Application Note April 2013 Ultra-High Definition TV Acquisition and Exchange
The TV industry is moving towards the next generation of high quality TV screen technology; Ultra High Definition Television (UHDTV). UHDTV is no longer just an experimental technology; it is part of every manufacturer s future plans. Ericsson s 2012 ConsumerLab TV and Video Consumer report highlights a high willingness among consumers to pay for UHDTV extreme quality as part of their TV and video service. UHDTV presents a stunning improvement on the quality and immersive experience of even today s HD TVs, and it s not difficult to see why broadcasters particularly those offering premium content such as sports broadcasters are looking at ways to adopt this technology in the not too distant future. As the prospect of 4K UHDTV grows over the next few years, operators and service providers will need to begin to recognize the importance of delivering the next-generation television experience that consumers demand. According to a January 2013 report from Deloitte, a number of 4K TV landmarks will emerge this year, including: > Around 20 4K TV set models from more than 10 vendors (by the end of 2013). > A range of 4K content made available (initially blockbuster movies) > The likelihood of multiple test broadcasts of UHDTV, with commercial services expected 2015. Going forward it s expected that the rapid decrease in the price of TV sets will boost penetration in consumers homes, hence increasing demand for premium content. A StrategyAnalytics [December 2012] report predicts a unit price reduction of over 90% over the next 5 years, increasing the installed base to over 20 million units worldwide, most above 60 inch in size. We are at important junction for UHDTV, where broadcasters must prepare for future success now. They must gain experience of UHDTV work flows and build a library of contribution quality UHDTV content ready for the launch of UHDTV direct to the home services, predicted for 2015. Ultra-High Definition TV Acquisition and Exchange 2
1. What is Ultra-High Definition Television? Ultra-High Definition Television (UHDTV) means television that is using any resolution that is above currently broadcast High Definition (HD) Television formats. Currently, the maximum transmitted HD resolution is 1920 pixels x1080 lines running at up to 29.97 frames per second, so any Horizontal x Vertical resolution above that is considered UHDTV. 1.1 What about 1080p59.94/50? 1920x1080 at frame rates greater than 29.97 frames/sec progressive are not transmitted to the home at the time of writing. Deployed consumer set top boxes are in general not capable of decoding this format. This format may have uses for contribution links, however, as it can be used as a mastering format for both 1080i and 720p standards, particularly using 10-bit 4:2:2 H.264 formats. Either UHDTV or 1080p50/59.94 may make sense for different operators and in fact there is no reason why both can not co-exist. 2. Why UHDTV? UHDTV as an origination format has a number of advantages: Master copy in a higher resolution > Long before HD was delivered to the home, content was filmed and produced in HD, because filming in SD would mean that no higher resolution would ever be available Better standards & format conversion > Deriving 1080i at 25Hz or 29.97Hz can be done with higher quality, starting with a UHDTV Level 1 master > 720p at 50 or 59.94 Hz can likewise be derived with higher quality Electronic pan/zoom > It is possible to use UHDTV Level 1 as an origination format, with the subsequent extraction of conventional HD content by electronically selecting part of the overall content. This has the potential to reduce operator cost in the field (ie at venues) Ultra-High Definition TV Acquisition and Exchange 3
3. Content Some content producers are already filming in UHDTV, for example some sitcoms are being recorded and produced in UHDTV, with subsequent down-conversion to conventional HD. Also, a few prototype outside broadcast events have taken place in UHDTV, not least the NHK/BBC Olympics feeds. Test uplinks have been transmitted by satellite, for example by EBU, Eutelsat and others. 4. Terminology, Resolutions & Frame Rates ITU has recently defined [1] two specific formats for use in UHDTV systems: UHDTV Level 1 > Resolution = 3840 x 2160 UHDTV Level 2 > Resolution = 7680 x 4320 4.1 UHDTV LEVEL 1 This is also known colloquially as 4K, however this is a somewhat ambiguous term. 3840x2160 is the only 4K resolution for television use. The resolution is exactly 2x horizontal and 2x vertical compared to the existing 1920x1080 HD resolution. The aspect ratio, like HD, is 16:9. Although there are several possible frame rates, including 23.976Hz for film content, the primary frame rates of interest for television content are 50.0Hz and 59.94Hz (so 2x frame rate of HD). The baseband data rate is therefore 8x that of conventional HD, so approximately 12Gbit/s. UHDTV Level 1 equipment exists today. The term UHDTV Level 1 is not commonly used, but instead UHDTV is generally taken to mean UHDTV Level 1. The majority of this document will consider only UHDTV Level 1. 4.1.1 About 4K The term 4K can mean either any resolution in the range 3840 to 4096 horizontal by 2048 to 2160 vertical. 4K became common as a term referring to the Digital Cinema resolution of 4096x2160. 4K resolutions used for film content are not necessarily 16:9 aspect ratio. 4.1.2 RED Digital Cinema RED introduced a variant of this resolution as a Digital Cinema recording resolution, 4520 x 2540 (sometimes also known as 5K ) 4.2 UHDTV Level 2 UHDTV Level 2 has twice the resolution (both horizontal and vertical) compared to UHDTV Level 1, so exactly 7680x4320. Baseband data rates at 59.94 Hz frame rate is 4x that of UHDTV Level 1, so approximately 48Gbit/s. UHDTV Level 2 is also known as Super Hi-vision. UHDTV Level 2 displays and cameras have been shown in prototype form, as well as links running over very high bandwidth connections. Ultra-High Definition TV Acquisition and Exchange 4
5. Connectivity to TVs An increasing number of consumer televisions are becoming available with UHDTV Level 1 resolution, with the early application being the internal upconversion of HD content to 4K. All major consumer electronics manufacturers have launched products. However, connectivity for true 4K operation is not necessarily so straight forward. 5.1 HDMI and DisplayPort HDMI 1.4 does allow 3840x2160 resolution to be delivered to the television, however it is limited to a maximum of 29.97 frames/second (therefore appropriate for the majority of today s movies). 50Hz and 59.94Hz UHDTV Level 1 exceeds the capability of HDMI 1.4. A new version of the HDMI specification is overdue, specifically with the aim of supporting frame rates higher than 29.97Hz for UHDTV Level 1. In the meantime, there are three options for consumer displays: Only support 24 frames/second UHDTV > Effective for movie content only Use 4x HDMI 1.4 > Each providing one quadrant (each quadrant 1920x1080 @ 59.94Hz) DisplayPort 1.2 > This already supports UHDTV Level 1 at 59.94Hz, but lacks the content encryption of HDCP, so does not have support from the content owners 6. The UHDTV Experience 6.1 Perceivable resolution Substantial research exists relating to perceivable resolution. Often this relates to optometrist sight tests, with reference to Snellen Charts. These relate to the angle of vision that a text character must occupy to be identifiable as that character. Perceivable resolution is, however, higher than this angle, as the ability to determine that two characters are different can be achieved at significantly smaller angles than that to identify exactly which character. This should not be a surprise, since latin-based characters consist of several black-white transitions within one character, so the perceivable resolution is in fact related to the ability to resolve those transitions. ITU has performed tests [2] to identify perceivable resolution, and the outcome is that detail is perceivable at approximately 0.018 degrees for humans, although it varies based on distance from the centre of view (with the highest acuity being in the central 2 degrees). This sets the viewing distance where a given resolution can be perceived, based on the screen size. 0.018 0 For a 1080-line HD screen, the resolution limit is reached when the viewing distance is approximately 1.6x the diagonal screen size. UHDTV Level 1 shows visible benefits over 1080- line HD at all distances closer than 1.6x diagonal, and reaches the limit of perceivable resolution at approximately 0.75x the diagonal size. Recent UHDTV Level 1 TVs are typically around 84 diagonal. At this size: > UHDTV Level 1 provides an improved viewer experience at distances closer than approximately 3.4m, progressively more visible until > Resolution reaches the limit at approximately 1.6m viewing distance. Figure 1: Viewing distance for UHDTV LEVEL 1 Ultra-High Definition TV Acquisition and Exchange 5
6.2 Field of view and the immersive experience The increased resolution and associated increase in screen size means that a UHDTV image occupies a significantly larger portion of the overall human field of view. Typically, we may expect that to be approximately 60 horizontal. An image that occupies the central 60 of the overall human vision range has a strongly dominant effect on the perception. This is because with a smaller field of view (e.g. 30 for typical HD), the surroundings of the TV set are obvious when watching the TV screen. With only HD resolution, increasing that field of view doesn t lead to the desired realism because the resolution is too low. So the combination of a large screen and UHDTV resolution provides a much more realistic immersive experience, because the distractions of surroundings are significantly reduced. The human visual system has a central viewing angle of approximately 90. This represents the region where, keeping the eye still, normal objects will be recognized and recalled. The overall viewing angle is greater than this, however the more peripheral vision has evolved to mainly only recognize large objects. The impact of the change in occupancy of the central viewing angle is shown in the comparison below: Central field of vision 90 0 Figure 2: Occupancy of central viewing angle for HD Central field of vision 90 0 Figure 3: Occupancy of central viewing angle for UHDTV As can be seen above the, change from the TV image from being the minority to the majority of the central viewing angle makes a dramatic difference to the perception, with the distraction of the surroundings being largely eliminated, leading to a significantly more immersive experience. Ultra-High Definition TV Acquisition and Exchange 6
6.3 Field of view and camera angle As screens become larger, the proportion of the overall human viewing angle occupied by the image is increased. One result of this is that the camera angle used to film the content should take into consideration the viewing angle. With UHDTV, it is possible to film with wider angle lenses, representing a closer match to the consumer s viewing angle for a large television. UHDTV s increased resolution allows this format change whilst keeping the resolvable detail. The end result is a more natural appearance to the images on the television. 6.4 Frame rates With increased screen size resulting in a larger viewed angle, for identically shot content, the angular velocity of motion for UHDTV is greater than for HD. As a result, motion is more perceivable as discrete steps in position. Frame rate increases can reduce this effect, however does increase both baseband and compressed video bit rates. It is expected that UHDTV will operate at 50.0 or 59.94Hz progressive frame rates for video content, and 24.0 or 23.976Hz progressive frame rates for native movie content (e.g. any UHDTV successor to BluRay). Colloquially, the video formats are often known as 4Kp50 or 4Kp59.94. These equate to 4x the 3G-SDI line rates, so approximately 12Gbit/s. 6.4.1 Frame rates above 60Hz There is some investigation into the potential for frame rates greater than 60Hz. This originates from the increased angular velocity of an object that travels across the whole width of the screen in the same time being higher with the increased field of view. However, as discussed in section 1.9 above, the production of UHDTV is likely to use wider camera angles, thus it is reasonable to expect that objects will take longer to travel the full width of the screen, thus partially compensating for the increase in screen size. As a result, it is not obvious that frame rates above 60Hz will be needed. 6.5 Colorimetry & bit depth Colorimetry, bit depth, pixel geometry, digital levels, for the baseband signals etc are defined in ITU-R BT.2020. See also 3.3 below for bit depth to the home. 6.6 Audio Existing audio standard, such as AAC, DTS, Dolby Digital and Dolby Digital Plus are equally applicable to UHDTV. At the time of writing, there is no clear drive to change the typical; audio configuration (i.e. from 5.1 or 7.1 surround sound), although there have been demonstrations of 22.2 surround sound, providing a full theater-class audio experience. Ultra-High Definition TV Acquisition and Exchange 7
7. Image tiling Tiling of UHDTV Level 1 into 4x 1920x1080 images provides a convenient way to use existing infrastructure to deliver UHDTV Level 1 content across a contribution network. An example of tiling a UHDTV Level 1 image is shown in Figure 1 below: 2160 Each quadrant is 1080p59.94/50 3840 Figure 4: Tiling of a UHDTV LEVEL 1 image into 4x HD quadrants Each quadrant is then a conventional 1920x1080 resolution image, at 50Hz or 59.94Hz that can be encoded and decoded with a 1080p59.94 capable HD encoder and decoder. An example of one quadrant is shown in Figure 5 below. 1080 1920 Figure 5: Example of one quadrant of a tiled image It is important that each quadrant remains exactly temporally aligned to avoid misalignment in the overall image when there is changing content. Ultra-High Definition TV Acquisition and Exchange 8
8. Ericsson UHDTV Acquisition and Exchange Systems Ericsson has developed a contribution system for UHDTV, covering 25, 29.97, 50 and 59.94Hz frame rates, allowing live content to be delivered across contribution links, using existing 1080p HD equipment. 8.1 Architecture The Ericsson UHDTV contribution system uses 4 separate encoders, each one compressing one quadrant of the UHDTV image, therefore each is a 1080p50 (or 59.94) H.264 encoder. At the receive end, 4 RX8200 IRDs are locked together using Ericsson Simulsync, providing the decoded output for the 4 quadrants, perfectly aligned. The encoders can be 4:2:2, 10-bit encoders, providing the optimal format for delivery of each quadrant. The encoders are multiplexed together into an MPTS (with ASI connections to the multiplexer), so that there is one TS carrying all the services. This allows adaptation to/from the carriage medium to be matched in latency. Each quadrant is 1080p59.94/50 Figure 6: Mapping UHDTV to 4 quadrants of 1080p Ultra-High Definition TV Acquisition and Exchange 9
Optional UHDTV record Or MPTS IP IP to ASI Quad 3G-SDI 2x AVP 2000 with 4x CE-xH42 encoding 1080p60 4.2.2 10b in synchronization 4x RX8200 with Simulsync phase lock Quad 3G-SDI or HDMI v1.4 Figure 7: Ericsson UHDTV contribution system Inputs to the RX8200s should be ASI or satellite in order to keep path delay mismatches to a minimum. If IP recovery is needed, then an external IP to ASI network adapter should be used. 8.2 Simulsync Simulsync encoders provide additional timing information along with the standard video data. RX8200s licensed for Simulsync use this additional information to ensure that the four quadrants are perfectly aligned. In order to ensure that the 4 outputs are timed exactly, the composite output of one of the decoders (typically chosen to be the top, left quadrant) is connected to the other 3 decoders, which align their outputs to that synchronization signal. 8.3 Monitor connections Two methods exist for connecting to full resolution UHDTV displays: > 4x 3G-SDI connections > 4x HDMI 1.4 connections In each case, the picture is split into 4 quadrants with the monitor aggregating the quadrants back into the overall UHDTV image. 8.4 Bit rates The baseband bit rate for UHDTV Level 1 is approximately 12 Gbit/s, as a simple scaling for resolution and frame rate. For compressed video, using H.264, it is reasonable to expect bit rates to be approximately 3-4x bit rates for the equivalent content in HD format. 8.5 1080p50/59.94 systems These are natively supported by existing Ericsson encoder and decoder products, some of which form the basis of the UHDTV system outlined above. Ultra-High Definition TV Acquisition and Exchange 10
9. UHDTV to the home The main goal for UHDTV is to enhance the viewing experience in the home. Since it operates at 4x the resolution and 2x the frame rate of conventional HD, it is clear that using existing technologies to deliver UHDTV to the home would require a large bandwidth. This may be appropriate for early adopters wishing to differentiate (using 4x 1080p50/59.94 H.264 compression), however in the long term a more efficient coding standard is required. 9.1 HEVC a new coding standard In January 2013, version 1 of High Efficiency Video Coding ( HEVC ) was approved as a final draft international standard, acquiring the designation H.265/ISO23009-1 as a result. HEVC offers bit rate efficiency improvements over H.264 of 35-50% depending on content, resolution, operating point, etc. In particular, the largest efficiency improvements are found at higher resolutions not surprisingly because HEVC introduced tools that were specifically targeted at increased picture resolutions (recognizing that 16x16 macroblocks were designed for content at around SD resolution). The use of HEVC means that it would be realistic to reach a point where a UHDTV service can be transmitted in the same bandwidth as an MPEG-2 HD service is today. 9.2 Decoding at the home More options will exist for decoding HEVC UHDTV in the home than was previously the case at the launch of HD H.264. STB silicon can reasonably be expected to be in place to allow services to be on air in time for the Rio 2016 Olympics, if not before. Next-generation games consoles are likely to have decode capability that would allow them to be used as alternative decode devices and it is likely that they will form a platform capable of decoding the next-generation disc formats for movie content. Both of these, of course, are dependent on the standardization and deployment of an updated HDMI specification that is capable of more than 30 frames per second. 9.3 Enhanced user experience Main 10 Profile Version 1 of HEVC includes both Main and Main 10 profiles. The only key difference between these profiles is the bit depth, Main 10 allows 10 bits per pixel, whereas Main only allows 8 bits. It is likely that Main (i.e. 8-bit) will be used on content targeted at tablets, mobile phones, etc, however one of the limiting factors for the consumer experience today with HD is the 8-bit resolution. This leads to contouring, most noticeable on content with slowly changing levels across a picture, particularly when associated with low levels of chrominance. The Main10 profile addresses this by including 10-bit resolution as an option for Main10 capable decoders, thus providing an improved viewing experience considered a key motivation for UHDTV television. As a side effect of the likely difference between different devices likely support for Main10, it is likely that a television connected to a Main10 capable STB or console will be the highest quality way to view content in UHDTV. Ultra-High Definition TV Acquisition and Exchange 11
9.4 Delivery mechanisms 9.4.1 DVB-Sx DVB-Sx is an update to DVB-S2 providing more optimized choices for modulation and coding values, with the aim of closer approaching Shannon s limit. This is expected to reach standardization in September 2013, providing up to 30% efficiency improvements over DVB-S2, depending on CNR and a number of other factors (although for DTH configurations, the saving is likely to be more in the range 10-15%). The timing of DVB-Sx makes it possible that UHDTV services via satellite could take advantage of increased transponder efficiency. 9.4.2 DVB-C2 DVB-C2 uses an OFDM-based modulation technique compared to more traditional QAM modulation used in existing cable modulation schemes. DVB-C2, like DVB-Sx, provides increased usable payload bitrate in the same bandwidth and again is a potential candidate for associating with rollout of UHDTV transmission over cable networks (although again like DVB-Sx, the modulation change is not essential to UHDTV services.) 9.4.3 Terrestrial delivery of UHDTV Terrestrial bandwidth for TV services is under pressure in many countries often because the spectrum is also suitable for mobile telephony services. Many governments expected and actioned a digital dividend ie recovering part of the spectrum from existing television services with the aim of re-selling that for other purposes. UHDTV did not feature in these plans and any additional expectation of a second digital dividend (ie where improvements in encoder performance releases more spectrum, then selling that for other purposes) may make UHDTV via terrestrial delivery difficult, even using HEVC (since legacy decoders will need to continue to receive the existing signal so UHDTV is a new demand on television spectrum). 9.4.4 Video On Demand UHDTV Video On Demand UHDTV potentially offers a first entry into UHDTV to the home. As previously described, a large quantity of movie content is already in existence at UHDTV resolutions, along with more studio television content being acquired and produced at UHDTV (even if not live). This provides a rich library of UHDTV content that is readily available and can offer the consumer with the right equipment a significantly improved viewing experience. Ultra-High Definition TV Acquisition and Exchange 12
CONCLUSION In summary, live production equipment and live contribution links are imminently available for UHDTV, with decode devices and televisions largely waiting for HDMI and silicon decoders. It is reasonable to expect that the Olympics in 2016 are likely to reach the home in live UHDTV. It is also very likely that live UHDTV contribution links will be used for the FIFA World Cup 2014 and may possibly reach the home in a few locations. 2013 Ericsson AB All rights reserved Ultra-High Definition TV Acquisition and Exchange - CONCLUSION 13
REFERENCES 1. ITU-R BT.1769 Levels 1 and 2 2. ITU-R BT.2246-1 reports 3840x2160 provides visible improvement compared to HD for distances in the range 1.5x to 3.0x picture height Ultra-High Definition TV Acquisition and Exchange - references 14