How to improve analog video using adaptive equalization in multi-channel surveillance systems
By: Chris A. Ciufo, Editor-in-Chief, Embedded, Extension Media Publishing
Despite the trend towards IPTV and VoIP, most CCTV surveillance cameras are analog. Signals degrade quickly and quality suffers. Here’s the fix.
Despite the trend towards IPTV and VoIP, most CCTV surveillance cameras are still analog. Signals degrade quickly and quality suffers, but adaptive equalization (AEQ) algorithms at the receiving end can recover signals despite long, often interrupted cable runs. DSP-based AEQ ICs with built-in A/D video decoders are ideal in multi-channel DVRs and eliminate costly analog amplifiers.
According to Security InfoWatch, a surveillance industry website, 2005 was supposed to be the year when IP-based security systems surpassed analog CCTV. It didn’t happen, either because of the Great Recession, the lack of enterprise infrastructure for digital packets, or plain old inertia at rest.
Data from analyst firm IHS predicts IP CCTV surveillance system installation revenue is expected to equal that of analog for the first time in 2014, but the surveillance world remains 70-80 percent analog, says Pericom Semiconductor. The percentage of analog is even higher if one factors in mobile installations like on-bus, in-train, and budget-based taxi and police cruiser cameras. A common denominator in all of these analog systems? Low cost and simplicity.
A typical single-channel analog surveillance/recorder system is shown in Figure 1. Analog cameras have improved dramatically with the addition of stereo audio, higher resolution, and even in-the-camera signal preconditioning. But the connections to the receiving display or video recorder via long analog cable runs always end up the same way: with lousy, lossy, signals.
Figure 1: Typical analog video surveillance system showing signal degradation after a long cable run. As we’ll see later in the article, adaptive equalization (AEQ) will clean up degraded analog signals.
Attenuation, noise and losses occur at lengths greater than 100m in even the best cables. But with low-cost COAX in low-budget installations, signals degrade at 1.6dB/100 feet (1.6dB/30.5m). Typical signal loss for several cable types is shown in Table 1.
Table 1: COAX cable losses. (Source: Google; various industry reference material.)
Note that at the receiving end in lieu of the digital recorder, a monitor might be substituted—as in a security monitor station—such as that for viewing a bus’s back-up or door monitor camera. In all cases, analog signals from the camera degrade with distance. It’s easy to imagine long cable runs in a parking garage installation, but in a mobile application like a train, bus or even a police cruiser, cable runs are tricky.
Routing is rarely direct from transmitter (the analog camera) to receiver (the recorder or display). Even in a vehicle, cable is snaked up door posts, under floors, across roofs, and often through multiple attenuation-causing connectors passing through actual firewalls, waterproof bulkheads, or flexible couplers (Figure 2). Clean analog source signals degrade with distance due to the cable’s channel effect. They’re attenuated by connectors, and in-camera or in-receiver circuit board traces. As well, analog signals are susceptible to EMI that induces competing signals on the cable and impairs the original waveforms and transitions.
Figure 2: Close-up of the articulation joint and coupler of the typical two-carriage city bus. Cables are routed wherever there’s space; adding connectors here can add noise and degrade analog signals. (Courtesy: Wiki Commons.)
Better Cameras; Weakest Link
As surveillance grows worldwide, camera manufacturers have focused (no pun) on increasing image quality by factors of four or more. As shown in Table 2, the CIF standard (360 x 240, NTSC) was improved by 4x when D1 (720 x 480, NTSC) came along. The WD1 (wide D1) standard added 34 percent more pixels (960 x 480, NTSC).
These high-res analog cameras add up to more information per frame at 30 fps, which means even more waveforms to stuff down long cable runs before termination at the receiving end. As before, the cables and connectors are the weakest link in the channel. Figure 3 shows the dramatic improvement in video realized by camera improvements.
Figure 3: Various standard camera resolutions show readability and field of view improvements. However, each successive improvement pushes more data down the same analog cable.
The result of stuffing data down analog cables, says signal integrity experts at Pericom, looks like the patterns shown in Figure 4. The left-hand image is fuzzy as a result of analog signals traveling across long cable. Although simulated for a 500m run, similar results occur in short runs with low-grade, low-cost cable, or in high EMI environments like factories. Similar results are seen where multiple cable connections are required transiting bulkheads such as in automotive or transportation applications. Simply stated: analog camera signals degrade in all but the most ideal installations. What’s the solution?
Figure 4: The test pattern on the left is noticeably fuzzier as a result of a 500m analog cable run. Adaptive equalization at the receiving end cleans up waveforms, resulting in dramatically improved surveillance video.
Getting Amped Up
With so many analog cameras around us, the solution to signal loss is often to add expensive analog amplifiers prior to the receiver display and/or digital video recorder (DVR). Amps do a good job of increasing the gain and improving video quality at the receiving DVR or console.
But it’s ironic that what starts as a low-cost surveillance system requires video amplifiers that cost between $5 - $10 per camera channel. With typical installations at 16-32 camera channels, the cost adds up to serious money in what was supposed to be low-budget surveillance. Clearly there must be a better solution.
DSP-Enhanced “Candid Camera”
Pericom Semiconductor, a company exclusively focused on signal integrity products, builds in adaptive equalization algorithms into several video decoder ICs. The company claims that proprietary adaptive filters improve signals by 2x, translating to either double the picture clarity or acceptable analog video at twice the usable cable length.
The family of ICs handles multiple analog audio and video channels with myriad programmable video image enhancements. Compared to an analog amplifier, a company spokesman told me, the ICs designed into a DVR receiver will save tens of dollars per video channel—while adding useful video features.
Digital decoders with adaptive equalization (AEQ), shown in Figure 5, do more than improve the gain on signals attenuated over long, lossy analog cables. They also reshape the waveforms (sharpness, hue, saturation, brightness, contrast), improve signal-to-noise ratio (SNR), and can switch to black and white (B/W) on extremely weak signals for extra analog clarity and image enhancement. A block diagram of a Pericom four-channel video decoder with AEQ is shown in Figure 6.
Figure 5: Long cable runs and analog video cameras: bad data is digitally enhanced using low cost adaptive equalization ICs.
Figure 6: One of Pericom’s four-channel video decoder ICs uses digital signal processing (DSP) to perform adaptive equalization (AEQ) to compensate for lossy analog camera cables.
So look sharp the next time you spot a video surveillance camera: it’s possible your smiling mug may be more clearly recorded than you think, thanks to the AEQ signal processing available for plain old analog video cameras.