Taking H.264 video compression mainstream

In the last year, we watched "trends" explode on the floors of tradeshows like ISC West and ASIS International Seminars and Exhibits. The year 2007 seemed to be the launch of video analytics from seemingly a couple dozen firms. If we could pinpoint one of the "trends" at ISC West 2008, it would have to be H.264.

If you're not familiar with H.264, the quick summary is this is a standardized video compression format and you might hear it go by alternate names such as MPEG 4 Part 10 or AVC (advanced video coding). In the security industry though, the de facto name of this standard has been H.264, as the name "MPEG-4 Part 10" is often confused with the numerous variations of MPEG-4 that had been adopted by camera and storage solution providers.

So why is it 2008 before you start to hear much about H.264 video compression, even though the standard has existed in its final draft since 2003?

According to Anders Laurin, executive vice president at Axis Communications, one of the main challenges has been processing power. To compress video better using this more powerful standard, he explains, you need more processing power. And since H.264 compression, which minimizes bandwidth and storage needs, is especially applicable for video transmission, that meant the processing needed to be done at the edge.

While many network cameras and IP video encoders had taken advantage of common compression formats MPEG-4 and MJPEG (Motion JPEG), most camera chipsets didn't allow the volume and speed of processing for H.264 to be a viable compression format at the camera level. According to Laurin, that's changing. His firm today introduced a new chip designed to allow the company to take H.264 into its cameras. About as small as a fingernail, the chip brings stronger processing, not only allowing the company to do H.264 at the camera level, but also allowing the camera to include basic video analytics to run at the camera level.

According to Fredrik Nilsson, the general manager for Axis in North America, H.264 had seen some earlier industry buzz and had been seen in some companies' products (including DVRs, where analog video was being digitized for storage), but he said the problem was that it was not typically allowing the kind of resolution or frame rate processing that purchasers of surveillance hardware had come to expect.

So what kind of performance can you expect in H.264? According to Nilsson, if you consider a Motion JPEG (MJPEG) stream as the benchmark, then H.264 can decrease bit rate by up to 80 percent. MPEG-4, comparatively, would only decrease bit rate by 50 percent. H.264 also side-steps some problems with MPEG-4, which had so many variations developed by manufacturers that the compression format didn't function well as a standard. The H.264 standard, according to Axis' Dr. Jumbi Edulbehram, allows the compression to be configured for special needs. That move means that product vendors won't need to side-step and bastardize the standard, as was often seen in the world of MPEG-4 products.