The Mainstream Migration to 4K Video Surveillance: Back to the Future?

This article examines the migration to HD surveillance over the past seven years to develop a plausible scenario for successive migration to 4K surveillance over the next seven years.

                               (a)                                                          (b)

Figure 1 Higher-Resolution Surveillance Migration Scenarios

In Figure 1:

• “Price” qualitatively reflects the cost of higher-resolution cameras relative to the then-dominant lower resolution formats.
• “Market Share” is the fraction taken from the lower-resolution formats.
• “LST Availability” indicates when specific cameras and DVRs with that Local Site Transport ("LST") have become available.
  • The light green boxes indicate when prototypes are being demonstrated
  • The dark green boxes indicate when cameras and DVRs are sold via distribution.
• Entries for 2016 and forward are projections, shown in italics, while entries through 2015 are estimates.

The two sides of Figure 1 are nearly identical, reflecting an expectation that the market migration to 4K will be similar to the migration to HD. Initially, the higher resolution is adopted only for those applications where the higher resolution is worth significant additional cost. As the technologies that enable the higher resolution reach maturity, the price comes down, and greater segments of the market migrate to the higher-resolution surveillance.

The HD and 4K scenarios shown in Figure 1 differ in one significant aspect, however: Two kinds of 4K cameras — Ethernet and EX-SDI — are available at the beginning of 4K, whereas only one kind of HD camera was available at the beginning of HD.

The technical challenges to mainstream adoption of 4K surveillance are significant. One way to appreciate the scope of the challenges is to consider the key constituents of a high-resolution video channel in detail.    

Figure 2 shows the components lying in the electronic video surveillance pathway from a camera’s field of view to forensic evidence available via the cloud:

Figure 2 Video Channel Architecture, moving evidence from Light to the Cloud, highlighting Local Site Transport ("LST")

The critical semiconductor components, shaded gray in Figure 2, command an ever-increasing share of camera and DVR end-product value. Meanwhile, each of the technical elements shown in Figure 2 must be as reliable, as convenient, and (nearly) as affordable as its lower-resolution counterpart in order for the industry to sustain our customers’ migration to the next higher resolution.

The scope of the obstacles to 4K migration becomes clear by considering the non-LST elements one by one:

CAMERA:

• Lens
  • Higher resolution demands a smaller blur spot (equivalently, a higher modulation transfer function) in order to resolve the smaller pixels.
  • Individual sensor pixel area gets reduced by more than a factor of four in order to double the number of pixels in each X and Y dimension. For a given sensor diagonal size (1/3” and 1/4” being predominant in surveillance), forming the individual lens elements demands ever greater levels of precision.
  • Lens manufacturers advance steadily up the learning curve at each successively higher resolution, enabling continual mainstream migration until the physical limit is reached. The physical lower bound to blur spot radius is the diffraction limit of light.  Ultimately, lens performance is a delicate balance amongst blur spot radius, sensor diagonal size, depth of field, and sensitivity.
• Image Sensor Chip
  • Higher resolution demands a smaller pixel.
  • Pixels do not scale in silicon the way that digital logic scales (as the pixels shrink geometrically, their ability to collect light decreases worse than linearly in the number of pixels, because the (fixed-size, not-light-sensitive) read-out circuits take up a larger fraction of the pixel area at the smaller resolutions.
  • Scaling resolution demands the image sensor achieve greater sensitivity from diminishing physical area per pixel.
  • Sensor manufacturers advance steadily up the learning curve at each successively higher resolution, enabling mainstream migration.
• Image Signal Processor (ISP) Chip
  • Higher resolution demands more pixel-bits processed every second.
  • Moore’s Law scaling directly benefits ISP chips, such that at each resolution, eventually ISP chips have enough processing capability to sustain a mainstream migration to higher-resolution video.

DVR:

• Video Codec Chip
  • Higher resolution demands more pixel-bits processed every second.
  • Moore’s Law scaling, driven by all digital processing applications, directly benefits codec chips.
  • At each resolution, eventually codec chips have enough processing capability so as to sustain a mainstream migration to higher-resolution video.
• Local-Site Storage Solution
  • Higher resolution forensic evidence demands more storage per second of archived video, at a given frame rate and recording quality.
  • Local-site storage solutions benefit directly from the economies of scale driving storage improvements for myriad I.T. applications.
  • At each resolution, eventually local-site storage is cost effective and thereby sustain a mainstream migration to higher-resolution video.

I.T.:

• LAN and Internet Access
  • Higher resolution places more demands on LAN and Internet interfaces.
  • LANs and Internet interfaces benefit directly from the economies of scale driving storage improvements for myriad I.T. applications.
  • At each resolution, eventually LANs and Internet interfaces become cost effective and thereby sustain a mainstream migration to higher-resolution video.

LOCAL-SITE TRANSPORT (LST)

LST is an under-recognized technical challenge area.  Existing LSTs impose severe limits on the ability to scale to higher resolutions:

• Conventional digital LSTs increase clock rate to stream higher-resolution video.  If the receiver ever fails to discern just one bit among the billions sent every second, mission-compromising image artifacts may become visible. To ensure that all bits get through the cable, high-resolution digital LSTs including EX-SDI and Ethernet compress content; that is, these digital LSTs remove bits from the transmitted evidence in order to accommodate bit-rate-limited LST.
  • Unfortunately, even the most advanced compression technology degrades video quality.
  • Furthermore, Ethernet imposes a strict distance limitation which is not overcome by even the most advanced compression technology.
  • Moreover, compression adds complexity (logic gates (hardware) and lines of code (firmware)) to all digital high-resolution video LSTs, thereby reducing reliability while simultaneously increasing design costs and manufacturing costs. Advanced compression technologies generally carry more hardware, more software, and therefore greater costs.

• Conventional analogue LST demands an increased carrier frequency to accommodate higher resolutions. It appears impossible for conventional cables to sustain signals at frequencies high enough for unimpaired conventional analogue LST of 8 MP video. Therefore, it may be impossible to scale analogue LSTs to 4K resolution. Although the analogue HD share of the 2 MP surveillance market has grown from 0% in 2009 to over 80% today, there is no known "next enhancement" that will enable conventional analogue LSTs to achieve higher resolutions.

Personal Opinion and Outlook on the Migration to 4K Surveillance

The earliest HD adopters had reason to presume that “IP camera” and “HD camera” are synonymous. At CPSE 2015, 4K IP cameras were exhibited alongside 4K EX-SDI cameras in neighboring stands, so buyers of highest-resolution digital 4K cameras already see a choice of LSTs. 4K EX-SDI delivers better fidelity, in my opinion, over longer distances than the Cat-5 spec. Moreover, 4K EX-SDI equipment appears likely to be more affordable than Ethernet 4K LST equipment. I expect buyers to choose both kinds of 4K equipment in 2016, and this greater choice suggests that the migration to 4K will be faster than the prior migration to HD.

---

Todd Rockoff bio:

Todd E. Rockoff, Ph.D., is General Manager of Rockoff Security P/L, a provider of semiconductor products for video surveillance equipment.

Dr. Rockoff founded HDcctv Alliance in 2009 and continues to serve as its Executive Director. Prior to the Alliance, Dr. Rockoff was VP Global Sales at EverFocus ODM, VP of Sales - Asia at Pixim, Director of Sales at Foveon, and Founder/CEO of Intensys, a Silicon Valley semiconductor company.

Dr. Rockoff received his Computer Science Ph.D. and M.S. degrees from Carnegie Mellon University, and undergraduate degrees in Engineering and in Mathematics from the University of Pennsylvania.