January 25, 2013 marked a very important day in the timeline of video compression standards, but few security people probably noticed. On that day, the International Telecommunications Union (ITU) approved H.265, known as High Efficiency Video Coding (HEVC). According to the ITU press release, “HEVC will unleash a new phase of innovation in video production spanning the whole ICT spectrum, from mobile devices through to Ultra-High Definition TV.”
It was Ultra High-Definition (UHDTV) that put a big footprint on The Consumer Electronics show in January with numerous displays of 3840 x 2160 (known as 4K) resolution and models announced ranging from 55 to 110 inches (MSRP for an 84” model is - $20 - 25K). As if this 8.3 MP resolution wasn’t enough, the term UHDTV also covers 7680 x 4320, or 8K, for 33.2 MP resolution.
While it is apparent that stunning, higher-resolution displays are in our future — both at home and in the control center — it is the power of the underlying compression that is truly exciting.
First, let’s take a brief look at how the most used video compression standards have evolved with their years of ratification:
1988: H.261 (ITU) – intended for ISDN; provided CIF and QCIF resolutions.
1992: MPEG-1 (Moving Pictures Experts Group) – VHS-quality video, CD quality audio (formed basis for MP3).
1996: H.263 (ITU) – extended H.261 to phone-based videoconferencing and eventually to MMS, streaming media and network-based videoconferencing.
1996: MPEG-2 (also ITU H.262) – supported interlaced video, DVD quality and resolutions up to 720 x 480. Only full NTSC and PAL frame rates were supported.
1999: MPEG-4 Parts 1, 2, 3, 6 – built on prior MPEG standards to eventually provide a full range of frame rates, data rates, and resolution to 720 x 576. MPEG-4 consists of a number of profiles, including Part 2, Advanced Simple Profile. Consider this 2001 quote form CNET about MPEG-4, 2 years after its ratification: “While Microsoft and RealNetworks wage a noisy fight over their audio and video technologies, an underdog with a strong pedigree and whiz-bang features is quietly entering the fray.”
2003: H.264 (also MPEG-4, Part 10, Advanced Video Coding) – provided the basis for HDTV and Blu-ray DVD. In 2008, Axis and Sony introduced CCTV camera lines based on H.264, with many more companies to follow.
So, now, 10 years later, we have H.265 which is to H.264 what H.264 was to MPEG-4 Advanced Simple Profile, with roughly 50-percent gains in efficiency and bandwidth usage. How is this possible?
First, the processing power was developed to support the mind-boggling number crunching required to encode video. The picture, as in previous standards, is split into blocks for processing. In earlier standards, those blocks were 8 x 8 pixels — that number grew to macro-blocks of 16 x 16 in MPEG-4 and H.264. In H.265 blocks, called Coding Tree Blocks, are 64 x 64, but may be sub-divided.
Within a frame, encoding relies on “intra-picture” prediction, which uses an estimate of change from block to block, and a similar technique is used between frames (inter-picture). Prediction signals are used in a complex mathematical transformation, filtering and coding process whose result is employed at the decoder to reconstruct frames from reference frames. It is the increased sophistication of prediction and subsequent quantization and provisioning for more efficient parallel processing that gives rise to the efficiency of the encoding/decoding process. We see this in terms of better resolutions and lower bandwidths. H.265 will support resolutions from QVGA (320 × 240) to 4320p, but supports progressive scan only. Interlaced video may be sent either by coding each field or frame as a separate picture.