latency - Latency is defined as the time it takes for an image captured at a camera source to be presented to a system user. Every camera has latency to a degree, because the data travels from one location to another. It is much greater for IP systems, however, because their signals need to be encoded at the source and then must travel through the network to the decoder in order to be presented to the user. Because analog signals are point-to-point (camera-to-monitor), they don't have the additional latency caused by network routing and the encoding and the decoding process. Because of this they are typically preferred in industries such as gaming and corrections.

Need for system to suit staff abilities - Many businesses operate without complex or extensive computer systems and do not have a need for the dedicated IT professional(s) required to provide timely and effective response to network emergencies on a system of any size or complexity.

Wider camera variety and choice - There are a large variety of Analog cameras (for instance, mini covert cameras and pan-tilt-zoom cameras in various sizes and shapes) to choose from. With IP cameras, not every vendor has many varieties and not every vendor's surveillance software supports others' cameras. In many organizations, physical security staff takes care of the surveillance system administration as well as the overall security plan design and implementation, and there are no existing IT needs that require IT professionals.

Vendor relationships and support - Vendor relationships have the power to greatly enhance or greatly damage the user experience and dissolve the integrator's margin. Integrators and users at installations that already have analog cameras have an existing relationship with the manufacturer, so support and trust in the product are already in place. If the camera or DVR manufacturer with whom a relationship exists doesn't have/support IP cameras, a new relationship of trust and access to support must to be developed.

Indications for choosing a hybrid IP/analog system

(A hybrid system will provide many of the advantages of the pure systems on a per-camera basis. The following are true only of hybrid systems.)

Ability to add IP camera recording to existing investments in analog - Where there is existing analog infrastructure, the hybrid approach allows the user to avoid expensive replacement of their existing analog cameras and wiring. Instead, the user can add IP cameras to their surveillance resources and record from both analog and IP cameras in the same DVR.

The right camera for the each site - Hybrid recording allows the user to choose analog cameras or IP cameras according to the recording requirements and conditions of each camera site. A site requiring the lowest possible latency will call for an analog camera, whereas a site using facial recognition or other data-hungry analytics will need an P camera source.

Minimal retraining expenses - Surveillance operators familiar with an existing analog-system user interface will be able to manage new IP cameras without disruptive and expensive retraining. Using hybrid DVR/NVR recording allows integration of IP cameras with little change to the end user's normal routine.

Abigail Hamilton
Jun 25th, 2008