Ultra HD Forum Guidelines: The UHD bible
During the past NAB show in Las Vegas, the Ultra HD Forum -a conglomerate of companies and experts dedicated to studying and promoting implementation of ultra high definition technologies- disclosed version one of its guidelines. Let us look at it in more detail.
By Yeray Alfageme, Service Manager at Olympic Channel
This document is the result of four years of work by the Guidelines Work Group. Previously, separate recommendations had been published about HDR, WCG and other technologies being applied to UHD. This is the first time that a combined document including everything is released, now that UHD is finally being taken seriously.
The goal of this document is to describe methods for creation and distribution of content toward consumers with uniform features in such a way that said content can be regarded as UHD. This comprises both distribution through the Internet and through satellite, terrestrial and cable, but not encoding systems or file formats. More specifically, the document covers:
- Production of recorded and live content:
- Grading of recorded content.
- HDR and WCG.
- Immersive audio.
- Distribution and compression.
- Master format.
- Mezzanine format.
- Encoding methods, codecs and recommendations.
- Transcoding methods, codecs and recommendations.
- Bitrate ranges recommended throughout the chain.
- Distribution and transport.
- Linear programming of content.
- Conversion between SDR and HDR, in various formats.
- Interconnection of distribution production systems (HDMI and similar).
- Backward compatibility with old systems.
Fall outside the document topics such as:
- Recording techniques (lighting, camera configuration, …).
- TV configuration.
- Encoder configuration.
- Quality control.
- TV technologies (OLED or Quantum dots, for example).
- Color grading techniques.
- Digital cinema.
For all this, they have based on the current standards as published by ITU, which are shown in the table 1.
The forum suggest regarding most technologies within the UHD universe as foundations, as for instance, the Wide Color Gamut (WCG). Are typically regarded as foundations all technologies existing by 2016 and therefore, with broad adoption and maturity levels.
Therefore, the following minimum parameters are set in order to be regarded as UHD:
- Resolution: Equal to or higher than 1080p and lower than or equal to 2160p, progressive at all times, as BT.2100 does not include interlaced scan.
- Wide Color Gamut (WCG): gammas greater than the specification BT.709.
- High Dynamic Range (HDR): ranges equal to or greater than 13 f-steps.
- Bit depth: 10 bits.
- Frame rates: up to 60 frames per second, preferably integer values.
- Audio: 5.1 or channel-based immersive audio, although stereo audio 2.0 is acceptable.
- Subtitles: CTA 708/608, ETSI 300 743, ETSI 300 472, SCTE-27 and IMSC1.
Being more specific, permitted values would be of Table 2.
As a result, we can combine these parameters in order to obtain content regarded as UHD, as for instance, HD-HDR, which would have a 1080p resolution and a BT.709 traditional color space but a HLG dynamic range, in the fashion in which many live sport events are produced.
As for distribution encoding, HEVC is regarded as the foundation codec for UHD as it supports both HDR and WCG in 10 bits, which enables reaching up to 4K resolution with an admissible bandwidth consumption. For production, contribution and mezzanine format, the AVC codec is the recommended choice as it supports a 10-bit depth in addition to the BT.2020 color space as well as PG and HLG gamma curves., which makes feasible using said codec for encoding HDR content. Additionally, upgrading current AVC encoders to support UHD is feasible through software updates.
All this combination of possibilities translates into the table 3 of formats available in UHD.
Besides these foundation technologies, the document also provides recommendations on additional technologies that can be used, as for example static metadata or absence thereof, in HDR content based on a PQ10 gamma curve, for instance.
The chart 1 included in the document shows a generic block diagram for a production chain, either UHD or not, and can be used as basis for analyzing each individual component.
Let us see, for each individual area, specific aspects on the various technologies.
High Dynamic Range (HDR) and Wide Color Gammut (WCG)
Undoubtedly, these two technologies are the ones bearing most implications -along with resolution- in generation of UHD content. In regard to HDR, the document provides a detailed description of features and peculiarities, both for PQ curves –as PQ10- and Hybrid Log-Gamma (HLG), HLG10. It makes recommendations based on the ITU-R BT.2100 standar and on the metadata to be used for this kind of signals, as specified by the SMPTE in the document ST 2086.
Before going into details such as luminance peaks or video over IP, on which the document also provides interesting contributions and explores the HDR10 environment, a factual standard in respect to consumer TV sets, along with its evolution -already dealt with in TM Broadcast- HDR10+.
In regard to HDR, the document devotes a full section to the other standard found in the industry, Dolby Vision. Due to the particular features of this technology, it deserves to be considered on its own and attention paid to the way in which it may be implemented in Broadcast.
Production of recorded and live content
Recommendations on production of content, either recorded or live, focus on the requirements to be met both by cameras and mobile units as well as production environments so as to achieve a proper UHD production. Subject of study are both the HDR profiles to be used -either PQ or HDR- and grading, as well as channel-based immersive audio, a foundation technology of UHD.
Essential aspects that are often taken for granted, such as encryption or watermark protection of content, also deserve attention. In this instance, the document provides a study case that sets an example and clarifies the use of content protection technologies as applied to UHD and to other formats.
Well-known standards such as Data Encryption Standard and DVB CSA can be used in UHD along with AES, at all times in their correct versions, such as DVB-CSA3 or DVB-CISSA.
Also considered are other digital watermarking technologies such as Digital Rights Management (DRM) or Conditional Access Systems (CAS) with the aim of detecting inadequate distribution of content, piracy.
Linear programming of content
When facing linear programming -in which various types of content are combined, even within the same image, such as graphics, advertisement or emergency marks or similar items- keeping consistency and even an adequate conversion between various contents is a crucial issue in UHD, as a wide array of possibilities is available for producing our signal.
The document covers both HDR and WCG parameter matching concepts for converting from one content type to another, as well as existing conversion possibilities between SDR or BT.709 content and HDR PQ10 or HLG10.
Distribution covers both what we regard as contribution between the point of origin for signal production and the Broadcast Center, as well as final distribution of content up to end consumers.
Study is made of notions relating transport over SDI, based on 2-sample interleave or by quadrants and IP, in addition to compression techniques based on said HEVC or AVC codecs, as well as adaptive bitrate (ABR) concepts for bandwidth optimization or audio and subtitle encoding.
To put it in proper context, note that under HEVC, 2160p HDR content would require a bandwidth of approximately 50-80 Mbps, while in AVC, it would require about 90-140. It is noticeable that HEVC is much better at optimizing available bandwidth, with the resulting savings in transport costs, either through satellite or fiber, or even through the Internet via streaming.
In regard to final distribution, suggestion is made to use MPEG-2 TS based on Phase 2 DVB UHD-1 specifications, even for distribution through IP. For OTT environments, MPEF DASH is the suggested format.
Decoding and Rendering
This section targets decoding and rendering at end consumers by providing set top boxes with adequate capabilities for being able to display the signal properly and connecting with the end TV sets.
Mention is made to rendering, as in the program’s final composition could be the case that certain contents -such as advertisements or graphics in the consumers’ desired language- are directly generated on the player and not at the broadcast center, so certain standards must be followed in order to achieve a correct display of the final image.
Interoperability of formats
When talking about interoperability of formats, concepts like conversions, scaling and similar terms always come to our mind. This section deals precisely with those topics.
Special consideration is made of scaling concepts, either ‘upward’ (i.e. upscaling of older content to UHD) or ‘downward’ (downscaling of UHD content to be played in older equipment not supporting WCG or HDR. As these are conversions going beyond definition because they also have an impact on image color or luminance, they should be performed carefully. Adequate workflows and equipment must be selected in order to avoid spoiling the quality of our image.
High Frame Rate (HFR)
Although it may seem that HFR is something of lesser importance in this whole environment, it must be kept in mind that this is the second technology -only behind definition- requiring a bigger increase in bandwidth for practical implementation. For this reason, proper implementation is critical. HFR has particular impact in broadcasting sport events and, most especially, care must be taken when aiming at playing more frames than supported in old devices. Backward compatibility must be taken into account at all times.
Next Generation Audio (NGA)
Audio, the forgotten issue. As it happens throughout the industry, not even 5% of the document is devoted to audio, something that is quite important indeed. Ultra HD offers Next generation Audio (NGA), a feature providing sound with immersive capabilities that are customizable depending on listener, flexibility for reproduction and rendering, and even different sound spaces, depending on the scene.
Because when dealing with NGA no mention is any longer made about channels although the forum recommends channel-based immersive audio within its foundation technologies, but about objects that are distributed throughout the sound space on the various channels.
To such purpose, recommendation is made to use at least 5.1 sound, although stereo sound 2.0 is also available. Please note that NHK, for instance, is already working in 22.2 sound spaces, with 24 independent audio channels, thus taking the full immersive feeling to the highest degree.
For all these, two standards –MPEG-H and Dolby AC-4- are available and widely covered in the document.
Content Aware Encoding (CAE)
Last, the forum makes recommendations on adaptive encoding, a novel technology that enables adapting the compression profile based on the content to be processed.
This technology combines novel concepts such as artificial intelligence –which enables automatic detection of the content to be compressed- and machine learning –which enables gradual optimization of the compression algorithm dynamically as more content is increasingly being processed.
This is particularly relevant when making distribution of content to mobile devices featuring limited bandwidth or in environments which infrastructure provides no assurance as to available capacity, although this technology is at a very early stage.
The full 187-page document is available from the UHD Forum website at http://www.utrahdforum.org. This is a must-read for all broadcast professionals and devotees who are excited about this new production format that is not the future, but the present.
It is somewhat funny to witness how an aspect as critical as it was definition goes unnoticed throughout the whole document and is instead taken for granted because, as we already remarked in previous articles, the new UHD standards are almost agnostic regarding definition. From HD to 8K, through 2K and 4K, any resolution is acceptable and changes nothing in regard to other topics such as HDR or WCG. Remember, better pixels instead of just more pixels.