<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Diego Gibellino on Medium]]></title>
        <description><![CDATA[Stories by Diego Gibellino on Medium]]></description>
        <link>https://medium.com/@gibellino?source=rss-28b254ee3fba------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*AiknqLhVrAoH9GsFeVOSRQ.jpeg</url>
            <title>Stories by Diego Gibellino on Medium</title>
            <link>https://medium.com/@gibellino?source=rss-28b254ee3fba------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 01:06:56 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@gibellino/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[A Future for MPEG]]></title>
            <link>https://medium.com/@gibellino/a-future-for-mpeg-18b214595658?source=rss-28b254ee3fba------2</link>
            <guid isPermaLink="false">https://medium.com/p/18b214595658</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[standards]]></category>
            <category><![CDATA[multimedia]]></category>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[mpeg]]></category>
            <dc:creator><![CDATA[Diego Gibellino]]></dc:creator>
            <pubDate>Sun, 22 Sep 2019 07:13:09 GMT</pubDate>
            <atom:updated>2019-09-22T07:13:09.829Z</atom:updated>
            <cc:license>https://creativecommons.org/licenses/by-nd/4.0/</cc:license>
            <content:encoded><![CDATA[<p><em>As the market of digital media evolves and requires innovation at an unprecedented pace, MPEG should also evolve to face the challenges ahead and to better serve this industry. There are ongoing discussions within ISO/IEC JTC1 about the structure that will better support the development of standards for still and moving pictures coding in the future. MPEG, in over 30 years, has achieved an impressive track of published standards with deep economical and societal impact along with a constant growth in terms of expert’s participation and brand recognition…it’s time for a new beginning?</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*LVS3Yu-pObA7fke-" /></figure><p>The Moving Picture Experts Group, <em>aka</em> MPEG, operates today as a Working Group (WG11) of the Sub-Committee 29 (SC29) of ISO/IEC JTC1. Its objective is to develop standards for the compression of video, audio, metadata and other data, along with the required systems technologies. In over 30 years of activities, MPEG has produced standards that created and shaped a whole Industry: MPEG-2, MPEG-4, AVC and HEVC (jointly developed with ITU-T VCEG), MPEG Audio Layer II/III, AAC, DASH, CENC and CMAF, to name a few.</p><p>These results are largely due to consistent efforts and investments that experts and Companies or Academia put into MPEG, but also to a specific <em>modus operandi</em> embraced by MPEG: a continuous evolution driven by the ability to conceive strategic visions for new media user experiences, design work plans matching industry needs, develop standards in close collaboration with client industries, demonstrate their performance and promote their use.</p><p>Recognizing that the number of activities, relationships, and membership of MPEG grows, and its scope becomes wider, a group of experts started to feel that a Sub-Committee status would better support MPEG’s objectives, while keeping structure slightly amended to improve the match of standards with market needs, leverage internal talents, and retain and enhance the MPEG brand.</p><p>A proposal to elevate MPEG to a Sub-Committee level in JTC 1 (SC 4x), organised into Working Groups and Advisory Groups is depicted in the figure below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kchSqY8IdSyfcfFQgMqqUQ.png" /><figcaption>A proposed new structure for MPEG as SC 4x</figcaption></figure><p>The proposed structure, marginally different than the current one, has been designed to keep as much as possible the current MPEG <em>modus operandi</em>:</p><p>The new SC will conduct its activities by organising meetings collocated with WGs and AGs (as today happens for the different subgroups within the working group) requesting to meet. Synergies and coordination among different groups will be ensured by joint meetings during physical meetings and common activities with external entities will be conducted with formal Joint Working Groups (e.g. JVET with ITU-T). Current subgroups within WG11 will become separate WGs with additional Advisory Groups for technical coordination, technical requirements, market needs and liaisons and communication (according to ISO Guidelines AGs don’t produce standards).</p><p>Meetings will be held with the traditional “MPEG week format” slightly adapted to accommodate SC 4x plenaries, but with no impact on WG/AG meeting times.</p><h3>Conclusions</h3><p>Organisations shall adapt and evolve to face new challenges and exploit new opportunities. MPEG is not different: it needs to change as a part of a process started over 30 years ago that proved to generate value for the industry. A group of experts designed a proposal to elevate MPEG to Sub-Committee level within ISO/IEC JTC1. This proposal tries to conjugate the current successful MPEG <em>modus operandi</em> with an improved formal structure and governance able to accommodate the increased level of work load, membership and complexity and the diversity of activities carried out by the group.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/424/1*aw3N4-Uvi-02fnlTHj_d4g.png" /></figure><p>I’m one of the experts who contributed to the definition of this proposal, which has been conceived to ensure MPEG continuity and at the same time provide the tools needed to face new challenges in the future. It should be however recognized that other views and different positions will be likely developed and presented by other interested parties as part of the internal discussion in the ISO/IEC JTC1 SC29 Ad-hoc group created for this purpose.</p><p>More information about this proposal and the ongoing process to design the <em>“Future of MPEG”</em> can be found on the <a href="http://mpegfuture.org/">MPEG Future</a> web site.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=18b214595658" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Versatile Video Coding hits major milestone]]></title>
            <link>https://medium.com/@gibellino/versatile-video-coding-hits-major-milestone-baeb13c8960a?source=rss-28b254ee3fba------2</link>
            <guid isPermaLink="false">https://medium.com/p/baeb13c8960a</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[multimedia]]></category>
            <category><![CDATA[compression]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[Diego Gibellino]]></dc:creator>
            <pubDate>Mon, 02 Sep 2019 08:27:28 GMT</pubDate>
            <atom:updated>2019-09-29T12:18:16.384Z</atom:updated>
            <cc:license>https://creativecommons.org/licenses/by-nd/4.0/</cc:license>
            <content:encoded><![CDATA[<p><em>VVC, the new video codec developed by the Joint Video Experts Team (JVET) of ISO/IEC MPEG and ITU-T VCEG, reaches a key milestone with the availability of a Committee Draft of the specification. Once finalized, in 2020, this standard will provide the same visual quality of the current state of the art technologies in video compression (e.g. HEVC) at half of the required bitrate.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*KEtqAlMoFORkm6g_" /></figure><p>Versatile Video Coding (VVC) is a new video compression standard under development by the Joint Video Experts Team (JVET). JVET is a collaborative group that includes experts from ISO/IEC JTC1 SC29 WG11 MPEG and ITU-T SG16 Q6 VCEG. During the last meeting in Gothenburg, the group decided to issue a Committee Draft (CD) of the specification. The availability of a CD is a major step towards the creation of a new standard because, according to ISO/IEC directives, is the first formal opportunity for National Bodies to vote and comment on the progress of this project. It also implies that the general architecture and the set of tools and technologies currently described in the CD text, once approved, will likely be part of the final standard though minor changes and refinements may also be included later in the process. The resulting international standard will be finalized in 2020 and will be available as ISO/IEC 23090–3 Coded representation of immersive media — Part 3: Versatile Video Coding and as ITU-T Recommendation.</p><p>VVC will provide an average reduction of roughly half the bitrate compared to the previous generation of codecs (e.g. HEVC) for the same visual quality (subjective quality). This means that a UHD video delivered today by an OTT streaming service with bitrates up to 15 Mbps, will likely require less than 8 Mbps once VVC implementations will be available and mature.</p><p>Current BD-rate figures on VVC coding efficiency, based on PSNR, tend to confirm that this 50% target gain can be achieved, showing bitrate reductions of 37% on average for UHD content (subjective results are typically much higher than those based on objective metrics). Another important aspect, is the trade-off between performance, in terms of coding efficiency, and computational complexity. The VVC test model (VTM), the reference software implementation adopted and developed by JVET, in version 5, shows an increase in complexity of roughly 10x for the encoder and less than 2x for the decoder. Values are related to encoding and decoding time compared to the reference software HEVC (HM) run on the same platform using the Random-Access configuration (the most relevant for broadcast and streaming applications).</p><h3>Beyond bitrate reduction</h3><p>Why a new video codec? Video is by far the most bandwidth demanding media that is distributed today through communication networks. In the last two decades video content has become the prominent type of data consumed over the Internet in a number of different forms. Usually we refer to a new generation of video compressors by its improvement in terms of coding efficiency. Bitrate reduction can directly produce cost reductions in terms of core network or CDN bandwidth, satellite&#39;s transponder or terrestrial mux occupancy and storage requirements but also has an implicit value increasing the reach of services, providing for example UHD content to customers that couldn’t have access to that service with legacy compression technologies, due to limited connectivity.</p><p>Bitrate reduction for the same visual quality, although important, is not the only motivation behind the creation of VVC. The MPEG output document “<a href="https://mpeg.chiariglione.org/standards/exploration/future-video-coding/requirements-a-future-video-coding-standard-v5">Requirements for a Future Video Coding Standard</a>“ outlines two major directions: the first is represented by the need of improving current services (broadcast, IPTV, OTT, videoconferencing, surveillance, etc.) supporting efficiently higher resolutions and content adaptation, reducing latency, providing better HDR content or mixed natural and synthetic content. The latter is to provide a versatile technology (it’s in the name of the codec, after all) to support new services that are today at early stages such as virtual and augmented reality (VR/AR), automotive and IoT media-related applications. The transition from traditional 2D video to immersive video enabling navigation of the content with 3 or more degrees of freedom requires new techniques allowing for example to selectively decode just a portion of the whole picture based on real-time information from head-mounted displays. These new ways of producing, organizing and consuming content have deep impacts on the design of new systems and video codec interfaces.</p><h3>What&#39;s inside VVC</h3><p>VVC is an evolution of the traditional block-based hybrid video coding architecture. The upcoming standard includes improvements, simplifications and new tools to push coding efficiency beyond the levels achieved today by HEVC. Almost all aspects of the coding process have been investigated and re-designed or refined:</p><ol><li>Partitioning</li><li>Intra prediction</li><li>Inter prediction</li><li>Transform and Quantization</li><li>Entropy coding</li><li>In-loop filters</li></ol><p>Picture partitioning is based on the concept of Coding Tree Units (CTU) and Coding Units (CU). The maximum size for a luma block in a CTU supported by the standard is 128x128. VVC defines a new quad-tree partitioning structure with nested multi-type tree structure to better adapt to local characteristics of the picture. CTUs are split in a quaternary tree structure, then each leaf in the coding tree can be recursively split in 4 different ways with 2 or 3 regions in horizontal or vertical directions. The resulting CUs are used for both prediction and transform operations (if they don&#39;t exceed the maximum transform size). It’s possible to define separate trees for luma and chroma components (CST).</p><p>Intra coding has been enhanced increasing the number of directional intra modes to 65 (compared to the 33 directional modes available in HEVC) in addition to planar and DC modes. The list of most probable modes (MPM) has been extended to 6 and a new wide-angle intra prediction mode has been defined for non-square blocks, based on the aspect-ratio of the block (while keeping the same number of prediction modes and coding methods). It&#39;s possible to apply to chroma components a cross-component linear model (CCLM) to derive the prediction from reconstructed luma samples in the same block. Other available intra coding tools are: mode dependent intra smoothing (MDIS) which uses a four-tap intra filter to improve prediction accuracy, matrix intra prediction (MIP), multiple reference line (MRL), that uses blocks in additional reference lines, position dependent intra prediction combination, and intra sub-partitions (ISP).</p><p>Inter prediction includes now several new techniques: extended merge prediction list for motion vectors and merge mode with motion vector differences (MVD) to better represent motion parameters obtained from neighbouring CUs, and the adaptive motion vector resolution (AMVR), allowing to code MVDs with different precision. Affine motion compensated predictions supporting zooming, rotation and other transformations in addition to simple translation motions, triangle partition to apply more precise predictions within a CU, and decoder side motion vector refinement (DMVR) which allows the decoder to search for better motion vectors around the ones signalled in the bitstream in the reference picture lists. Other available tools are: symmetric MVD (SMVD), bi-prediction with CU weights (BCW), subblock-based temporal merging, bi-directional optical flow (BDOF), and combined inter and intra prediction (CIIP).</p><p>VVC can use large block-size transforms up to 64x64 that are useful in particular at high resolutions and introduces new primary transforms: namely DST-VII and DCT-VIII. Another tool is the low frequency non-separable transform (LFNST), or reduced secondary transform, that can be applied between the transform and the quantization at the encoder side and between de-quantization and inverse transform at the decoder side. Sub-block transform (SBT) can be used for inter-predicted CUs, allowing to code just a portion of the residual block. Concerning quantization, the maximum QP in VVC is extended from 51 to 63 and a new method called dependent quantization (DQ) is included in order to reduce as much as possible the distortion of the reconstructed vectors by defining the available value levels for transform coefficients considering previous levels in reconstruction order. Chroma residuals can be coded jointly.</p><p>The Entropy Coding process adopts an improved version of CABAC. The core engine has been enhanced and the initialization algorithm has been re-designed: the decoding decision in VVC uses a new two state model for each context instead of a look-up table (LUT) as in HEVC.</p><p>A new in-loop filter is added after the deblocking and the Sample Adaptive Offset (SAO) filters already specified for HEVC: the Adaptive Loop Filter (ALF). ALF reduces blurring and ringing artifacts, and in general, those artifacts that are not removed by the two previous filters. The luma mapping and chroma scaling tool (LMCS) is also introduced before the in-loop filters to reshape luma components and improve rate distortion for both SDR and HDR content.</p><p>The High Level Syntax (HLS) of the video codec has been revised: in VVC CTUs can be organized in slices, bricks, tiles in order to improve robustness and parallel decoding operations and sub-pictures to enable independent extraction and decoding of portions of a picture (useful for example in omni-directional viewport-based application scenarios). Signalling is achieved with syntax structures such as sequence parameters set (SPS) and picture parameters sets (PPS), video parameter set (VPS) and the new adaptation parameter set (APS) and decoding parameter set (DPS). APS carries ALF, luma mapping with chroma scaling parameters and scaling list matrices, while DPS conveys bitstream constraints such as profile, tier and level, facilitating splicing operations. Non-normative SEI Messages are defined in a separate specification.</p><p>Another interesting feature in VVC is the reference picture resampling (RPR) mechanism: pictures in the reference lists can be stored at a different resolution from the current picture and then resampled in order to perform regular decoding operations. The inclusion of this technique supports interesting application scenarios such as real-time communication with adaptive resolution, adaptive streaming with open GOP structures, and enhanced omni-directional viewport-based streaming, for example allowing different temporal structures for different part of the picture. RPR allows also to support, from the first version of the codec, scalability (in particular spatial scalability). This is a major change compared to AVC and HEVC, where scalability was defined and introduced only after the completion of the first version of the standard, and supported by separate profiles.</p><p>VVC has been designed to support omni-directional applications, therefore provides a specific tool to improve the quality of the reconstruction of pictures containing equirectangular projections. This tool is called horizontal wrap around motion compensation and reduces the artifacts generated by the regular padding process for out of bounds samples, using information of the neighbouring samples in the projection domain.</p><p>A specific tool, namely the Intra Block Copy (IBC), is available for screen content coding. This tool, originally included in the HEVC Screen Content Coding Extensions (HEVC SCC), takes advantage of repeating patterns in text and graphics in the same picture, significantly improving the compression efficiency for this kind of content, in particular for All Intra configurations.</p><p>The set of tools listed in this section is included in the Committee Draft text submitted for ballot, however, modifications may be required later to address comments expressed by National Bodies or further refinements may be deemed necessary by the group and implemented during the next iterations and development stages.</p><h3>Conclusions</h3><p>VVC promises to renew the technical leadership of previous video compression technologies jointly developed by ISO/IEC MPEG and ITU-T VCEG. It will provide a solid base to build the next generation of immersive media applications and it will push TV, streaming, gaming and video communication services to new levels of engagement and user experience. However, a key aspect remains unclear (being outside JVET scope): the licensing terms for the usage of this technology. Companies behind HEVC have been criticized for disclosing public licensing terms in some cases years after the finalization of the standard, generating uncertainty in the market due to the existence of multiple patent pools with different licensing fee structures. To avoid this situation, companies active in VVC decided to create the <a href="https://www.mc-if.org/">Media Codec Industry Forum</a> (MC-IF), an independent industry consortium whose objective is to facilitate the adoption of this new codec, also discussing the most appropriate licensing options and their (hopefully) timely availability. Licensing terms will be certainly important to ensure a wide commercial adoption of VVC, also considering that other competing video compression solutions are now becoming mature.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=baeb13c8960a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Introducing MPEG-5]]></title>
            <link>https://medium.com/@gibellino/introducing-mpeg-5-5ed1de2be71b?source=rss-28b254ee3fba------2</link>
            <guid isPermaLink="false">https://medium.com/p/5ed1de2be71b</guid>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[streaming]]></category>
            <category><![CDATA[data]]></category>
            <category><![CDATA[multimedia]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Diego Gibellino]]></dc:creator>
            <pubDate>Mon, 04 Mar 2019 14:49:49 GMT</pubDate>
            <atom:updated>2019-03-06T19:19:18.285Z</atom:updated>
            <cc:license>https://creativecommons.org/licenses/by-nd/4.0/</cc:license>
            <content:encoded><![CDATA[<p><em>MPEG-5 Essential Video Coding (EVC) is a new codec under development in MPEG that promises to become a viable alternative to HEVC. With licensing terms clearly defined and at least a royalty-free basic set of tools, may it challenge AOMedia AV1 and VVC as video format of choice for building your next video streaming service?</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mopHrXr9DCSHAyFW8hzTZg.jpeg" /></figure><p>Coding efficiency has always been a major driving factor for the development of innovative algorithms in the video compression field. The market constantly demands increased capabilities in terms of compression ratio and this is becoming even more important with the introduction of new formats such as UHD 4K and 8K, enhanced technologies like High Dynamic Range video (HDR) and High Frame Rate (HFR), and new service scenarios (e.g. virtual reality). Today, digital video is distributed mainly in AVC, with HEVC adopted for 4K resolutions and beyond. If Versatile Video Coding (VVC), the new video codec under development by ISO/IEC MPEG and ITU-T VCEG Joint Video Experts Team (JVET), is the natural candidate to keep high the level of technical excellence achieved by AVC and HEVC, other kind of requirements from the industry are emerging.</p><p>A number of companies and entities are interested in solutions that potentially are less efficient than the state of the art, but provide in return other benefits in terms of flexibility and clear licensing model implications. This is particularly critical for certain markets or service segments (e.g. video streaming), or where licensing costs and uncertainty thereof may jeopardize an entire commercial initiative if not taken into account up-front.</p><p>In addition, alternative codecs such as AOMedia AV1, developed outside the ISO/IEC and ITU-T domains are starting to gain traction providing both a royalty free licensing schema and very good performances in terms of coding efficiency.</p><p>MPEG decided to launch a new initiative around video coding, with a set of very specific objectives collected in a requirements document [1]:</p><ul><li>Provide a technology with coding efficiency similar to that of HEVC</li><li>Support real time encoding and decoding scenarios</li><li>Offer clear licensing conditions and make them available in a timely manner</li><li>Finalize the text of the standard in early 2020</li></ul><p>A Call for Proposals [2] was announced in October 2018, with responses analyzed during the MPEG meeting of January 2019. Evidence gathered during the evaluation process showed that proposals received were able to achieve up to 24% bitrate savings with respect to HEVC 10 Main for the same quality (on BD-rate basis) with all the tools enabled, and up to 30% with respect to AVC when only basic royalty free tools were used. The first version of the EVC working draft is available as an internal output document in MPEG [3], as well as the related test model (ETM 1.0). The codec is currently structured in a baseline profile, including tools with expired patents or royalty free, and a main profile adding a few new tools. One interesting feature is that EVC will be able to selectively activate or switch off individually each tool beyond the baseline profile. This mechanism, supported directly in the bitstream, would ensure flexibility and, at the same time, would allow to remove tools that cannot be licensed under reasonable terms after the finalization of the standard (although EVC contributors are invited to provide licensing terms within a reasonable time frame).</p><p>EVC activities have really just started and it’s very early to say whether this new MPEG codec will become a success. It’s however important that MPEG has recognized that an alternative should be built and made available as ISO/IEC Standard. In 2020, the date planned for its finalization at Final Draft International Standard (FDIS), EVC will try and find a place in a codec lineup that already include VVC and AOMedia AV1 (with the latter that, at that time, will be mature and already available in production for some time). Much will depend on the level of licensing complexity and uncertainty that VVC will inherit from HEVC and on the support for optimized implementations of AV1. EVC (as VVC) will also offer the benefit of being a recognized international standard, an important requirement for some specific markets with regulatory constraints (e.g. broadcast TV).</p><p>It seems that the future for video services will be a<em> multi-codec</em> world without a single technology ruling across all the different service segments (as we saw in the past with AVC). Operators and service providers shall carefully select the most suitable compression technology for each kind of service offer or deployment scenario, and content preparation workflows and infrastructures need to be ready to support such a change.</p><h3>References</h3><p>[1] “Requirements for a New Video Coding standard”, N17928; MPEG 124, Macau, October 2018 <a href="https://mpeg.chiariglione.org/standards/exploration/future-video-coding/requirements-a-new-video-coding-standard">https://mpeg.chiariglione.org/standards/exploration/future-video-coding/requirements-a-new-video-coding-standard</a></p><p>[2] “Call for Proposals for a New Video Coding Standard”, N18075; MPEG 124, Macau, October 2018</p><p>[3] “Working Draft 1 of Essential Video Coding”, N18283; MPEG 125, Marrakesh, January 2019</p><figure><a href="https://usejournal.com/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_image"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*f2IVAl0TbsfES9cFGYr40g.png" /></a></figure><p>📝 Read this story later in <a href="https://usejournal.com/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_read_later_text">Journal</a>.</p><p>🗞 Wake up every Sunday morning to the week’s most noteworthy Tech stories, opinions, and news waiting in your inbox: <a href="https://usejournal.com/newsletter/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_text">Get the noteworthy newsletter &gt;</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5ed1de2be71b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How 5G will change your TV for good]]></title>
            <link>https://medium.com/@gibellino/how-5g-will-change-your-tv-for-good-96ad1e42eea2?source=rss-28b254ee3fba------2</link>
            <guid isPermaLink="false">https://medium.com/p/96ad1e42eea2</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[broadcast]]></category>
            <category><![CDATA[wireless]]></category>
            <category><![CDATA[mobile]]></category>
            <category><![CDATA[5g]]></category>
            <dc:creator><![CDATA[Diego Gibellino]]></dc:creator>
            <pubDate>Thu, 31 Jan 2019 12:35:06 GMT</pubDate>
            <atom:updated>2019-02-04T18:35:01.621Z</atom:updated>
            <cc:license>https://creativecommons.org/licenses/by-nd/4.0/</cc:license>
            <content:encoded><![CDATA[<p><em>The introduction of 5G is setting off a storm of hype around this technology: more bandwidth, reduced latency, and greater reliability and security are just some of the benefits the next generation communication system will be able to provide. But there is more: a new model for TV production, distribution and consumption.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*p9g3FKDow9TiMilH" /></figure><p>Whether you were at CES 2019 in Las Vegas or not, you should already know that this year the keyword of the show has been 5G. Announcements by major mobile operators and phone manufacturers seem to suggest that commercialization of 5G will start sooner than expected, with an increasing number of deployments planned from 2019–2020 onward, initially based on the “<em>5G early drop</em>” 3GPP specifications.</p><p>5G is designed to improve every single aspect of the LTE mobile network, including a new radio access technology (5G-NR). Requirements cover three main classes of services:</p><ul><li><strong>Enhanced Mobile Broadband (eMBB): </strong>access to multimedia content, services and data by humans</li><li><strong>Massive Machine Type Communications (mMTC):</strong> data exchange among low-power devices with minimal throughput and relaxed constraints in terms of delay.</li><li><strong>Ultra-Reliable Low Latency Communications (URLLC): </strong>connectivity with strict levels of availability and latency (autonomous vehicles, public safety, industrial control, robotics and drones, remote medical surgery)</li></ul><p>As defined by 3GPP, these three, very different, classes shall all be offered by a single mobile network. The 5G platform is designed to be tuned and dynamically configured in multiple logically partitioned networks with different characteristics and customized QoS levels using a mechanism called <em>network slicing.</em></p><p>Delivery of video streaming will exploit eMBB capabilities, that will be also the first target for early market deployments. 5G will be the platform for distributing high quality content to mobile devices and, with the so-called <em>fixed wireless</em> <em>access</em>, to connected TV or set top boxes in homes.</p><p>But this new platform, with its high spectrum efficiency, can represent an interesting way forward also for traditional TV broadcast terrestrial delivery (currently using DVB-T/T2, ATSC or ISDB-T technologies). In Europe, with the progressive transition of the 700 MHz spectrum to mobile broadband (to be completed by 2020–2022 according to EU rules) and the increasing demand in terms of bandwidth to provide high quality content, broadcasters are evaluating alternatives.</p><p>Will 5G networks be able to support broadcast linear TV services and their requirements? Is a progressive convergence to IP-based TV services in synergy with mobile operators, a sustainable model for the future of TV? And if yes, will it happen now?</p><h3>A bit of history</h3><p>In early 2000s, a collective effort to provide a means to enable TV on mobile phones produced the first DVB Handheld specification (DVB-H). The specification, published as ETSI standard in 2004, defined an additional module to be included in compliant mobile devices. The DVB-H demodulator, conceptually implementing a superset of the digital terrestrial receiver functionalities, was based on the same principles, trying to protect broadcasters’ investment and transmission equipment by ensuring compatibility with DVB-T transmissions.</p><p>This initiative didn’t take off. Commercial services were launched in a number of countries in Europe but scarcity of devices with DVB-H stacks and, above all, lack of a sounding and profitable business model, forced operators to shut down them.</p><p>A different approach was pursued by 3GPP with the Multimedia Broadcast Multicast Service (MBMS). In this case, the specification extended 3G UMTS network capabilities to support point-to-multipoint (PMP) distribution though severe constraints were imposed by limited bandwidth supported and static allocation of the spectrum.</p><h3>Evolved MBMS</h3><p>The Evolved MBMS (eMBMS), is an extension of the 4G LTE system (starting from 3GPP Release 9 specifications). With eMBMS, an operator can decide to allocate a part of the bandwidth available in downlink in a cell (up to 60%) to broadcast traffic that will serve multiple devices with the same content: video, audio, files, etc... Data is delivered in a one-way fashion, each operator is responsible for the allocation of the multicast portion within his network, and content cannot be shared among different networks (needs to be replicated).</p><p>Content delivery is based on MPEG-DASH adaptive streaming but with a single representation or quality level per content. eMBMS doesn’t use HTTP, but instead adopts the FLUTE mechanism, designed to efficiently distribute files to devices (chunks in case of adaptive streaming).</p><p>Subsequent additions defined the<em> MBMS operation on demand</em> (<em>MOOD)</em>, a mechanism to dynamically switch from unicast delivery to multicast when the number of devices accessing a particular content goes over a defined threshold. For example, when at least 3 users are watching the same content via unicast in an area, it may be more convenient to provide that content once via multicast, avoiding wasting resources, and transparently migrate those devices and users to this new stream.</p><p>A typical service scenario for eMBMS adoption, in addition to mobile TV, is “Venue casting”, where in a relatively limited geographical area like a stadium, a campus, or an exposition hall, a portion of the spectrum is reserved to multicast just for specific events. The audience can access additional content in real-time, provided as one or more linear channels available on their devices and showing, for example, instant replays or different views of the event. Each channel is delivered as a multicast stream potentially serving thousands of users without incurring in network congestion issues and service degradation.</p><p>The first eMBMS commercial service was launched in 2014, after several trials carried out by operators around the world. eMBMS requires the deployment of additional elements in the LTE core network architecture. Although these operations can be performed via software updates, a compliant user device is still required in order to access services. Verizon, in USA, attempted to commercialize a large scale eMBMS service (Go90) but eventually discontinued it. eMBMS adoption today, where existing infrastructures are already installed, is mostly confined to selected areas and enabled on event basis only.</p><p>Why isn’ t this model working? The first consideration is that launching widely available mobile linear TV services with eMBMS today doesn’t scale in terms of revenues to justify investments in the infrastructure either with Ad-funded or subscription-based models. Comparable live services delivered with the unicast OTT approach are still more profitable considering costs and potential user base. Another critical aspect is the lack of support in devices. eMBMS API has been recently included in Android, starting from version 8.1, but the number of devices supporting those capabilities is by far too limited, with some products disabling the features even if the underlying chipset would be eMBMS capable.</p><h3>Towards a mobile network for Broadcast TV</h3><p>In 2017, with Release 14, 3GPP introduced the <em>Enhanced TV (enTV) </em>feature: a new set of eMBMS extensions implementing a full broadcast TV stack in the LTE mobile network. Major additions are:</p><ul><li>A model supporting High Power High Tower (HPHT) transmitters, in addition to cellular cells, to create overlay networks able to cooperate and provide the required coverage in both rural and urban areas</li><li>Larger Inter-Site Distance (ISD) at high spectral efficiency</li><li>Support for up to 100% eMBMS carrier allocation to create dedicated broadcast networks</li><li>A receive-only mode, in which devices can view broadcast content even without a SIM</li><li>A transport mode that allows reuse of broadcast services without transcoding thus ensuring backward compatibility and minimal effort to migrate from legacy transmission systems</li><li>A shared broadcast network, where different operators can aggregate their MBMS networks to create a common distribution platform without the need to replicate content</li></ul><p>These additions made LTE, in its latest version, able to support both mixed multicast/unicast and standalone terrestrial broadcast services, defining a truly converged platform for content distribution. The European Broadcasting Union (EBU) and several broadcasters played an important role in those developments, ensuring that EU digital TV broadcast requirements, including public service regulatory constraints such as Free-to-air channels availability, were met.</p><h3>5G &amp; TV</h3><p>A number of research projects are studying, prototyping and evaluating enTV in the context of LTE and upcoming 5G deployments: examples are <a href="https://5g-xcast.eu/">5G-Xcast</a>, a H2020/5G-PPP project partially-funded by the EU; 5G Today, a project currently ongoing in Bavaria, Germany, and the 5GTN+ project in Finland. One of the objectives is to find the best solutions to exploit 5G enhanced capabilities, in particular using the new radio interface (5G-NR) for broadcast TV services.</p><p>3GPP is working on the evolution of enTV with a study on <em>LTE-based 5G terrestrial broadcast</em>, which should highlight gaps to match requirements already defined for 5G, and allow required updates to be integrated in Release 16 (scheduled for end 2019), whereas the design of a 5G-native eMBMS broadcast mode has been postponed after 2019. This means that broadcast capabilities on 5G-NR (for TV, but also for other services such as IoT or V2X) will not be available as part of the first 5G deployments.</p><p>In parallel, an increasing interest has been shown for the adoption of 5G as communication infrastructure for production and transport of audiovisual content. 5G, and the adoption of new, more efficient video compression technologies, can provide a great opportunity to simplify and enhance content production and contribution, reducing costs. Today live events with remote production require complicated and expensive setups that include wireless cameras with video relays to OB vans at the venue and satellite or fixed links to studios. This scenario can be greatly simplified using 5G. 5G capabilities will allow the use of a single, reliable communication link with high capacity and very low latency, to connect cameras directly to remote production studios. The same approach can be used also for studio-based production, without the need of local wired infrastructures, and it’s aligned with current market trends that are aiming at shifting audiovisual production to all-IP workflows.</p><p>These requirements, that include also synchronization of feeds and management of QoS and security, are part of an ongoing 3GPP <em>Feasibility Study on Audio-Visual Service Production</em>, promoted by EBU and other members.</p><h3>Conclusions</h3><p>One of the key features of 5G, besides the unprecedented bandwidth capacity, reduced latency and reliability, is its flexibility in supporting very different service scenarios. We can consider broadcast TV as yet another vertical that will be supported by 5G using the <em>network slic</em>i<em>ng</em> capability, with the additional benefit of taking advantage of unicast streaming where and when convenient. The mixed multicast/unicast approach, has in this context, a major role to play. If 5G spectrum efficiency can be appealing for a transition of the digital terrestrial television to a mobile network, it’s the dynamic orchestration of broadcast and unicast capabilities that promise to change the game. With the expected explosion of high-quality video consumption, available to any connected device, a fine-grained and on-demand management of the broadcast network resources will become necessary for operators, as an additional tool for network optimization and scaling.</p><p>The full picture of TV on 5G is yet to be developed and it will not be shipped with the first wave of 5G services, but significant technologies are already mature and can be applied successfully in the transition from LTE networks. Wide support on devices remains a critical aspect. Convergence will impact distribution of TV and linear services to users but also production and contribution, with envisaged benefits in quality, flexibility and reduction of costs.</p><p>Business arrangements between broadcasters or content providers and mobile operators will likely be required to realize this convergent scenario. Different models can be considered: ad-hoc deployments leveraging HPHT overlay networks, partnerships, or the addition of TV broadcast distribution “<em>as a service</em>” to mobile operators’ portfolios. Decisions and resulting market scenarios will depend largely also on regulatory directives in terms of spectrum allocation, public service availability, and net neutrality.</p><figure><a href="https://usejournal.com/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_image"><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*f2IVAl0TbsfES9cFGYr40g.png" /></a></figure><p>📝 Read this story later in <a href="https://usejournal.com/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_read_later_text">Journal</a>.</p><p>🗞 Wake up every Sunday morning to the week’s most noteworthy Tech stories, opinions, and news waiting in your inbox: <a href="https://usejournal.com/newsletter/?utm_source=medium.com&amp;utm_medium=noteworthy_blog&amp;utm_campaign=guest_post_text">Get the noteworthy newsletter &gt;</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=96ad1e42eea2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[3 Video Streaming Technologies for 2019]]></title>
            <link>https://medium.com/@gibellino/3-video-streaming-technologies-for-2019-d51b6b1223d5?source=rss-28b254ee3fba------2</link>
            <guid isPermaLink="false">https://medium.com/p/d51b6b1223d5</guid>
            <category><![CDATA[tech]]></category>
            <category><![CDATA[streaming]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[ott]]></category>
            <category><![CDATA[compression]]></category>
            <dc:creator><![CDATA[Diego Gibellino]]></dc:creator>
            <pubDate>Thu, 03 Jan 2019 14:02:23 GMT</pubDate>
            <atom:updated>2019-01-03T14:09:34.910Z</atom:updated>
            <cc:license>https://creativecommons.org/licenses/by-nd/4.0/</cc:license>
            <content:encoded><![CDATA[<p><em>The new year is here and there is no better time to list 3 of the most promising technologies I expect to see in production in the video streaming arena in the next months…</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*lLvo4eR4vwIbUukd" /></figure><p>The Video Streaming Industry keeps experimenting and introducing innovative technologies, however the transition to production is not always as smooth as we would like or expect. There are implications in terms of costs, support by consumer devices, impact on established architectures, or, in some cases, a technology is simply not good enough for the market yet. The three technologies described below have an opportunity window in 2019. We will see how many services will embrace them.</p><h3>Low Latency for OTT Live Services</h3><p>In 2018, an increasing number of players offering OTT commercial services included live events in their catalog. In addition to traditional IPTV, Cable, Satellite and Digital Terrestrial distribution, service providers operating mainly on the Internet, managed to acquire rights for Prime-time sport events such as Football, NFL, etc... These events were distributed through adaptive streaming, mainly using MPEG DASH and HLS, to a wide range of different devices, from smartphones to connected TVs, taking advantage of one or more CDNs (trying) to cope with the traffic demand.</p><p>It turned out, however, that a major issue for this kind of services was the end-to-end latency or delay experienced by the user, in particular when compared to traditional broadcast distribution. Currently deployed HTTP-based adaptive streaming distribution architectures are usually behind broadcast of about 15–30 seconds, reaching 1 minute or more in some cases. This is considered a key aspect in terms of user acceptance, because subscribers are certainly not very happy to pay to watch a team score when their neighbors are already celebrating the goal.</p><p>Technologies exist to reduce latency, and in particular, MPEG DASH provided extensions since the publication of the Second Edition of the Standard, in 2014. However, only in 2018, with the introduction of Common Media Application Format (MPEG CMAF) support in products, we have seen implementations tackling this issue in all the important building blocks of the architecture: Encoder, Packager, CDN, and Player. First results are promising, with end-to-end delays comparable to those provided by satellite or digital terrestrial (5 seconds or even below) and demonstrators able to reach ultra-low latency (below 1 second).</p><p>In 2019 I expect to see production-ready solutions able to exploit CMAF low latency mode and commercial services able to deliver Prime-time sport events with reduced latency. It will be interesting to look at how far we can go in real services finding the right trade-off between latency, solution stability and robustness.</p><h3>Advancements in Video Compression</h3><p>This year we’ve witnessed the very first signs of the upcoming battle for the next generation video codec: VVC Vs AOM AV1.</p><p>In April 2018, after a first exploration phase and a formal Call for Proposal, ISO/IEC MPEG and ITU-T VCEG launched the Joint Video Experts Team (JVET), whose objective is to develop the video codec successor to HEVC: Versatile Video Coding <em>aka</em> VVC. The VVC standard will be finalized by the end of 2020 and preliminary results show that it will be possible to increase the coding efficiency of HEVC of another 40–50% in terms of bitrate for an equivalent visual quality, with a reasonable increase in codec complexity. VVC is designed to fulfil the requirements of modern and next generation video services, including broadcast and unicast streaming delivery of Standard Dynamic Range (SDR) and High Dynamic Range (HDR) content, and 360 Video.</p><p><a href="https://aomedia.org/">The Alliance for Open Media</a> (AOM) is an initiative created in 2015 by a number of important players in the video streaming and technology markets like Google, Mozilla, Netflix, etc…, with the objective of creating a royalty free video codec solution able to improve the current state of the art solutions. In 2018, AOM published version 1.0 of AV1 along with the software codebase available to all interested parties (aomenc).</p><p>A number of scientific papers and studies compared AV1 and the software developed by JVET in the exploration phase, called Joint Exploration Model (JEM). Others compared AV1 and Reference software for AVC and HEVC, or AV1 with implementations such as x264 and x265. Relevant data were presented and discussed at IBC this year and in other conferences but figures, as always with codec comparison, largely depend on the target parameters and usage scenario considered by the authors (e.g. fixed QP, rate control, 1 or 2-pass etc…). However, AV1 has the potential to become at least an alternative to HEVC in terms of coding efficiency, although some additional work on optimization and hardware support is still needed and already planned.</p><p>2019 will not be the end of the battle, but it will certainly shed some more light on the contenders: AOM AV1 is expected to reach some level of maturity through commercial implementations and better hardware support, answering to concerns on encoder complexity raised by some studies. Commercial services will start using AV1 (some sample content is already available), at least for VoD content, taking advantage of browsers such as Chrome and Firefox that already provide decoding capabilities.</p><p>VVC, on the other hand, will enter in the most important phase for its development, where critical decisions in terms of tools selection and complexity will be made. The decisions in JVET will define VVC performances, and probably, the video landscape in the next 5–8 years. It is therefore good to have some competition: it will help to raise the bar even more.</p><p>One of the major concerns about the future of VVC is the licensing terms that will be defined for it. The development of the Standard is still ongoing, therefore nothing exists at this point yet, but VVC will likely require the payment of royalties to a group of patent owners. This mechanism proved to work well for previous codecs such as AVC or MPEG-2 Video but showed some shortcomings for HEVC. HEVC licensing is very difficult to understand and manage, with three patent pools and a number of companies that keep their IP aside. This situation is one of the reasons why HEVC deployment in the market is proceeding at a relatively slow pace since the Standard publication, in 2013.</p><p>A group of companies recently created the <a href="http://www.mc-if.org/">Media Coding Industry Forum</a> (MC-IF), a forum that is willing to promote standard Media Coding solutions to the Industry and where parties involved can discuss issues related to the licensing of intellectual property rights relevant to the deployment and use of future Standards (e.g. VVC). How this and other initiatives may influence the licensing terms of forthcoming coding standards has to be seen, but it is definitely a step forward in the right direction.</p><h3>Content Adaptive Encoding (CAE)</h3><p>Content Adaptive Encoding is not a new thing. Netflix described this technique in 2015, with a blog post titled<a href="https://medium.com/netflix-techblog/per-title-encode-optimization-7e99442b62a2"> Per-Title Encode Optimization</a>. The key concept is that video content may have different levels of visual complexity. Some content, when encoded, requires more or less bits (and bitrate) to achieve good or acceptable quality. If we are able to automate this decision each content may be encoded with just the right number of bits, and for adaptive streaming, the quality of each level of representation, also collectively known as the bitrate ladder, can be customized accordingly.</p><p>How you can estimate content complexity efficiently and automatically in your transcoding workflow makes the difference. It is possible to use brute force algorithms based on a number of trial encoding during the ingestion phase and then analyze the resulting bitstreams with your preferred objective metric. Another approach is to adopt sophisticated Artificial Intelligence algorithms, typically based on Neural Networks, to extract meaningful video features from the ingested content, apply a model and constantly refine and improve it to match your content catalog. Impacts in terms of computation power, storage space, modifications to the workflow and digital asset management structure and publishing process may be different, and the approach to follow should be carefully evaluated and selected.</p><p>Some service providers designed their own internal solution or they are currently working on it, but there are also products provided by encoder or technology platform suppliers, already integrated in their transcoding workflow. Whatever approach you prefer, be prepared to see CAE in production at scale in 2019.</p><h3>Conclusions</h3><p>Low latency, new video codecs, and Content Adaptive Encoding are three technologies that will likely be adopted by an increasing number of commercial video streaming services in 2019.</p><p>Solutions and products exist (except for VVC, of course, which is planned for end 2020), or are expected to reach maturity in the next months , allowing integration in production environments.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d51b6b1223d5" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>