By John Blem

Not being new to these trade shows -- ISC West in the U.S., and IFSEC in the U.K., I realize that the media wants to report on the latest and greatest.  When a particular concept grabs hold, it can seem like they’ve all hopped on the newest bandwagon to be on the forefront presenting the latest and greatest.  Let’s remember that just a couple of years ago IP was the hot topic, and that is now becoming the accepted standard for surveillance as an undeniably vast improvement over analog technology. 

THE BUZZ AROUND H.264

Currently, the buzz is about the new video compression standard H.264.  This standard has been jointly developed by the two international standards bodies ITU and ISO/IEC and is also known as MPEG-4 Part 10 Advanced Video Coding (AVC).

Demands in the video surveillance world are increasing for more storage and bandwidth without compromising the high frame rates and high resolutions that are desired these days.  More effective compression methods are, therefore, required.  An H.264 encoder can reduce the size of a digital video file by more than 80% against the Motion JPEG standard while maintaining the same visual quality.  Compared to MPEG-4 Part 2 Simple Profile (SP), H.264 typically achieves a 40-50% reduction in the size of a video file.

The market uptake of megapixel cameras is growing but the issue has been the storage requirements that such high resolution cameras can generate.  The use of H.264 will have a significant impact on this going forward.

My personal view is that H.264 will almost obliterate the use of MPEG-4 Part 2 over a period of just a few years.

Video management solution providers will be implementing this new functionality in the near future, as will all major camera manufacturers eventually.  Several of these were already able to present at the shows last year -- Milestone XProtectTM Enterprise 6.5c had been supporting H.264 in summer 2008, as a seamless upgrade to existing versions of the software.

HARDWARE REQUIREMENT

On the flip side, H.264 is in the early stages right now.  While it provides a great tool for savings in network bandwidth and storage costs, it will only perform with high performance cameras.  The compression standard is more sophisticated than previous standards, with the decoder complexity around two times higher than that of MPEG-4 Part 2 SP, which will translate into higher computational requirements.

Even though H.264 is new for the IP surveillance industry, it has been available as a standard for about 5 years and many areas have absorbed it already -- e.g., next-generation, high-definition DVD (Blue-ray) format.

TECHNICAL BACKGROUND

H.264 is a block-based, hybrid video coding standard using motion compensation, where motion compensation refers to the use of displacement vectors that are included to aid in the prediction between pictures.  The typical high correlation between successive pictures in a video sequence allows for coding the displacement of the different parts of a picture and the prediction difference between the actual picture content being coded and the displaced content from other coded pictures -- the so-called reference pictures.  This is called inter-prediction.

There are two main types of inter-prediction -- prediction based on one reference picture (P-type macroblock) and bi-prediction (B-type macroblock) that utilizes a combination of two reference pictures.  To facilitate random access and to increase error robustness, the standard also allows for intra-coding where the coded data does not depend on other pictures as is the case for inter-prediction.

The H.264 standard divides a picture into a number of macroblocks that each consists of a 16-by-16 pixel portion of the picture.  The macroblocks are grouped into one or more slices, usually in raster scan order.  Therefore, one picture may be coded as just one slice or as a number of slices.  The use of slices allows different error resilience tools, different macroblock coding types, and such things as the ability to compress the fields of an interlaced picture separately from each other as individual slices.

Color video is coded with the luminance signal separate from the chrominance (color) signals, usually with sub-sampling of the chrominance signals relative to the luminance signal for more efficient coding due to the nature of the human visual system.  In overall terms, this is not fundamentally different from previous video coding standards (including MPEG-4 Part 2), which are also block-based hybrid video coding standards.

NEW CODING TOOLS AND ENHANCEMENTS

Yet H.264 does introduce a number of new coding tools, along with enhancements to other coding tools.  The biggest significance of these is an in-loop adaptive de-blocking filter that helps reduce blocking artifacts, storing more than two reference pictures for better prediction, sub-division of macroblocks into smaller blocks (down to 4-by-4 pixels), prediction in intra-coding, and the use of an integer transform instead of the DCT transform used in previous standards.

Built into H.264 is the concept of a Network Abstraction Layer (NAL) that on top of the Video Coding Layer (VCL) represents the compressed video efficiently to allow for easy integration with a number of different protocols and transport mechanisms -- very suitable for IP networks.

The net result of all the coding technology advancements incorporated into H.264 is that it out-performs all previous video coding standards, and is, therefore, the current state-of-the-art within digital video coding.

In summary, is the media buzz on H.264 worth it?  Video compression standards are changing with its advent -- being able to maintain or even lower your bit rate and storage requirements while maintaining a high resolution stream is valuable.

John Blem is CIO and co-founder of Milestone Systems  (www.milestonesys.com).