Decoding the Digital Picture

Surveillance technology has undergone a revolution in the past several years. Systems once made up of low resolution analog cameras hooked up to VCRs and monitors now consist of analog and IP cameras attached to encoders, networks, and ever-evolving digital video recorders (DVR) and network video recorders (NVR). These advances create not only great potential but also problems, including increased demands on unprepared corporate networks and unheard of bandwidth and storage needs. This overview looks at some integral aspects of video surveillance systems to help managers make more informed decisions about the system features that will best meet their needs.

 IP Video

The benefits of moving to digital networked cameras are clear, but most companies can’t afford to scrap their legacy systems. Instead, companies are creating hybrid systems by encoding the analog video to digital and sending that data over the network.

This way, “you can still have all the functionality that you have with an IP camera,” but with far fewer dollars invested, says Dilip Sarangan, research analyst at Frost & Sullivan.

As companies get to a point in the lifecycle of their surveillance systems where it’s time to replace equipment, they more closely assess their options in the digital arena. “You see a cycle of about every two to three years of a technology refresh,” says Brad J. Wilson, CPP, president of RFI Communications and Security Systems, which is a member of SecurityNet, a group of integrators.

Even when they buy new equipment, many companies aren’t yet ready to go digital. Only 20 percent of camera sales were IP last year, according to JP Freeman Co., with analog holding strong at 80 percent.

Many companies are hedging their bets, however, by laying the groundwork for future adoption of completely IP-based CCTV networks. Wilson says, for example, that clients have him pull Category 5 (typically used for Ethernet) cabling to cameras even if they’re not IP and will instead be using coaxial cables initially. That move prepares the networks for the eventual replacement of analog cameras. When it’s time for the switch to IP, the infrastructure will already be in place.

Another way companies transition toward digital and networked systems is to have DVRs that can record from both analog and IP cameras, concurs Bob Banerjee, IP video product marketing manager for Bosch Security Systems.

Compression

When the average security director thinks about the components of a surveillance system, the encoder (often located in another piece, such as the DVR or camera) may not come to mind, but it is an integral piece of any system. One of the most important jobs of the encoder is compression. Compression is often categorized as spatial and temporal. Spatial compression is when software looks for similarities within a single frame. For example, in a photo of a tree against a blue sky, the software will “compress” the similar areas, such as the sky, by recording a shorter description of it, rather than describing each similar blue pixel.

The temporal option encodes only the changes from one frame to the next, as most frames are similar. Temporal CCTV options appeared most popular among those interviewed for this article.

Within the temporal sphere, the favored compression option is MPEG-4, Part 2, which is an improvement over an earlier temporal incarnation, MPEG-2. Where MPEG-2 would yield 6 megabit per second (mps) streams, MPEG-4, Part 2 gives a 50 percent improvement at 3 mps, with full bandwidth and resolution. (Note: mps is a measure of space being used up on the network per second, not a speed, so streams at 3 mps take up half as much bandwidth as those at 6 mps.) Therefore, MPEG-4, Part 2 cuts down on bandwidth needs, says Chip Howes, CEO of Steelbox Networks Inc.

A newer, but still not widely adopted, temporal standard is H.264 (also known as MPEG-4, Part 10), which is designed to be 50 percent more efficient than its predecessor MPEG-4, Part 2, according to Banerjee.

With H.264, “I can get the same quality video at half the bit rate, or, I can get double the quality at the same bit rate,” explains Barry Keepence, CTO of IndigoVision. Keepence adds that this compression standard also doubles the storage capacity, essentially doubling the time it will take to fill a storage medium of a certain capacity. This means that if a system recorded 30 days worth of video before, it can now hold 60 days.

So if it’s that great, why isn’t everyone rushing to install it? As with much cutting edge technology, cost is an issue. In this case, the problem is that companies would need to invest in more powerful—and thus more costly—cameras and PCs to run it.

“It takes a lot more computing power to do H.264 than it does to do the original MPEG-4. And I think most of the industry is just struggling to find the horsepower,” Keepence says. He estimates that cameras and encoders would need four times the processing power to run H.264 than to run other codecs (the industry terms for technology that encodes and decodes data), which adds to the cost. But some companies may find that the bandwidth and storage savings justify the extra cost of the additional processing power.

Keepence’s company spent millions developing its own hardware (rather than software) that uses a large amount of parallel processing to get the necessary bandwidth to do H.264. “But now we have it, we’ll have it forever,” Keepence says. That is, of course, until the next best standard comes along.

In addition to front end processing power, Banerjee says a problem with H.264 is that most PCs are too weak to run it for surveillance video. Keepence estimates that the PC would need 10 percent more power to run H.264 surveillance than another codec. Banerjee says it’s a “no-brainer” that once the PCs powerful enough to run H.264 come down in price, H.264 will become the CCTV standard. For now, however, Bosch, for example, sells mainly MPEG-2 and MPEG-4 Part 2 encoders.

On the spatial side, one compression option is a wavelet. The wavelet allows the compressed picture to become “fuzzy,” rather than “blocky” as occurs under the MPEG standard, according to Banerjee. This fuzziness creates a more pleasing picture under high compression.

 However, while wavelets use a low amount of processing power to encode and view on a PC, they consume high bandwidth. Additionally, it should be pointed out that wavelets are not standards-based but are proprietary, meaning that they will likely not be interoperable among vendors. This might make it impossible to mix and match equipment during integration and might explain why wavelets are not more popular.

Individual frames can also be compressed using JPEG, which compresses the frame, and MJPEG, which are JPEG shots filed together.  JPEG 2000 is a type of JPEG compression used in CCTV. Banerjee says that at lower frame rates, JPEG consumes the same bandwidth as MPEG options, but at higher frame rates, the consumption increases compared to MPEG.

Howes is a proponent of JPEG 2000, because it offers low latency, which is the lag time between what the camera sees and when you view it remotely. Howes explains that with JPEG you are receiving every single frame or picture, not just the changes between pictures, which can contribute to latency in MPEG and other types of compression. 

Additionally, JPEG 2000 can provide multiple resolution streams. Therefore, you can look at the video in either full or partial resolution. Howes adds that JPEG is cheaper than MPEG and H.264 because it requires less processing power.

Howes’ company offers a router-type box (digital matrix storage switch, or DMSS) that works particularly well with JPEG compressed data, and he admits that “if you don’t have a box like the DMSS then [JPEG] is not going to be as beneficial,” because it does take up more bandwidth on a normal system.

Whatever the compression choice, the system’s performance should be tested out in person, says Keepence. “H.264 doesn’t mean you’re buying a good product. MPEG-4 doesn’t mean it’s a good product. All it means is it meets some standard,” Keepence says. “It’s the implementation of the standard that makes the difference.”

 Frame Rates

As much as cameras have evolved, the concept of frame rate hasn’t changed. Frame rate is the number of frames recorded or displayed per second, and it’s one of the critical issues that must be considered. A lower frame rate will ease bandwidth demands and reduce storage needs, but it will also reduce the video quality. How much that matters may depend on the video’s intended use.

Broadcast quality frame rate is 30 frames per second. In CCTV, frame rates can be decreased per camera to allow a network to handle the feeds from more cameras within the limited bandwidth, says Roy Bordes, president and CEO of Bordes Group, Inc.

For many applications, it would be okay to go as low as 7.5 frames per second, but if you are using software to do certain analytics on the video, you’ll want to record in higher frame rates so that the algorithms will have more to work with, says Wilson. It’s also important to note that the compression rate and the frame rate together affect the total data loss per image.

It’s possible to set different frame rates and different compression standards for times of day or specific locations or certain conditions, such as when an alarm goes off.

Many companies are now offering load balancing in their products, and that includes frame-rate reduction. For example, the cameras/encoders may be transmitting a low frame rate for general surveillance but when an event occurs or something changes in a scene, that rate might bump up to 30 fps.

Resolution

Resolution affects CCTV systems in several ways. The most direct relationship is that the higher the resolution, the better the picture quality. But there are implications to this simple statement, such as if the resolution is high enough, as with megapixel cameras, one camera may be able to do the job of several, because operators can digitally zoom in on specific quadrants without the picture becoming too pixilated (more on this later).

There are basically three camera options in the CCTV world:  analog cameras, nonmegapixel IP cameras (which, according to JVC security division engineering manager Adrian Parvulescu, refers to cameras that produce similar video quality to analog cameras but have an Ethernet connection and include a computer that encodes the video and can serve other IP functions), and megapixel cameras.

Analog cameras are restricted by the video standards in each country (for example, in the United States the cameras are covered by NTSC standards), while digital cameras are independent of those standards, which allows them to keep becoming more powerful. Thus, 4CIF, sometimes called full TV frame, is the maximum digital resolution (equivalent to 704 x 480) when analog signals are converted to digital. (CIF, which stands for common interchange format, is the digitized version of the television standard for resolution). You need 4CIF for good facial detail, but each level of CIF increases storage costs.

Digital cameras can far exceed 4CIF, going into the realm of megapixel cameras, where the horizontal by vertical resolution equals one million or more. Companies can now find from one, two, and three megapixels all the way up to the 21-megapixel camera called the SentryScope, by Spectrum San Diego, Inc.

Megapixel cameras offer the ability to zoom in for greater detail in wide areas. For example, in a parking lot shot, if an end user wanted to digitally zoom into a picture to get a license plate number, it would be more effective to do so with a megapixel camera because there is more detail in the picture. With megapixel cameras, “because of the wide screen abilities, you can actually replace about three standard resolution cameras,” Sarangan says.

Having only one camera from one vantage point might be too restrictive, however, notes Banerjee.

It should be noted that many cameras offer an optical zoom (not to be confused with a digital zoom), from a few times magnification all the way up to twenty and more than thirty times. Resolution is not an issue with optical zoom. Parvulescu points out, however, that when the field of view is that magnified, every movement is exaggerated and the picture could become useless if it is too shaky. To avoid this problem, the end user must make sure that there is an image stabilizer built into the camera.

No one can deny the image quality improvement on a surveillance camera with one, two, or more megapixels. However, resolution like that does not come cheap, and you’ll pay in more than just dollars: There are also bandwidth and storage costs with each megapixel. Megapixel cameras take up more bandwidth in part because they use JPEG compression rather than MPEG, notes Banerjee.

The amount of bandwidth needed is the issue during data transmission. Then there’s the question of storage. “Probably the largest issue is resolution versus disk space,” says Clark Cummings, CPP, security director of FirstBank. The higher the resolution, the clearer the picture but the greater the storage requirement.

A company has to weigh those competing considerations in light of its own needs. As with all other aspects of CCTV, the choice of resolution depends on the application. Not enough people understand this concept, says Bordes. “A camera viewing down a hallway will require standard resolution whereas a camera being used for quality control on a product line may require much higher resolution,” he says.

Analog cameras have a range of horizontal and vertical resolutions measured in TVL (television lines), where 640 x 480 might be considered high resolution. Bordes says standard resolution, which he defines as around 330 to 360 lines of resolution, tends to be adequate for video surveillance. He adds that most cameras come with a specific resolution setting, so you have to know your needs when you make the purchasing decision.

Opinion varies widely on the necessity of megapixels in the surveillance field. Many agree that when dealing with critical infrastructure and certain environments, such as chemical plants, resolution should not be compromised. But in other applications, that quality may not be as necessary. “You can get great clarity, but do you really need that for event monitoring?” asks Wilson.

Most end users do not choose the 4CIF option for their video. “This tells me that [nonmegapixel] CCTV cameras are good enough for almost everyone, and that megapixel cameras will see niche sales,” says Banerjee.

Vlado Damjanovski, editor of CCTV Focus magazine and author of the book CCTV Networking and Digital Technology, agrees that there might not be a rush to megapixels for most applications. “There is still plenty of room and life ahead for standard definition CCTV,” he says, adding, “if a good choice of cameras, lenses, and compression is made, the standard definition CCTV can be processed better and quicker (more images per second) than megapixel.” Damjanovski does see the market heading to high definition TV (HDTV) one day as that technology becomes ubiquitous.

One application that may justify or benefit from high-resolution cameras is intelligent video analytics. Experts in this field say that if the megapixels improve, so can their algorithms. Fredrik Nilsson of Axis Communications, for example, says that his company will pursue advanced recognition of objects after cameras get better. “With better megapixel cameras, it’s going to be easier, because with megapixels, you’re going to get more data to evaluate,” he says.

Vidient President Steve Goldberg gives the example of someone 500 yards away. “If you have a low resolution…digital camera, you’re only going to get one or two pixels that cover that person,” Goldberg says. “But if you had many, many pixels, you can actually see if that thing’s got a head and arms and legs at 500 yards away. It’s like having binoculars.”

Storage

Once a company has settled on resolution, compression, and other aspects of the system, there is the question of where to put all the data that is streaming out of these cameras and encoders.

Fortunately, the technology keeps improving for storage systems. In digital surveillance systems, the options range from DVRs and NVRs to the larger-capacity network-attached storage (NAS) units.

Standalone DVRs have been the replacement to the analog VCR, but some experts predict that DVRs will soon go the way of the VCR, as they are replaced by their networked cousin, the NVR. In addition to being on the network, NVRs tend to have more storage capacity than DVRs.

There are inherent benefits to being directly on the network. Data can be managed with more flexibility and sent out to extended drives, according to Jeff Fenton, engineering manager of SFI Electronics, a SecurityNet member. NVRs can also receive input from more cameras.

One downside is that NVRs are completely dependent on the network. If the network is down, so is the NVR.

There is a long list of features that end users should look for in an NVR, says Wilson. For example, it should include basic video functions, such as live viewing and possibly remote pan-tilt-zoom control. Additionally, it can include multiple recording options, such as event, manual, and continuous. It should also provide basic features such as the ability to play back video and audio.

Searching should allow for “smart” searching through the archive and export of images with watermarks and timestamps. NVRs should be able to simultaneously manage IP and analog video streams and have the ability to work with equipment from various vendors.

Sometimes neither DVRs nor NVRs are enough. A company may need more storage, especially if it has many cameras and plans on keeping data for the long haul. For many companies, that means using network-attached storage (NAS), which is basically an array of hard disks. These external devices provide centralized storage that is network accessible. NAS is proliferating with the increase in multimedia data storage needs.

NAS is a good choice for clients who do remote monitoring, because it can centralize data from various recorders and locations, Sarangan says. There is a wide range of NAS storage choices that offer varying total storage capacity amounts, including up to several terabytes.

The main concern for security managers when dealing with NAS is “scalability,” or the ability to grow. “Scalability of NAS is not limited by the number of internal or external ports of a server’s data bus, as a NAS device can be connected to any available network jack,” Wilson says.

Some standard features of the NAS include operating systems, remote management capability, and universal connectivity, notes Wilson. He recommends adding a RAID (redundant array of inexpensive disks) to allow for a drive failure without the loss of data.

The NAS could either be factored into the original system or added on later.

As CCTV merges more with IT, Bosch has employed iSCSI (small computer systems interface, which is a commonly used massive storage option, over IP) RAID storage. It is network-based but communicates directly with cameras and encoders, so it does not rely on the NVR or the network in order to record.

Retrieval. Bordes says advances are being made not only in storage ability but also in the quality of future replay. He cites Hie Electronics, which won Frost & Sullivan’s 2007 North American Video Surveillance Product Innovation of the Year Award. Hie’s TeraStack solution puts together removable storage towers of 125 DVDs which fit into chassis that hold up to 50 terabytes of data.

TeraStack, which runs on 800 watts (equivalent to the amount of energy needed to run a small microwave), is able to store data from up to 64 cameras for 9 to 12 months. In addition, the stored video will all be available for search and replay at high resolution and 30 fps, while other systems are often limited to shorter time periods. 

Some storage media have data management capabilities built in to facilitate retrieval and playback. Another option is the router-type system put together by Steelbox, called the DMSS. 

Steelbox built the product from the ground up. It works between the video source and the video recipient, and it sends the video streams—which can vary in terms of frame rate, compression, resolution, and other factors—from up to 512 cameras into different recipient areas, including NAS. “It’s like video plumbing, in essence,” says Howes.

DMSS substitutes for an NVR and aspects of the video management system. Howes highlights the scalability of DMSS and the number of channels it can deal with. Additionally, he says it allows people to rewind or search any video stream at any point without affecting other aspects of the system or other individuals also viewing video.

Intelligent Video

Intelligent video analytics, which have been on the scene for several years now, take surveillance to the next level. Rather than merely capturing video, the system can be programmed to alarm for certain suspicious or prohibited activities, such as leaving a package, walking the wrong way in a restricted location, or driving too fast.

Some of the analytics that companies are mastering are the identification of left behind objects and tripwire detection. Motion detection, though not technically an intelligent analytic, is also getting grouped into this category.

In general, intelligent video is still an early market. But proponents believe that intelligent video may eventually facilitate a sea change in the way CCTV is used, making it more of a proactive tool than a forensic one. “You could do it [analytics] in real time and it’s providing you actionable intelligence,” says Sandy Jones of Sandra Jones & Co., a security consulting firm. Goldberg says analytics can turn a false sense of security into “true security.”

As with many cutting-edge technologies, however, when purchasing a system that purports to offer intelligent video, it’s essential to sift through the hype.

Many analytics still have too many false alarms to be helpful, according to Nilsson of Axis Communications. “Ninety percent accuracy for some applications isn’t good enough,” Nilsson says. Those 10 percent false alarms add up to consumers not trusting applications.

There are exceptions to that, he says. In some cases, even just 90 percent accuracy is desirable over nothing. For example, he cites tripwire installations in subways that develop alerts for people going onto the tracks. “It’s much better than not catching anything today, because most of the time, no one is watching those cameras,” Nilsson says.

A trend cited by several integrators and others in the CCTV world is the continued move to put analytics at the “edge,” near or in the camera or at the encoder. The other option is to have analytics conducted at the “back end,” such as at the NVR or DVR.

There are several advantages to having analytics in the camera. First, they would be working on uncompressed video, giving the analytics’ algorithms more pixels with which to work. Second, the move of the analytics to the front end allows for less processing and storage power needed on the back end, according to Fenton.

Essentially, the decision of whether the footage is important can be made before the video is streamed over the network for live viewing or for storage. “There’s no point in recording 23 hours and 55 minutes of video with nothing going on,” says Alan Matchett, CPP, of Johnson Controls.

 Vidient, for example, has a standalone device that is installed on the camera side of the network and carries out compression and analytics, among other functions. When an alert goes off, selected clips of live video can be streamed over the network, rather than all of the video. Thus, the system can serve as a real-time security tool but not take up as much network space as it would if all of the video had to be streamed constantly over the network.

There are some drawbacks to having analytics in the camera rather than in a “centralized architecture” back at the server, however. For one, each camera would need to be fitted with the software, and you would need licenses for all, according to Sarangan.

Additionally, although prices of intelligent analytics are noticeably decreasing, analytics do still increase the price of the camera, sometimes by hundreds or thousands of dollars. “I don’t think everyone can really afford to have every camera with analytics on,” says Sarangan. “They might pick the one door or the one point of entry or one location where they have noticed that there’s the highest crime or highest threat.” Sarangan says with that approach, some cameras are left out. In contrast, systems that use centralized analytics, all the data will be reviewed.

And most importantly, some experts just don’t think the cameras have enough processing power to adequately perform all the analytics a client might request.

One compromise is a combination solution that includes analytics in multiple spots, starting at the edge. Howes calls this the “winning” solution. He thinks optimized analytics will be done at the edge and then “as new analytic capabilities come along, they’re going to be perfected in servers.” 

An example of analytics that Howes believes must stay server based for now is any type of facial recognition. To begin with, such biometric applications require the use of databases for comparisons. Additionally, certain analytics require more processing power than what is generally available in camera technology today. Some companies have moved to fix the processing problem, however. Goldberg says Vidient has added parallel processing power to its device.

Banerjee agrees that some analytics require too much processing power for cameras, but he argues that the market for those analytics is small.  Banerjee adds that his company created cameras that can handle such applications as objects left behind or removed and people or objects moving in the wrong direction.

The capabilities of IP-connectivity, storage systems, and intelligent analytics will continue to evolve in the coming years. The challenge for security directors, as always, is to assess which advances in technology offer enough of a boost in system performance for their own application that they merit the investment.

---

Laura Spadanuta is assistant editor at Security Management.