HEVC, the key to delivering an enhanced television viewing experience Beyond HD

Transcription

1 SMPTE Meeting Presentation HEVC, the key to delivering an enhanced television viewing experience Beyond HD Sophie Percheron, Marketing Product Manager ATEME, Bièvres France, Jérôme Vieron (PhD), Advanced Research Manager ATEME, Bièvres France, Written for presentation at the SMPTE 2013 Annual Technical Conference & Exhibition Abstract. Broadcasters and operators objectives remain the same over time: reduce transport costs, reach more customers, and improve the TV viewing experience. Ten years after the beginning of the SD to HD transition, history is about to be repeated once again. With the 50% bitrate savings promised by the new HEVC (High Efficiency Video Coding) standard, these compression gains will not only reduce the bandwidth but also allow us all to more importantly improve the viewing experience, conveying more information and emotional impact though higher spatial and temporal resolution. 1080p60 transmission has already been demonstrated to be a better user experience than 1080i30, inspiring market leaders to commence the transition towards producing 1080p60 channels. However the required bandwidth still remains the bottleneck to complete and efficient end-to-end deployment. This paper will demonstrate why HEVC will be the key to unlock the progressive only broadcast chain deployment; and why it will not lead to forcing a premature end of life to interlaced television. Finally we will discuss future (premium) services brought about by the beyond HD viewing experience: live events, sports, concerts, and other newly immersive experiences offered by UHDTV. Keywords. HEVC, beyond HD, Ultra-HD, 4K, UHDTV, enhanced viewing experience, video compression, interlaced, progressive The authors are solely responsible for the content of this technical presentation. The technical presentation does not necessarily reflect the official position of the Society of Motion Picture and Television Engineers (SMPTE), and its printing and distribution does not constitute an endorsement of views which may be expressed. This technical presentation is subject to a formal peer-review process by the SMPTE Board of Editors, upon completion of the conference. Citation of this work should state that it is a SMPTE meeting paper. EXAMPLE: Author's Last Name, Initials Title of Presentation, Meeting name and location.: SMPTE. For information about securing permission to reprint or reproduce a technical presentation, please contact SMPTE at jwelch@smpte.org or (3 Barker Ave., White Plains, NY 10601).

2 Introduction The analogue television switch-off has been completed in many developed countries over the past few years, and the availability of more digital broadcast outlets has paved the way to even more immersive audio and video formats. The broadcast industry is on its way to all-hd. But increasing demand for enhanced end-user experiences with high fidelity content (especially for home theatre and premium events like sports) are rising. End users are calling for Beyond HD. In parallel, with the recent availability of 4K-UHD (Ultra-High Definition) screens (TVs, tablets, and mobiles) on the market, the end-user remains with a new offer of high display capacity devices, but is facing an almost un-existing associated service offer. Some first new services experiments were conducted and demonstrated this year. Despite the availability of new devices, a massive deployment of real services is not feasible through the current end-to-end delivery chain because of bandwidth limitations. Video compression must be more efficient to enable the delivery of such Beyond HD services. The first question is do we really need to go Beyond HD? What benefits are brought, and what are the implications of Beyond HD? Can new technology improve the consumer viewing experience while not requiring a bandwidth increase? Broadcasters and operators objectives remain the same over time: improve the TV viewing experience, reach more customers, and deliver new services offers in the same or less bandwidth to improve the costs. After reviewing the existing leverages to provide an enhanced viewing experience, we will propose a description of Beyond HD facets, and will describe the current limitations to broad deployment of these new services. Then a brief description of HEVC (High Efficiency Video Coding), the next generation video compression standard, will be given with latest results demonstrating its direct benefits for Beyond HD services. Finally we explain why the HEVC technology will ease the all progressive broadcast chain while unlocking the development of the next enhanced services generation, without thereby announcing farewell of interlaced. New trends: from capture to devices In very large measure, today s broadcast television production chain still operates in interlaced mode. With the advent of digital, the latest generation of cameras captures in progressive mode. Nevertheless, most still output an interlaced feed as the whole progressive broadcast chain is not yet ready. We estimate that 70% of the worldwide HD content production is still done in interlaced (i.e.1080i25/30) and only 30% in the progressive (i.e. 720p50/60). Some market precursors already undertook initiatives to get whole production chain in 1080p50/60. As devices with enhanced display capabilities, increased decoding capacity (e.g. multicore architectures), and ability to manage higher spatial resolution, frame rate and bitrate; are about to arrive on the market. New UHD big screen devices are well on the way, but we will also see premium HD services soon reach tablets, mobiles and OTT devices owners. These services may well reach a mass market much faster on the small screens, as consumers around the world are increasingly driven to upgrade their phones, tablets on two year cycles. We believe that end users will request to benefit from an enhanced viewing experience on all these screens. With the development of TV Everywhere, the market may soon call for Beyond HD. The access to these higher forms of premium content and new enhanced services is currently prohibited by the bandwidth limitations on managed and unmanaged networks. The development of new infrastructure like 4G and/or standards like DVB-S2/T2 could (partially) circumvent these limitations and free the access to Beyond HD services. 2

3 What is Beyond HD? Beyond HD is the next-generation TV systems that will be able to offer an enhanced end-user experience through new audio-visual technologies that will provide a more realistic experience, stunningly better image quality, and more accurate visual details as compared with today s current HDTV systems. Keys to improving the viewing experience By commonly associating Beyond HD with 4K-UHD, end-users and TV suppliers narrow the focus on improvements to spatial resolution only. But Beyond HD is not only a matter of spatial resolution. Improving the end-user viewing experience can be realized through other characteristics of the video, including frame rate, the depth and the realism of color displayed. The choice of a preferred frame-rate has a hot topic of debate for many years. It has been clearly demonstrated that 60Hz is better than 30Hz for, but several studies [1][2] indicate that frame-rates even higher than 100Hz are needed to avoid temporal artifacts. During IBC 2013, EBU 1 presented the results of the HFR (High Frame Rate) project which demonstrated the increase in viewing comfort observed from 60Hz to 120Hz, but also from 120Hz to 240Hz. In the production of HDTV programs [3], all the colors captured using video cameras are firstly defined within the gamut formed from the CRT-centric reference RGB-primaries. With the arrival of new display s technologies (e.g. LCD, LED) as well as new cameras with extended characteristics, colorimetric becomes of major importance to improve the user experience. First wide color gamut can be considered as in BT.2020 [4] which covers 75.8% of the CIE 1931 color space whereas the BT.709 [3] is only covering 35.9%. Bit depths of 8,10,12,16 bits are also important parameters to control the accuracy of the color sampling. Chroma sampling rates also need to be considered. Finally, a High Dynamic Range (HDR) would allow for a more realistic perception between the brightest and darkest elements in a scene. Authors would like to add unusual characteristics to this set of parameters implied in the enduser viewing experience: the Film Grain. Beyond HD and premium services for home theatre would take into account such characteristic that is finely tuned by film directors. Film grain plays a significant role in the atmosphere of a scene. It was observed that, in terms of visual perception, film grain reinforces the feeling of realism for the audience. Various approaches, whether proprietary or standardized, are considered in order to deal with extended color management. However, whether this is based on full or scalable metadata, higher bit depths, higher frame rates, or enhanced spatial resolutions such choices will increase the amount of data to be processed and raises important technical and bandwidth issues all along the production and the broadcast delivery chain. Beyond HD services Beyond HD can have different facets and provide different benefits depending on the parameter sets applied to the images. HD+ (1080p50/60) Experiments have demonstrated that content with fast and large camera motion requires enhanced viewing comfort. Sports content or live event viewing experience challenge today s 1 European Broadcasting Union 3

4 HDTV with image fluidity and motion quality may result in an unsatisfying experience for the end-user, especially as screen sizes increase. Moving from 1080i25/30 to 1080p50/60 production and distribution can enhance the end-user viewing comfort, but of course requires a doubling of the information to be displayed. Hence, we can definitely consider a 1080p50/60 service as a Beyond HD service. We propose to call it HD+ in the sequel. Content like sport, theatre, and live concert may be good candidates for upcoming HD+ services. Is there also a potential need for what we might call HD++ services with even higher frame rates (e.g. 120Hz), alternate color spaces, and higher dynamic range? Nevertheless, we will keep in mind that broadcasting a feed at higher frame rates (50/60, 100/120 or even higher) will definitely lead to increases in the required bandwidth. Would doubling the frame rate require additional bandwidth in nearly similar proportions? UHDTV A better sense of realism, immediacy and intimacy with the viewing experience are definitely keys to television success. Reaching into the emotions of the audience is essential. UHD provides additional features that can touch upon all these emotional aspects by increasing the fidelity of image detail and movement, increasing viewer involvement and providing an enhanced sense of being there (cf figure 1). Figure 1: UHD view of a large scene (left) and character details at various resolutions (right). As discussed earlier, UHDTV is not only a matter of spatial resolution, but also a matter of wide color gamut, higher dynamic range and higher frame rate, together enabling a radical improvement in viewing experience. In this context, the ITU 2 standardized UHDTV1 (as pixels) in [4]. However this standard does not clarify which subset of features should be used for broadcast applications. We believe that early UHDTV broadcast would be performed with 3840x2160 pixels pictures, assuming a 10 bits signal transmitted at 59.94Hz, with multi-channel audio up to 8 channels. Would UHDTV fit on current distribution architecture? Very first Beyond HD experiences and limitations UHD experimentations on satellite During this calendar year, satellite operators the world over (Europe, North America and Asia) experimented with UHD transmission. Most of them experienced the satellite broadcast of Quad 2 International Telecommunication Union 4

5 HD channels on their transponders. As an example, broadcasting via DVB-S2 an UHD/Quad HD channel encoded in H.264/MPEG-4 AVC required at least 40Mbps in order to provide the video quality expected with UHD. Limitations are reached: knowing that practically a transponder broadcast via DVB-S2 at around 70 Mbps, only 1 UHD/Quad HD channel would fit. If first UHD offline encoders are arriving on the market, high power processing required for realtime encoding prevent UHD Live (i.e. single Tile) deployment soon. The only solution experimented so far for live is to broadcast Quad HD content produced by 4 encoders fed with 1080p50/60 on 3G-SDI links. Then, 4 synchronized IRDs are needed to decode those streams and 4 links (3G-SDI or HDMI) are finally used to transport the uncompressed signal to the display. Even if such architecture can be considered for contribution matter, bandwidth constraints remain very important and neither the set-up (i.e. 4 decoders or 4 STBs) nor the bandwidth is realistic for distribution matter. Alternative solutions have to be considered. VOD content providers launch a first 4K-UHDTV service Beginning of September 2013, a market actor launched a 4K-UHD video store. This service is a world s first and is not likely to be broadly deployed as there are limitations due to the use of a proprietary UHD TV set and an associated proprietary Media Player. To launch such an UHDTV video offer, independent Video On Demand (VOD) providers, broadcasters or operators would need to provide their UHD first-mover customers new equipment since current legacy STB, PC s, game boxes, etc. do not support UHD decoding. With the current compression technology such services favor end-users having high bandwidth network since the download duration may lead to a limited usage and/or a limited audience. HEVC: New generation video compression Several solutions exist to address network congestion issues and bandwidth constraints: from infrastructure evolution (e.g. DVB-T/S to DVB-T2/S2 standards) to compression technology evolution. Thus, the new generation video compression standard HEVC (High Efficiency Video Coding) has been developed by the JCT-VC -Joint Collaborative Team Video Coding - a joint working group of both ISO/IEC Moving Picture Experts Group (MPEG) and ITU-T Video Coding Experts Group (VCEG). This new standard was finalized in January 2013[5][6]. On performance History is once again repeating itself; ten years after the 50% coding gain provided by H.264/MPEG-4 AVC over MPEG-2 [12], yet another 50% improvement from HEVC over H.264/MPEG-4 AVC has been attained. The latest objective (PSNR-Bjontegaard) and 5

6 subjective (Mean Opinion Score) tests [7][8], on various broadcast content between HEVC and the AVC High Profile, have already shown a gain in compression better than 40% for HD content. Subjective tests indicate that the HM encoder (i.e. HEVC reference software) is saving at least 50% bitrate compared to the JM (H.264/MPEG-4 AVC reference software) for equivalent perceived quality. Similar performance has been exhibited on a large set of content using both our H.264/MPEG-4 AVC and HEVC encoders. Moreover, in the context of the 4EVER project 3, we conducted a wide study considering both scanned and native 4K-UHD content presented in [9] which exhibited coding gains in a range of 38% to 50% based on subjective measurements. Yet another hybrid codec? Despite similar overall architecture, the HEVC standard contains a set of advanced tools that together make this new codec more flexible and much more efficient than H.264/MPEG-4 AVC. This codec is particularly tailored toward high resolution images with block sizes that could be as large as 64x64, and high frame-rates with advanced motion vector prediction (cf Figure 2). On profiles Figure 2: Coding structure comparison between H.264/MPEG-4 AVC and HEVC. Unlike H.264/MPEG-4 AVC, the goal with HEVC was to consider only a reduced number of profiles. Thus (so far) only 3 profiles have been defined: Main, Main 10 and Still Picture. The Main profile covers the wide set of current consumer distribution applications from mobile devices to HDTV and supports an 8 bits 4:2:0 signal. The Main 10 profile targets premium and Beyond HD applications by offering a 10 bits support, therefore suitable for contribution and high end use cases. Finally, the Still Picture profile concerns I pictures only. Unofficially, the current HEVC standard is called HEVC v1 since different research work related to further extensions/amendments has been undertaken with finalization planned for early In addition to the Scalable and the Multi-view extensions, the Range Extension (Rext) is of major importance for the purpose of professional and Beyond HD applications. 3 4EVER - For Enhanced Video ExpeRience - ( is a French consortium supported by Europe (FEDER), French Ministry of Industry, French Regions of Brittany, Ile-de-France and Provence-Alpes-Cote-d Azur and related Competitivity Clusters. It is 3 year collaborative R&D project running from 2012 to

7 Indeed, these extensions not only target bit depths of 10 bits or higher, they look to provide for 4:2:2 and 4:4:4 chrominance sampling, but also a wider color gamut. HEVC is being considered as the next generation s premium compression technology. Broadcasters, Movie Studios and other high-end content producers are closely following these activities. HEVC performance for Beyond HD In order to assess the performance of HEVC on Beyond HD content, we performed a detailed study and the main results are reported here. Our main objectives were to evaluate the impact of frame rate, spatial resolution and bit depths on the compression efficiency of HEVC. All the reported results are based on objective quality measurement (PSNR). HEVC vs H.264/MPEG-4 AVC Figure 3: HEVC vs H.264/AVC performance on Beyond HD: 1080p (left) and 2160p (right). When comparing HEVC and H.264/MPEG-4 AVC performance on Beyond HD content like 1080p50/60 or 2160p60 sequences, we observed a compression gain of between 38% and 45%. The results illustrated on figure 3 have been validated on a large set of sequences. On Frame rate Figure 4: Frame rate impact on HEVC performance: Hz (left) and Hz (right). In order to evaluate the impact of the increase of frame rate on the bitrate, we investigated various video sequences captured in 50/60 Hz and 120Hz with different motion and texture characteristics. Figure 4 illustrates some results obtained on 1080p and 2160p sequences. A bit rate increase between 19% and 42% is observed when going from 25/30Hz to 50/60Hz, while a lower bitrate increase between 17% and 28% is observed when comparing 60Hz to 120Hz. Note that the impact of the spatial resolution seems negligible on this percentage. 7

8 Broadcast quality First results provided in [9] indicated that the HEVC coding gain would allow a broadcast quality of UHDTV at video bitrates below 13Mbps. These results were obtained with the HEVC Main profile using 4:2:0 8 bits for 60Hz content and have been confirmed across a wide set of 8 bits and 10 bits video sequences from 24Hz to 60Hz, leading to our finding that broadcast quality for UHDTV with HEVC can be achieved at bitrates between 11Mbps and 18Mbps. The Main 10 profile After confirming that bitrate savings between H.264/MPEG-4 AVC and HEVC are similar for both Main and Main 10 profiles, we performed some comparisons between various 10 bits sequences. A bitrate saving in average of 5% is observed for Main 10 versus Main, confirming the relevance (if needed) of such a 10 bits profile for the Beyond HD experience. On film grain One contrary result of the good performance of using HEVC for UHD content is the impact on film grain. Indeed, within the range of bitrates between 10 and 20 Mbps, the good grain of filmbased sources is identified as noise and tends to be eliminated from the signal, which of course may not be what the director. This is why we recommend the use of the HEVC Film Grain Characteristic SEI message to allow the add of controlled grain on the decoder/display side. Why is HEVC the key to unlock the Beyond HD broadcast chain? Beyond HD services will definitely provide end-users with an enhanced viewing experience. However, so far, bandwidth limitations prevent any massive and commercially viable deployments in the current infrastructure. With the arrival of HEVC, as demonstrated in the previous results section, Beyond HD services will become a reality. With the emancipation of TV Everywhere usage, the increasing penetration of OTT devices and the rise of high video quality expectations, Beyond HD services in concert with HEVC compression could be deployed with only a very limited (or even no) impact on the bandwidth costs for broadcasters, programmers and network operators. Thus, services like HD+ (1080p50/60) could be delivered at bitrates slightly below today s 1080i data rates with H.264/MPEG-4 AVC. Higher bit depth and/or frame rate signals (e.g. 10 bits, 120Hz) could even be broadcast at similar bitrates as those used today for transmitting HDTV in MPEG-2. UHDTV deployments of services at 60Hz with 10 bits signal could be envisioned at bitrates between 11 and 18 Mbps. Using HEVC, up to 4 UHD channels could fit on a satellite transponder (DVB-S2). Benefits brought by the combination of Beyond HD and HEVC (i.e. improved end-user experience and increased number of services) should be monetizable! HEVC would enable the emancipation of the Beyond HD formats. With such ideal metaformats all sub-formats for all screens can be easily addressed. In the context of multi-screen, using such progressive formats instead of 1080i25/30 would simplify the broadcast chain deployment, while limiting and avoiding complex processing steps like deinterlacing etc. One more step towards the progressive only broadcast chain!? HEVC does not mean interlaced end of life About 70% of the HD television broadcast around the globe is interlaced, most specifically 1080i25/30. Moreover standard definition SD (i.e. 480i or 576i) is only broadcast as an interlaced format. The legacy interlace format is still the most dominant TV format in the world. 8

9 As described above, HEVC offers very attractive compression benefits, but lacked any specific tools designed for interlaced formats. Indeed, unlike the efficient PAFF (Picture Adaptive Frame- Field) and MBAFF (Macroblock Adaptive Frame-Field) coding tools of H.264/MPEG-4 AVC, HEVC was originally designed for progressive applications only. One may agree that ultimately the switch from interlace to progressive formats for all of television is only simply a matter of time, but time is not now. There are many technical, operational and commercial challenges for today s interlace houses to solve before they can willingly make a move to progressive. Agreements have to be obtained from content producers/owners: many of them categorically refuse to see their interlaced content be modified into progressive. Interlace coding: objective and subjective results As co-chair of the HEVC working group on interlaced coding, we have made various contributions in collaboration with renowned television experts from across the globe, leading to the integration of new tools within the HEVC standards to allow for full support of interlaced coding. New SEI and VUI messages were introduced. Even without dedicated compression tools, HEVC makes it possible to compress interlaced content either as Frame coding (progressive only) or as Field coding (field only at a doubled frame-rate). Objective and subjective measurements were performed on legacy interlaced content. As only HD content was considered in the MPEG working group, we decided to evaluate HEVC using the set of broadcast content used to evaluate MPEG-2 and H.264/MPEG-4 AVC compression technologies. Comparisons were performed using the JM18.5 and the HM12 software. Figure 5: Illustration of Bitrate gains of around 30% on SD (576i) and HD (1080i) content. As illustrated on figure 5, HEVC objective bitrate gains on legacy interlaced content are found to be around 30%, both in SD and HD. Hence for instance, SD could be broadcast at 1.25Mbps instead of 1.8Mbps and HD at 5.5Mbps instead of 8Mbps. In addition, the 4EVER Consortium performed a subjective evaluation of HEVC on typical 1080i content reported in [10]. The conclusion of this study is that a subjective coding gain of about 50% is observed assuming the same perceived video quality. Interlace coding is still alive Such results show that the HEVC standard allows for the efficient coding of interlaced content. Thus, legacy television applications could benefit of HEVC by lowering the bitrate by at least 30% (or alternatively, increasing video quality). Following this conclusion, in September 2013 the DVB (Digital Video Broadcasting) group decided to integrate the support for HEVC interlace coding. We shall see interlacing for quite some time still in the broadcast industry. 9

10 What should speed up the arrival of Beyond HD to the market? First-mover programmers, broadcasters, or video platform providers may use the occasion of upcoming sporting events to accelerate the market delivery of Beyond HD: the Sochi Olympic Games from Russia in February 2014, the Brazil World Cup in June 2014, and the Rugby World Cup from the UK in 2015 are likely targets. These events would be fantastic laboratories for testing the latest service innovations, and in a competitive landscape, operators and programmers should play the innovation card to market a differentiated offer. The two ends of the Beyond HD broadcast chain, from cameras to TV sets and video projectors have been in limited availability for almost a year, and are starting to become affordable. However, native UHD support for most other devices like switchers, encoders and STB is still expected. Moreover, the post-production workflow requires adaptation to higher bitrates, process and storage capacities: the first challenge in deploying Beyond HD is in baseband. Fortunately, the technical landscape evolves very quickly and some of these technical limitations are beginning to fade away. Despite the announcements of equipment and standards like 6G or 12G-SDI, the natural trend for producers, operators and broadcasters (with long term desires but short-term capital buying cycles) is to work around the bottlenecks by using what they already have in new ways, like aggregating multiple 3G-SDI links. On the end-user side, the arrival of HDMI 2.0 (with a support of 4Kp60 up to 10bits) could de facto set up the UHDTV broadcast format phase 1. Extended gamut, HDR and HFR should arrive in a phase 2, but when? Regarding decoder building blocks, one can expect the arrival on the market early 2014 of STB supporting UHD services. But we believe that the upcoming Smart TVs able to natively manage UHD will definitely speed-up the arrival of UHD services. In addition, infrastructure evolution, like DVB-S2 extension, is planned before the end of the year and can increase the spectral efficiency from 30% to 50%, enabling users to transmit even more or better signals from satellite at higher bitrates. A possible Beyond HD roadmap? 10

11 Conclusion After describing the various levers that can enable the Beyond HD viewing experience, potential future (premium) services like HD+ (or HD++) have been discussed that could benefit television users in the very near term. But we also discuss the implications the new immersive television experience offered by the promise of UHDTV. New test results provided in this paper demonstrate HEVC s high compression performance and its direct benefits for Beyond HD services deployment. Finally, we explain why this compression technology can ease the all progressive broadcast chain deployment; and why it will not lead to the premature end of life for interlacing. Despite the fact that we are convinced that both HD+ and UHD bring real enhanced viewing experiences, the market, to date, seems completely consumed with the prospects and promises that UHD offers. Indeed, UHDTV appears to be easier to market to the end user. The UHDTV value proposition is based on real differentiating elements: new TVs offer larger screens, new immersive experiences, and forms of new premium content for home theatre, live concerts and sporting events. UHDTV should be monetizable! Even if the HEVC compression technology will unlock the development of the next enhanced services generation without requiring any (major) infrastructure modification, the availability of HEVC Beyond HD encoders and decoders remains the question. Thus, HEVC offline encoders for UHD content are just made available but what about professional HEVC live UHD encoders? Acknowledgements The authors would like to thank their colleagues J.M. Thiesse, A.L. Lavaud, C. Daguet, and the members of the collaborative project 4EVER for their contributions to this paper. References 1. T. Yamashita et al "Super Hi-Vision video parameters for next-generation television, Proc SMPTE Conf, R. A. Salmon et al, Higher Frame Rates for More Immersive Video and Television, BBC White Paper WHP 209, ITU Recommendation BT.709, Parameter values for the HDTV standards for production and international programme exchange," ITU Recommendation BT.2020, Parameter values for Ultra-High Definition television systems for production and international programme exchange," ITU-T Recommendation H.265, High Efficiency Video Coding, April ISO/IEC FDIS , Information technology -- High efficiency coding and media delivery in heterogeneous environments -- Part 2: High efficiency video coding, G.J. Sullivan, J. Ohm, W-J. Han, and T. Wiegand, Overview of the high efficiency video coding (HEVC) standard, IEEE Transactions on Circuits and Systems for Video Technology, 22(12): , J. Ohm, G.J. Sullivan, H. Schwarz, Thiow Keng Tan, and T. Wiegand, "Comparison of the Coding Efficiency of Video Coding Standards-Including High Efficiency Video Coding (HEVC)," IEEE Transactions on Circuits and Systems for Video Technology, vol.22, no.12, pp , Dec J. Viéron and M. Parmentier, "4K TV capture, an early experience sharing, Proc SMPTE Conf, Z. Agyo, C. Daguet, J. Fournier, J-C Gicquel, F. Henry, J-M. Thiesse, J. Vieron, on behalf of the 4Ever Project, Subjective comparison of HEVC and AVC for HDTV content, MPEG document, m29210, April