Waiting for H.265

Jan. 18, 2016
What's happening in the interim?

In 1996, Bill Gates wrote an article predicting that the real revenue to be made on the Internet would be in content generation. As a visionary, he understood bandwidth challenges would eventually be conquered enabling an infrastructure that would rival the power of broadcast television. If you look at today’s up-and-coming security professionals, they’ve probably had more picture and video experiences by their early twenties than most of their employers have over the course of their entire lives. For these young professionals, content has become more than just motion pictures, advertising and television shows. It is a way of life that changed a whole generation’s expectations of their work environment. Interestingly enough, licensing of that very content might be holding back the adoption of H.265 – not only in the security industry, but also in any other industry you can name.

How close are we to broad-based adoption of H.265?

H.264 is the heavyweight champion of deployed video codecs. It holds the lion’s share of recorded video across the security industry. But it took a long time to reach that status. Specified in 2003, it didn’t reach broad-based adoption for another five years. It then took nearly another year for the recording ecosystem to develop support before it really took off in 2009. At first, network camera manufacturers offered products with dual CODEC support enabling security practitioners to deploy MPEG4 cameras initially and then enable H.264 when their recording platform offered a solution. This same scenario will play out again and again until Moore’s Law stops being relevant. Many in the industry, including myself, have predicted that same pattern will emerge with H.265, but at a much-accelerated rate. And we have a five-year window in which to prove ourselves right.

Jointly developed by the ISO/ITU and MPEG, H.265 was officially ratified in April 2013 with an additional version updated in 2015. Though primarily developed for consumer electronic devices, the compelling advantages offered by this innovative compression technology also make it a prime candidate for adoption in the world of physical security. So where are all the H.265 products?

Depending on how you want to look at it, we’re closing in on year three or year one of its ratification.  At ASIS and ISC East, only a handful of camera manufacturers offered H.265 solutions and there was close to zero support from VMS developers. What good is video if you don’t have the means to record it? Not to worry. Those who have H.265 cameras will be happy to sell you their video management system that supports their unique implementation of H.265. Is this starting to sound familiar? It should. It’s the same way our industry approached the adoption of H.264.

But wait a minute. Given our experience with H.264, shouldn’t the repurposing of this next generation of compression standards for the security industry be going a lot faster?  Why is this taking so long?

Wrestling with licensing and royalty structures

One area that has everyone on edge is the licensing and royalty structure proposed by two organizations representing patent pools for H.265. The first, HEVC Advance, was formed in March 2015 and represents 500 patents. Their licensing model includes any combination of encoder/decoder, but also demands a royalty on revenues generated on the content developed by said HEVC-enabled device. This royalty on the content has prompted strong backlash from companies like Google, Amazon and Netflix who have joined others to develop the Alliance for Open Media, which aims to provide a royalty-free alternative to H.265.

At first glance, you might think the royalty portion of the HEVC Advance licensing model would not affect the security industry – largely because we don’t draw revenue from the content we generate. Who is going to sell video evidence of security lapses? But hold on, did you forget the cloud? Video Surveillance-as-a-Service (VSaaS) provides customers with on-demand access to content that will be generated by cameras using H.265. Almost sounds like the basic concept for a private Netflix account of your video surveillance system and it certainly generates revenue for the hosting provider. HEVC Advance has made choosing video codec support a bottom line discussion for companies offering VSaaS solutions.

An alternative is the MPEG LA which collects the licensing fees we have all been paying for using H.264 and previous versions of CODECs managed by the group. MPEG LA uses a more traditional model that only licenses the encoding or decoding. Most companies view this as simply the cost of doing business in the video industry, which enabled MPEG LA to largely avoid the backlash unleashed on HEVC Advance. MPEG LA may lose out if the Alliance for Open Media is able to gain substantial traction in the consumer electronics market. Because what happens in the consumer electronics market tends to heavily influence what eventually happens in security.

Either way, immature licensing models tend to breed confusion in markets, which ultimately slows down the adoption. For an industry that is historically slow at adopting new technology, the physical security market might not be as close to broad-based adoption as we think.

Is the push for H.265 worth the effort?

If we have to involve the lawyers and the VMS community requires a complete re-write of their software to support it, is H.265 worth the effort? Yes it is. With claims of a 50 percent decrease in bitrate compared to H.264, H.265 represents true cost savings especially for security professionals with limited budgets. Implementing H.265 can cut your existing storage requirements in half. Think of what you could do with the additional space and savings:

  • Increase the resolution on your existing cameras to gain more pixels on target,
  • Increase the frame rate of the recordings to increase situational awareness, or
  • Deploy more cameras at the same resolution and frame rate.

All three provide higher performance across your existing investment giving you distinct security advantages.

More pixels on target provide the option of widening the field of view while still maintaining the operational requirement of detection, recognition or identification. A wider field of view provides more coverage and possibly less cameras for a given scene.

By increasing the frame rate operators can review video recordings with more fluid motion providing greater situational awareness as well as the opportunity to “frame grab” the perfect image for possible prosecution. The third option seems inevitable.

Over time, most security operations find themselves increasing camera count as more opportunities for surveillance present themselves. With H.265, that expansion can be accomplished without overtaxing bandwidth consumption and other resources.

Introducing new intelligent compression techniques

H.264 is likely to be the video standard of choice for the foreseeable future based on the barriers of entry for H.265 mentioned above. However, there are still options for reducing bit rates further with the existing algorithms. One such technique comes from Axis Communications, called Zipstream — a radically more efficient implementation of H.264 that can reduce bandwidth and storage requirements by an average of 50 percent or more when compared to existing H.264 solutions. That should sound familiar as it is the same savings figure that H.265 is touting.

Axis is not the only company offering bandwidth savings by optimizing H.264, but this particular iteration has a unique approach that dynamically allocates regions of interest inside a camera scene. In more traditional solutions, the user defines a static region of interest. The problem with that approach is two-fold: the bad guy is likely to move out of a static region of interest and if you try to compensate for that fact by making the region of interest too big, you miss out on the compression savings.

Getting the H.265 ball rolling

Once the market works out the obstacles with licensing and VMS compatibility, manufacturers and customers will begin rolling out H.265 solutions. I would expect that more companies will launch H.265 cameras in 2016, but broad adoption may not occur for another few years. On the other hand, video coding enhancements such as Zipstream, will most likely be applied to whatever compression standards are launched in the future.

About the Author: James Marcella has been a technologist in the security and IT industries for nearly two decades. He is currently the Director of Technical Services for Axis Communications.