The Moving Picture Experts Group

High Efficiency Video Coding

Standard: 
Part number: 
2
Activity status: 
Open
Technologies: 
White papers: 

The overall amount of video data rate that is to be transferred over networks will continue to grow at a very fast pace, driven by the increased number of services and users and the increasing resolution of video from SD to HD and beyond. HEVC is the new generation of video compression technology with higher compression capability than AVC High Profile. HEVC supports all commonly used progressive scan picture formats, ranging at least from QVGA (320x240) to Ultra HD resolutions such as 8Kx4K, as well as picture formats of arbitrary size.

MPEG High Efficiency Video Coding (HEVC)

 

MPEG doc#: N11922
Date: January 2011
Authors: Jens-Rainer Ohm and Gary Sullivan

 

Market demands for video services with increasingly high quality (temporal and spatial resolution, color fidelity, amplitude resolution) continue to be evident, and video has become a major and continuously-increasing percentage of digital network traffic world-wide. It is difficult in the transmission networks of today to carry HDTV resolution with data rates appropriate for high quality to the end user – and this is especially the case on the Internet, where video-over-broadband services have now emerged a major phenomenon. According to various reports, video has become a majority of network traffic world-wide, and its traffic load continues to grow at a rapid pace. Ultra HD resolution (such as 4Kx2K and beyond) is expected to emerge in the near future and will be supported by next generation displays. Further data rate increases will put additional pressure on all types of networks (cable and wireless/mobile), and data rates for video content (when current compression technology is used) are increasing faster than the network infrastructure is able to carry economically. In the meantime, technology designs for substantial further improvement of video compression have begun to emerge. Therefore, definition of a new generation of video compression standard that has sufficiently higher compression capability than the existing AVC standard [1] has become a vital necessity.

As an intended solution to fulfil this need, the new HEVC standard is being developed by the Joint Collaborative Team on Video Coding (JCT-VC), which was established by MPEG and ITU-T VCEG with the goal to issue this new standard as twin text by both organizations. The development work was started after a joint Call for Proposals (CfP) was issued in January 2010.

The coding layer of the current HEVC draft specification and the corresponding HEVC test model (HM) software and encoding algorithms is still based on the traditional hybrid coding approach as found in previous standard designs, combining motion-compensated prediction between video frames with intra-picture prediction, closed-loop operation with in-loop filtering, 2D transformation of spatial residual difference signals, and adaptive entropy coding. Similar concepts of network abstraction layer (NAL) and high-level syntax (such as frame buffer management, sequence and picture parameter sets) are used as in the AVC standard.

In contrast to previous standards, larger block structures with flexible mechanisms of sub-partitioning are introduced. For this, the HEVC draft specification defines variable-block-sized coding units (CUs) which define a sub-partitioning of a picture into rectangular regions. The coding unit structure replaces the macroblock structure as known from previous video coding standards with an analogous structure of variable size, and each CU contains one or several variable-block-sized prediction unit(s) (PUs) and transform unit(s) (TUs). At the level of the PU, either intra-picture or inter-picture prediction is selected, the techniques for which are augmented by various new design elements for higher compression capability. Each TU is processed by a spatial block transform and quantization of the resulting transform coefficients. An adaptive loop filter (ALF) is applied within the prediction loop prior to copying the frame into a reference decoded picture buffer, providing improved objective and subjective quality. The ALF in the HEVC design has been added as a feature in addition to a deblocking filter with similar capabilities as the deblocking filter found in AVC. Context-adaptive entropy coding schemes are employed in a similar fashion as in AVC.

As of January 2011, the capabilities of the HEVC test model (HM) resemble or somewhat improve upon the capabilities of the best performing proposals submitted in response to the CfP, which, as reported in the CfP test report [2], in a significant number of cases showed similar quality as the AVC anchors when encoded at roughly half of the bit rate. Due to careful selection and combination of proposed technologies, this result has been achieved with a relatively moderate increase of complexity relative to AVC technology. The design is also configurable to provide appropriate trade-offs of compression capability versus implementation complexity.

Further improvements are expected within the ongoing development process. The first version of the standard is expected to be finalized by early 2013. Doubling the compression performance compared to AVC then would mean that Ultra HD video could be provided using compressed bit rates similar to those used for HD representation today. Further applications for stereo 3D, multiview, and scalable video capabilities based on the HEVC design can be easily foreseen.

A broader overview about the envisaged application areas of HEVC, the types of content to be supported and the technical properties of the envisaged design can be found in the HEVC Vision, Applications, and Requirements document issued by MPEG in January 2011 [3].

References

[1]   ITU-T and ISO/IEC, ITU-T Rec. H.264 | ISO/IEC 14496-10 Advanced Video Coding (AVC), May 2003 (with subsequent editions and extensions).

[2]   ISO/IEC JCT1/SC29/WG11 (MPEG), “Report of Subjective Test Results from the Call for Proposals on High Efficiency Video Coding”, doc. no. N11275, Dresden, DE, April 2010.

[3]   ISO/IEC JCT1/SC29/WG11 (MPEG), “Vision, Applications and Requirements for High Efficiency Video Coding (HEVC)”, doc. no. N11872, Daegu, KR, January 2011.

 

Amendments

N14701, Text of ISO/IEC 23008-2:2013/DAM4 3D Video Extensions

This amendment specifies multiview and depth video coding, referred to as 3D high efficiency video coding (3D-HEVC). The reader is referred to Annex I for the entire decoding process for 3D-HEVC, which is specified there with references being made to clauses 1-9 and Annexes A-I.

Other documents