Live 8K Video Production

with 5G mmWave
 Table of Contents
1. Background: Business Needs.................................................................................. 1
2. 8K Video Use Cases, Protocols and Demands....................................................... 2
2.1. 8K Video Wireless Backhaul Use Case Description........................................2
2.2. Video Compression Formats............................................................................ 3

2.2.1. Visually Lossless Compression..............................................................3

2.2.2. Heavy Compression Protocols............................................................... 5

2.3. 8K UHD Video Radio Transmission KPIs.......................................................7

3. mmWave for 8K Video Transmission.................................................................... 8
3.1. End-to-end Transmission Schemes.................................................................. 8

3.1.1. Network Architecture Selection............................................................. 8

3.1.2. Core Network Schemes..........................................................................9

3.1.3. Radio Network Schemes...................................................................... 10

3.2. Uplink Transmission Technology...................................................................10

3.2.1. DSUUU Frame Structure..................................................................... 10

3.2.2. MU-MIMO...........................................................................................12

3.2.3. High-Power User Equipment (HPUE)................................................. 12

3.3. Protocol Adaption and Link Aggregation...................................................... 13

3.3.1. Transmission Protocol Background..................................................... 13

3.3.2. ISSP Protocol Introduction...................................................................13

3.3.3. ISSP Advantage Roundup.................................................................... 15

3.4. Research on Key Ecosystem Companies....................................................... 15

3.4.1. mmWave Mobile Communication Industrial Chain Maturity............. 16

3.4.2. UHD Encoder Maturity........................................................................16

3.4.3. 5G-based UHD Video Mobile Backhaul Maturity...............................18

4. Typical Transmission Cases.................................................................................. 18

 4.1 5G mmWave-based 8K Video Field End-to-End Transmission Demo...........18
4.2 Extended Discussion over 5G mmWave Network Applications.....................19

4.2.1. Various Mobile Live Broadcast and Coverage.....................................20

4.2.2. Field Production / In-Vehicle Production.............................................21

4.2.3. Remote Production...............................................................................22

4.2.4. Slow Livestream with Landscape Cameras......................................... 22

5. Conclusion...............................................................................................................24
References................................................................................................................... 26
Acknowledgements.....................................................................................................27
1. Background: Business Needs
“Information presented with video in ultra-high definition” has already become a development
mega-trend in the global information industry thanks to full 5G deployment. In terms of growth
and size, ultra-high definition (UHD) will account for 35% of video-on-demand (VOD) IP traffic
by end of 2022, and the UHD video industry will surpass RMB 4 trillion in China; both China
Central Television (CCTV) and Japan’s NHK achieved 8K event broadcast at the Tokyo Olympic
Games in 2021. The 2022 Spring Festival Gala and Beijing 2022 Winter Olympics have been
broadcasted via live 8K streams. It is hence inferred that UHD production and transmission will be
used to live broadcast the most compelling sport events in the future, including the Olympic
Games and FIFA World Cup. With the continuously rising uplink demand in the UHD video
service, China’s UHD industry has a fantastic opportunity for development, while also facing huge
challenges in uplink/downlink network speeds and capacity.

From the industrial chain perspective, development of the UHD industry will fuel product upgrade,
replacement and business process reengineering along the long UHD industrial chain, including
chips, video production equipment, storage equipment, network transmission equipment, display
panels and devices. In hardware, development of the UHD video industry and 8K in particular will
drive projectors, VR/AR and outdoor large-format screens to become a new type of gateways; in
application, development of converged technologies such as 8K+AI, 8K+VR/AR,
8K+holographic and 8K+interaction will propel further growth in 2C markets such as home and
culture & entertainment, and explosion in ToB fields like safety & security, healthcare, industrial
control and precision manufacturing. In China, as competitiveness and coverage of UHD devices
continue to increase, 60-inch or above large-format display panels will be fully UHD, the basic
device capability that catapults UHD experiences into homes already gets ready and the robust
foundation will strongly push industry and consumption upgrade, the future of explosive growth in
the UHD market is increasingly promising and China will become the world’s largest UHD
market.

In the Technical Requirements for 8K UHD TV Program Production and Broadcast (Interim
Provisions) [1] issued in January 2021, China Media Group (CMG) has stipulated:

1) 8K equipment in videos systems at 8K UHD studios or outside broadcast vans (OBVan)

directly use 8K baseband signals or convert 8K baseband signals into uncompressed IP signals
compliant with SMPTE-2110 standards for production in the systems. For signals sent to the
master control room, 4-stream 12Gbps video IP multicast signal flows are used for 8K UHD
signal transmission in accordance with the SMPTE-2110-20 standard. Light weight compression
protocols defined in the SMPTE-2110, such as Jpeg XS, are used for production domain
transmission.

2) In technical requirements for 8K UHD TV broadcast signals and interactive

video-on-demand (VOD) coding compression, AVS3/H.266/H.265 standards are used and video
bitrates range between 80-120Mbps.

1
As shown in the above data, although 8K uncompressed signals in broadcast standards can
guarantee ultra-low latency, the requirement for transmission rates using the baseband (12G SDI
x4) or SMPTE-2110-compliant IP streams can only be met with optical fiber transmission as the
current wireless transmission cannot carry such signals. The 80-120Mbps transmission rates
currently used in 8K broadcast and interactive VOD coding standards are basically close to the
upper limit of 5G uplink rates in today’s public network environment; however, the encode/decode
latency using the existing encode/decode modes for Internet video delivery is at the second level,
which cannot show the low-latency nature of 5G; furthermore, it is hard to meet the mastering
requirement in view of too low bitrates.

Therefore, the use of JPEG XS encode/decode technology for light weight compression of 8K
video into hundreds of Mega bps, close to 10bit recording rates for 8K broadcast video and
significantly higher than 8K bitrates for broadcast transmission, can retain more image details and
quality. Meanwhile, the relatively low encode/decode latency, in collaboration with the 5G
network, translates to the high-speed 5G+8K transmission capability, which is not available in
previous solutions.

2. 8K Video Use Cases, Protocols and Demands

2.1. 8K Video Wireless Backhaul Use Case Description

To transmit video streams at hundreds of Mbps and even multiple Gbps, high-bandwidth optical
fiber channels are usually used for production domain transmission of traditional 8K video. The
biggest limitation of this transmission technology is that camcorders and encoders need to be
placed at positions set in advance and cannot move flexibly. Therefore, this technology is not fit
for sport event recording with high mobility or outdoor programs moving in a large range.

5G technology using mmWave has broken the bottleneck of transmission bandwidths and can
offer Gbps-grade uplink bandwidths, making it possible for 8K high-bandwidth video backhaul on
uplinks in mobile communication. The below figure shows 8K UHD video transmission on 5G
links. 8K camcorders transmit data captured to encoders for encoding, and data encoded is sent as
uplink data via routers and CPEs. Signals are delivered to the International Broadcast Center (IBC)
through MEC and communication machine rooms.

Figure 1 5G links for UHD video broadcast and backhaul

2
2.2. Video Compression Formats

Video compression formats can commonly be divided into two major categories in accordance
with different transform domains, i.e., discrete cosine transform (DCT)-based and wavelet
transform (WT)-based. These two types of technologies have different characteristics. Today’s
relatively high compression ratio standards such as MPEG H.264, H.265 and H.266 and China’s
AVS series standards are all based on DCT/DST transform; light weight compression standards for
production domain such as TICO, JPEG-XS and JPEG-2000 commonly use wavelet transform.

IP data rates of 4-stream 12Gbps SDI signals, which 8K camcorders often output, are commonly
at hundreds of Mbps and even multiple Gbps after being converted with light weight compression
encoders for production domain. These signals can only be transmitted over optical fibers or
ethernet before 5G mmWave networks are commercialized, as they are far beyond the capabilities
of current 5G low-band (sub-6 GHz) and 4G commercial networks. 8K video signals are
commonly transmitted in heavy compression formats like H.265. To adapt to the capabilities of
4G and 5G commercial networks, relatively high compression ratios are commonly used to control
data rates at tens of Mbps.

Figure 2 Comparison of image quality primarily encoded/decoded based on wavelet

transform and DCT algorithms

2.2.1. Visually Lossless Compression

As network transmission bandwidths continuously improve, either wired private networks or

wireless networks including radio microwave technology and 5G mmWave technology continue to
be driven and implemented, and data transmission link bandwidths increase correspondingly. The
industry and users have higher requirements for live broadcast and video transmission quality,
especially in cultural, entertainment and sport events. However, lower is better for the video
transmission latency requirement.

3
While 8K video gains traction in the broadcast industry, UHD HDR (high dynamic range) and
BT.2020 broad-color-gamut video processing becomes a foundation for today’s high-end
professional video processing, and high quality of video also poses high requirements for
transmission link bandwidths.

In the wake of TICO format, JPEG-XS has become an optimal format for long-distance,
high-quality and ultra-low latency transmission of 8K or 4K UHD video currently in broadcast
and high-end video industries. The JPEG-XS (ISO / IEC 29170) standard specifies image
compression based on low-complexity sub-frame discrete wavelet transform (five-layer horizontal,
two-layer vertical), which combines with entropy coding and finally does bitrate allocation as
required. Image encode and decode can be separately controlled with 9~16 scanning lines, and do
not need image group-related prediction transform, fundamentally determining that end-to-end
latency of JPEG-XS encode and decode is minimum in currently-known algorithm formats. It is
workable to adapt to storage equipment or network transmission bandwidths by taking into
account transmission bandwidths and video quality requirements on the precondition that overall
video quality is maintained.

In broadcast-grade professional program production, typical JPEG-XS compression ratios set by

large television stations in China and other countries range between 10:1 and 16:1, JPEG-XS
video streams and relevant audio and supplementary data are transmitted separately with the
SMPTE ST2110 protocols, fully demonstrating flexibility of RTP data multicast for high-end
video and audio. Furthermore, master image quality of end-to-end latency as low as up to 0.3ms
can be realized to meet requirements for video program production and long-distance transmission
at studios.

Major characteristics of complete products are determined by JPEG-XS on the algorithm basis:

1) Video encoded and decoded in JPEG-XS has high reproduction, and in particular video
encoded and decoded after multiple iterations boasts stable quality. In typical video
sequence tests by B&M Modern Media Inc. among Japanese users, the PSNR value of
video second encoded and decoded in JPEG-XS is close to that first encoded and
decoded. This is also the greatest advantage of wavelet transform-based intraframe
coding format JPEG2000 over other DCT/DST transform-based interframe compression
and coding formats.

Table 1 Comparison of image reproduction iteratively encoded and decoded in JPEG-XS 4K

12G-SDI – actual PSNR test value
Item PSNR first encoded/decoded PSNR second encoded/decoded

JPEG-XS 8:1 52.25 52.19 (Almost Same)

2) Low latency of end-to-end video transmission

4
 3) The algorithmic theoretical value of JPEG-XS encode/decode latency is lower than
0.1ms, and considering Line Buffer on the decoding device, the end-to-end latency of
link products actually finished can be lower than 0.3ms. Actual latency in network
transmission systems will produce different values between 3ms~40ms, depending on
link conditions in private wired or wireless networks and reference phase precision in
ST2059 network clock and video transmission systems.

4) JPEG-XS encoded data at constant rates has stable requirements for transmission link
bandwidths

5) After compression bitrates of JPEG-XS video streams are set in accordance with image
transformation, there is no need to consider other relevant prediction information
quantization and storage, as bandwidths taken by data are constant. IP transmission link
bandwidths for UHD signal systems with SMPTE ST2110-22 protocol are reduced to
one of dozens of those in traditional broadcast signal transmission. This has greatly
decreased the cost of UHD systems and also brought high reliability and flexibility to IP
interaction and central management systems. As link bandwidths available in radio
transmission can increase to hundreds of Mbps and even 1 Gbps, high quality, ultra-low
latency and great stability of JPEG-XS encoded/decoded data deliver a fresh effect in
high-end video transmission applications.

6) 8K/4K UHD signals transmit video and audio data streams with the SMPTE ST2110
protocols, while gathering remote equipment or interaction information for sync
transmission. JPEG-XS streams and sync information or interaction information on
remote devices can be separately transmitted synchronously with ST2110. Such
innovation makes applications like virtual reality (VR) and UHD remote control more
lifelike and timelier.

7) JPEG-XS supports CPU-based encode and decode processing in FPGA hardware and
software, but it has relatively high requirements for interface speeds and processing
resources. The higher the video resolution is, the higher the computing frequency is, the
higher the processing bandwidths are and the more hardware resources taken are. With
the emergence of software products with support for JPEG-XS, international
organization VSF has issued more detailed standards and guidance specifications for
JPEG-XS streams transmitted through the ST2110-22 protocol with continuous updates,
which cover JPEG-XS-22 packaging formats and various data ranges, relatively unify
JPEG-XS-22 data formats and drive JPEG-XS encode/decode compatibility in software
based on the hardware and server foundation.

2.2.2. Heavy Compression Protocols

Heavy compression protocols are commonly designed for public IP links in considerably changing
transmission environments, featuring high compression ratios and superior link adaptability. When
deep-compression video coding modes are adopted in relatively complex changing environments,
highly-efficient video compression coding schemes should be used for video compression to the
5
maximum, which can effectively reduce video bitrates and lower requirements for network uplink
bandwidths. In live stream eco-system H.265 (HEVC) and AVS are wildly used for different
terminals, which are developed by MPEG and AVS alliance.

1. H265

H265, also known as High Efficiency Video Coding (HEVC), is the new generation of encoding
and decoding standard as a successor to H.264, which was jointly developed by ISO/IEC Moving
Pictures Experts Group and ITU- T Video Coding Experts Group (VCEG), similar to H.264. With
50% higher compression efficiency than H.264, it was mainly released in response to the current
trend of video applications continuously moving towards high definition, high framerates and high
compression. The first release of HEVC protocol standards was announced in July 2013 to more
than double the compression ratio on the H.264 standard complexity. As the new-generation video
coding standard, HEVC is still a prediction + transform hybrid coding framework, but makes
revolutionary uplifts in many aspects over H264. Its key technological highlights include:

1) Flexible coding structure

H.265 expands macroblocks from 16×16 pixel in H.264 to 64×64 to facilitate high-resolution
video compression. Meanwhile, it uses a more flexible coding structure to improve coding
efficiency, including Coding Unit, Predict Unit and Transform Unit.

2) Flexible block structure — Residual Quad-tree Transform (RQT)

Residual Quad-tree Transform (RQT) is an adaptive transform technology, an extension and

expansion of Adaptive Block-size Transform (ABT) in H.264/AVC. For interframe coding, it
allows for transforming the size of blocks to make adaptive adjustment in accordance with motion
compensation blocks; for intraframe coding, it allows for transforming the size of blocks to make
adaptive adjustment in accordance with characteristics of intraframe prediction residuals.
Compared to transforming small blocks, transforming big blocks can offer a better energy
concentration effect and retain more image details after quantization.

3) Sample Adaptive Offset (SAO)

Sample Adaptive Offset (SAO), located in the encode and decode loop and behind Deblock, sorts
reconstructed images and adds one offset to or subtracts one offset from the pixel value in each
category of images for the purpose of reducing distortion, which thus increases compression ratios
and reduce streams. The use of SAO can reduce streams by 2%~6% on average, and only
increases performance consumption of the encoder and decoder by approximately 2%.

4) Adaptive Loop Filter (ALF)

Adaptive Loop Filter (ALF), located in the encode and decode loop and behind Deblock and SAO,
is used to restore images reconstructed in order to minimize the mean-square error (MSE) between
reconstructed images and original images.

5) Parallel Design

6
As the chip architecture has nowadays evolved from single-core performance to multi-core
parallelization, HEVC/H265 was introduced with many optimization ideas for parallel computing
in order to adapt to the implementation of highly parallel chips, mainly including Tile, Slice and
WPP parallelization, which can fully leverage the multi-core parallel processing advantage of
modern chips.

2. H266 Coding

H266, also known as Versatile Video Coding (VVC), was born in the continuous evolution of
UHD technologies in the current UHD era with framerates gradually moving from 30fps to 60fps,
120fps and even 240fps and video resolution also continuously developing from previous 1080P
and 4K to today’s 8K UHD. VVC was jointly developed by MPEG and ITU, and many leading
companies around the world get involved including Qualcomm, HHI, Huawei, Samsung, Sony,
Intel, Nokia, Ericsson, Huawei, Tencent and Alibaba.

The requirement for VVC encoders is to include 30% (or more) lower bitrates than HEVC (H.265
/ MPEG-H) without compromising quality. VVC has the potential to support immersive content,
with various resolutions from 4K to 16K, and 360-degree panoramic video. VVC is similar to
HEVC in video compression, but makes improvements in partitioning, prediction and entropy
coding.

3. AVS3

Audio Video coding Standard (AVS) was developed by Audio Video coding Standard Workgroup
of China. Two generations of coding developed by the workgroup, including AVS1 and AVS2,
were issued as China’s national standards respectively in 2006 and 2016. Today, AVS standards
are widely applied in China’s radio and television fields, and expanding into Internet video and
monitoring fields. AVS3, the third-generation standard in AVS series, benchmarks VVC in key
design criteria, whose technological advantages are mainly reflected in two aspects: 1)
Optimization of traditional signal processing-related technologies, such as block partitioning and
motion prediction etc.; 2) Optimization and exploration using intelligent algorithms, such as
optimizing signal prediction and coding using neutral networks. The Audio Video coding Standard
Workgroup completed the draft benchmark document of AVS3 in March 2019, whose
performance surpasses AVS2 by 30% or so. The China Media Group Technical Requirements for
8K UHD TV Program Production and Broadcast (Interim Provisions) was issued on January 21,
2021, specifying the AVS3 standard is adopted for video delivery in encode and decode
technologies and AVS3/HEVC/VVC standards are adopted for interactive VOD video.

2.3. 8K UHD Video Radio Transmission KPIs

The Technical Requirements for 8K UHD TV Program Production and Broadcast (Interim
Provisions) issued by China Media Group (CMG) in January 2021 have stipulated:

1) 8K equipment in video systems at 8K UHD studios or on outside broadcast vans (OBVan)

directly use 8K baseband signals or convert 8K baseband signals into uncompressed IP

7
 signals compliant with the SMPTE-2110 standards for production in the systems to
realize high-bandwidth, low-latency processing and transmission. For signals sent to the
master control room, 4-stream 12Gbps video IP multicast signal flows are used for 8K
UHD signal transmission in accordance with the SMPTE-2110-20 standard.

2) In 8K video recording, XAVC-I Class 300/480/MXF (10bit) is used as the compression,

coding and packaging format with bitrates of 500Mbps X4 (Class 300/MXF) or
800Mbps X4 (Class 480/MXF).

3) In terms of technical requirements for 8K UHD television broadcast signal coding and
compression, the AVS3 standard is used, 8K UHD signals (7680×4320/50P/HDR) with
benchmark 10-bit profiles and 10.0.60 levels are used, with video bitrates not lower than
120Mbps.

4) In terms of technical requirements for 8K UHD interactive VOD, video is coded in

AVS3/H.266/H.265 with total bitrate greater than 80Mbps.

The above data shows that AVS3/H.266/H.265 coding standards with bitrates ranging between
80-120Mbps are used in broadcast domain, basically close to the upper limit of 5G uplink rates in
today’s public network environments. Screen-to-screen latency, as low as seconds and far lower
than that based on 4G and Internet, can offer viewing users with excellent experiences. In
production domain, the SMPTE-2110 light weight compression standard is used for compressed
8K signals compliant with the broadcast standard to retain more image details and quality and
guarantee tens of milliseconds of latency. Bitrates after compression at hundreds of Mbps are far
higher than 8K bitrates for transmission on the broadcast side. For huge challenges on the 5G
mobile communication system, mmWave technology is needed to enhance uplink capacity.

3. mmWave for 8K Video Transmission

3.1. End-to-end Transmission Schemes

To support the UHD video backhaul architecture shown in Figure 1, considerations are taken in
network architecture, core network and radio network scheme selection as follows.

3.1.1. Network Architecture Selection

In overall network architecture, NSA or SA networking may be selected, depending on support

capabilities and networking requirements of commercial devices. In the NSA networking scheme,
radio base stations need to support both LTE and 5G mmWave, corresponding radio network
administration is also required to support both 4G and 5G at the same time, and the core network
is selected to access virtual EPC. In the SA networking scheme, high low-band NR-DC or
mmWave standalone networking can be selected in radio base stations, corresponding radio
network administration only needs to support 5G and the core network is required to access 5GC.

8
 Figure 3 Network architecture

3.1.2. Core Network Schemes

Different configurations need to be chosen for the core network in accordance with different
network architectures. In the NSA network architecture, an operator can choose to access live 4G
EPC or build new EPC, and there is no difference between EPC and traditional 4G networks. In
the SA network architecture, an operator can choose to access live 5GC or build new private
network 5GC. The 5G core network supports a separation architecture for control/user panels to
realize distributed deployments.

Based on China Unicom’s latency requirement for live 4K/8K video broadcast at the Winter
Olympics, the operator can consider introducing a MEC scheme in the core network, where the
core network links with IBC video servers, as shown in the following system architecture figure.
ZTE’s 5G core network supports Multi-access Edge Computing (MEC) technology, UPFs are
deployed at the edge data center (Edge DC) closer to users in a bid to get close to users to the
greatest extent and access application networks such as Internet nearby. Traffic at the user panel
exports from the nearby UPF, enters application networks, and no longer exports from the core
network through the backbone network. This can effectively reduce network pressure and shock
from mmWave users’ high-bandwidth applications on our networks, while being able to offer
lower-latency services and further improve user experiences.

Figure 4 Core network scheme

9
3.1.3. Radio Network Schemes

Different configurations are chosen for radio networks in accordance with different network
architecture maturity. In the NSA network architecture, radio networks need to access 4G and 5G
radio base stations, so an operator can choose to overlap 5G mmWave base stations on live 4G
base stations or build all new 4G and 5G mmWave base stations. In the SA network architecture,
radio networks only need to access 5G base stations. If NR-DC is selected for networking, the
operator can choose to overlap 5G mmWave on live 5G low-band base stations or build all new
5G high-band and low-band stations. If SA FR2 is selected for standalone networking, an operator
should choose to build all high-band mmWave stations.

For live stream broadcast of top sport games, it is suggested to introduce a local traffic offload
scheme at the base station level, where the Node Engine in radio base stations links with IBC
video servers, as shown in the following network architecture figure, applicable to NSA or SA.
ZTE supports the built-in compute power engine Node Engine in base stations that moves
compute power down to base stations by offering compute, storage, network and accelerator
resources on the base station BBU chassis. In the containerized micro-service architecture, the
engine moves local traffic offload and radio network capabilities down to base stations, offers
PLMN and slice ID-based private network models, builds base station-grade industry edge
computing, allows data not to move out of the campus and realizes “direct access with one hop on
air interface”, while supplementing with Edge QoS refined guarantee that can meet low-latency
performance requirements needed for VR services.

Figure 5 Radio Network Scheme

3.2. Uplink Transmission Technology

3.2.1. DSUUU Frame Structure

To address the issue of insufficient mmWave uplink capacity, more slot resources may be
allocated for uplink service devices to increase uplink network capacity, such as using DSUUU

10
frame structure technology. In DSUUU frame structure configuration and DDDSU frame structure
configuration, there are same symbol configurations in any single D, S and U slot. For example,
the D/U slot respectively contains 14 full DL/UL symbols, S slot DL symbol: no special symbol:
UL symbol 10:2:2. The proportions of UL/DL symbols in DSUUU and DDDSU frame structures
are calculated as follows:

Table 2 Proportions of UL/DL symbols in DSUUU and DDDSU frame structures

DSUUU frame structure DDDSU frame structure

Total symbols (every five

DL GP UL DL GP UL
slots)

70 24 2 44 52 2 16

Symbol % 34.28% 2.86% 62.86% 74.29% 2.86% 22.85%

As shown in Table 1, the proportion of UL symbols in the DSUUU frame structure is 2.75 times of
those in the DDDSU frame structure, so more UL data and control signaling can be transmitted in
the unit time. Meanwhile, the UL peak rate of mmWave networks can be calculated with the
following formula:

UL data rate (Gbps) = the number of UL symbols (in one subframe cycle) * the number of spatial
multiplexing layers * the number of modulation orders * coding rate * PRB number *12/subframe
cycle (ms) * (1-UL control overhead)/10^6.

J
 ( j) BW ( j ), 
N PRB 12 
data rate (in Mbps)  10 6   v ( j)
 Layers  Qm  f
( j)
 Rmax   
 1  OH ( j ) 
j 1  Ts 

For a mmWave device with 2 spatial multiplexing layers and 64QAM modulation, the theoretical
UL peak rates of DSUUU and DDDSU frame structures can be calculated as shown in the
following table.

Table 3 Comparison of theoretical UL peak rates of mmWave devices in DSUUU and

DDDSU frame structures
DSUUU frame structure DDDSU frame structure

UL (200MHz) theoretical
1.12 0.81
peak rate of device (Gbps)

UL (400MHz) theoretical
2.25 1.62
peak rate of device (Gbps)

As shown in the above table, compared to the traditional DDDSU frame structure, the mmWave
DSUUU frame structure can significantly enhance UL capacities of mmWave networks and
devices, mmWave networks can theoretically accommodate multichannel light weight
11
compression 8K (600Mbps) video and effectively meet 8K video application needs in various
scenarios such as concert and stadium.

3.2.2. MU-MIMO

Multi-user MIMO can schedule devices that meet certain conditions to send resources on a same
time frequency, and distinguish signals from different users with the multi-user detection
algorithm on the receiving devices. Theoretical analysis shows, transmitting signals from different
devices with the same video resources can increase, even if not double, system throughput over
the approach that users harness different resources.

As shown in Figure 6, the single-user MIMO sending device produces one dedicated beam for a
single user and suppress side lobes in other directions as far as possible to avoid interfering other
users. The multi-user MIMO sending device produces multiple main lobes directed to serve users
at the same time. Each user’s main lobe signal is interfered by side lobes of other users’ side lobes
that time, leading to reduced SINR and a certain impact on single-user throughput. However, as
multiple users harness same time-frequency resources, the system throughput is still significantly
improved.

Figure 6 Comparison of MU-MIMO and SU-MIMO system throughput

3.2.3. High-Power User Equipment (HPUE)

User equipment at the cell edge or zones with poor coverage commonly have relatively low link
budgets, and cannot achieve ideal SINR even if full-power transmit is adopted. This time,
increasing the maximum transmit power of these equipment can effectively improve SINR of
receiving equipment and correspondingly increase the UL throughput.

3GPP has nowadays set a basic equipment transmit power of 200mw (23dBm), and user
equipment at some TDD bands can reach 400mw (26dbm). Now, 3GPP is developing
higher-power level standards for non-smartphone devices such as CPE.

12
3.3. Protocol Adaption and Link Aggregation

3.3.1. Transmission Protocol Background

IP-based video transmission protocols have nowadays become mainstream, including TS over
UDP, RTMP and SRT. Some of these protocols are based on UDP for guaranteed transmission
latency, and some are based on TCP/IP for guaranteed transmission accuracy, but they have a
common characteristic: designed for a single transmission link. In other worlds, only one
transmission path between two network nodes is chosen. Such transmission protocols designed for
Internet IP links is less fit for transmission channels with varying attenuation like mobile
communication because of more dramatic changes in mobile communication paths and even link
disruption in a short period when users enter a coverage shadow. In the meantime, mobile
operators often limit public network users’ traffic to guarantee universal access for users, which
also goes against transmitting a large amount of data. Therefore, the use of public networks in
mobile communication to transmit high-bandwidth video needs one type of multichannel
aggregation transmission protocols. The Inverse StatMux Stream Protocol (ISSP) designed by
TVU is an excellent multichannel aggregation protocol.

3.3.2. ISSP Protocol Introduction

The ISSP protocol, TVU’s transmission protocol using Inverse StatMux Plus (IS+) technology,
can use one single signal source for transmission by inverse multiplexing multiple channels and
then re-aggregating them on the receiving equipment. Single-channel signals are divided into
multiple channels and transmitted on multipath channels. In other words, one data stream can be
split into multiple small data packets that can be transmitted on all available network links. In
actual applications, the bandwidth of each connection is monitored and a corresponding amount of
data packets are transmitted in accordance with availability of every connection, and data packets
are re-combined after error correction on the receiving equipment and restored to original data.
Then, transmissions will be ordered and re-aggregated on the receiving equipment.

In the case of fluctuant 5G networks, aggregated data bandwidths will be more stable through
effective multipath data load balancing, and more suitable to high-stability live broadcast business
requirements in the radio and television industry, shown in Figure 7 and Figure 8 as follows.

13
 Figure 7 ISSP data transmission with multichannel aggregation

Figure 8 Stable bandwidths in aggregated links

Meanwhile, the ISSP protocol is an end-to-end transmission scheme added with more transmission
optimizations for audio and video services in 5G and Internet on top of basic multi-link
multiplexing:

 Application Layer Forward Error Correction (AL-FEC): The technology adds redundant error
correction packets to data on the transmitting side and corrects any error of lost packets with
redundant packets on the receiving side. Unlike traditional FEC, intelligent FEC in ISSP can
dynamically adjust the proportion of redundant packets in accordance with network
conditions to achieve more effective network utilization.

 Automatic Repeat-reQuest (ARQ): The technology can notify the transmitting side, when the
receiving side cannot restore data packets, of retransmitting lost data packets in the live
broadcast buffer time, thus ending with no impact on normal operation of live broadcast.

 Bandwidth prediction: Do real-time bandwidth prediction in accordance with the existing

network condition to facilitate notifying the front-end encoder of bitrate adjustment, while
14
 reasonably allocating bandwidths for every transmission link in coordination with
multi-network aggregation technology.

 Multi-network aggregation: The core technology in ISSP can effectively use all IP links for
balanced transmission of data loads. Data is split into small data packets fit for transmission
currently on different links on the transmitting side, and these small data packets are
consolidated again at the receiving side. In combination with the above-mentioned FEC and
ARQ technologies, it further improves live broadcast quality and stability.

 Dynamic data adjustment: The core technology in ISSP can dynamically split, adjust and
allocate data to different network IP links in accordance with real-time network bandwidths in
combination with bandwidth prediction. Similarly, data packet allocation includes the
combined use of the above-mentioned FEC and ARQ.

3.3.3. ISSP Advantage Roundup

According to systematic test and comparison analysis with other single-channel transmission
protocols, ISSP has a multi-network, multi-channel aggregation capability that can dynamically
allocate data, is more suitable for radio network transmission with relatively dramatic fluctuations
like 5G, and can effectively address a variety of problems arising from link fluctuations in 5G
mobile communication networks.

 Compared to traditional transmission protocols, the addition of intelligent FEC and intelligent
ARQ mechanisms ensures that every transmission link has the capability to resist packet loss,
jitter and latency in networks. It can effectively restore data and guarantee smooth live
broadcast in case of any sudden small-range fluctuation in 5G mmWave networks.

 Through the multi-link aggregation mechanism, it can effectively aggregate multiple 5G

mmWave network links and allocate data transmission loads in real time. In this way, data can
be retransmitted and normally restored in case of any sudden large-range 5G mmWave
fluctuation (such as 5G signals being blocked) or even signal loss (such as handing off
between base stations in motion or failure with some 5G CPE/module).

 The automatic bandwidth prediction and allocation mechanism can guarantee that the above
data error correction, retransmission and data aggregation can be automatically completed in
accordance with business needs. In this way, it can guarantee effective use of every 5G
network and automatic security of the live broadcast service.

3.4. Research on Key Ecosystem Companies

“Information presented with video in ultra-high definition” has already become a development
mega-trend in the global information industry thanks to full 5G enablement. In terms of growth
and size, ultra-high definition (UHD) will account for 35% of video-on-demand (VOD) IP traffic
by 2022, and the UHD video industry will surpass RMB 4 trillion in China; both China Central
Television (CCTV) and Japan’s NHK achieved 8K event broadcast at the Tokyo Olympic Games

15
in 2021. It is already made clear that the 2022 Spring Festival Gala and Beijing 2022 Winter
Olympics will be broadcast live in 8K UHD. It is hence inferred that UHD production and
transmission will be used to live broadcast the most compelling sports events in the future,
including the Olympic Games and FIFA World Cup. With the continuously rising uplink demand
in the UHD video service, China’s UHD industry has a fantastic opportunity for development,
while also facing huge challenges in uplink/downlink network speeds and capacity. In this chapter,
industry maturity is studied in several aspects such as the mmWave communication industry, UHD
encoder and 5G-based UHD video backhaul.

3.4.1. mmWave Mobile Communication Industrial Chain Maturity

The research report[3] released by GSA in May 2021 shows, 5G mmWave networks have moved
fast in the world.

 One hundred and eighty operators in forty-five countries/territories are investing in mmWave
(in the form of tests/trials, acquisition of licenses, planning deployments or engaging in
deployments).

 One hundred and thirty-two operators in 22 countries/territories have been assigned mmWave
spectrum (often on a regional basis) enabling operation of 5G networks.

 Twenty-eight operators in 16 countries/territories are known to be already deploying 5G

networks using mmWave spectrum.

 Nineteen countries/territories have announced formal (date-specified) plans or assigning

frequencies above 24 GHz between now and end-2022.

 One hundred and twelve announced 5G devices explicitly support one or more of the 5G
spectrum bands above 24 GHz. Seventy of those devices are understood to be commercially
available.

3.4.2. UHD Encoder Maturity

With the maturity of the UHD video industry, 8K/4K video shooting and production have come
standard in major events, starting with the 2018 FIFA World Cup. Meanwhile, pure 8K/4K
hardware products based on the JPEG-XS encode/decode format have iterated for several
generations. For example, there are already a host of actual application cases of B&M Modern
Media Inc.’s VFN28 encoder in large events internationally. UHD video captured at a competition
venue is compressed by the production domain encoder, edited and packaged on an outside
broadcast van (OBVan) nearby the venue and at the remote IBC Signal Center with the lowest
latency, and distributed or uploaded to the satellite network for transmission in H.265 and other
formats. The broadband optical fiber laid between the venue and the OBVan and the Signal Center
enables visually lossless, almost real-time 8K/4K video transmission.

Figure 9 shows an 8K signal transmission application for the sports Games. The distance between
8K camcorders on the sport venue and the off-site OBVan is 20km, and the distance between the
16
sport venue and the IBC exceeds 50km. After video is encoded/decoded and transmitted in
JPEG-XS format, the signal latency between the sport venue and the OBVan and IBC is controlled
within 40ms.

Figure 9 Application of JPEG-XS 8K/4K-based remote transmission products in major

events
China Central Television (CCTV) conducted link tests for 8K signal radio microwave
transmission in April 2021. B&M Modern Media Inc. remotely transmitted 8K video at
compression ratios of JPEG-XS 34:1 or 40:1 between different locations in view of radio
microwave transmission bandwidth limitations. The final latency of the whole radio transmission
system in actual tests was 50ms.

Figure 10 Actual application of JPEG-XS 8K-based radio microwave transmission

As now 5G mmWave commercial device-based transmission can provide approximately 900Mbps
uplink data bandwidths, transmission signals at about 600-700Mbps can be carried theoretically
on 5G mmWave networks when he JPEG-XS compression ratio is set at 60:1.

Heavy compression protocols like H.265 and AVS are widely adopted in production domain,
MPEG series protocols including H.264 and H.265 are commonly used on mobile devices, and the
AVS protocol is broadly applied to the traditional radio and television broadcast fields such as
television set and set top box (STB). Chinese encoder companies like Arcvideo already have the
capability to offer H.265 and AVS2/AVS3 encode/decode products with 8K support, which is
tested on the market in China.

17
3.4.3. 5G-based UHD Video Mobile Backhaul Maturity

As 5G networks are growing mature, 5G is also increasingly applied in radio, television and media
service applications. The most common 5G backpack livestream is used by national,
provincial/municipal and even district/county-level television stations for news, sports,
entertainment and livelihood channels to realize various real-time, portable 5G+4K/HD live
broadcast.

The multi-network aggregation 5G Rack Router can offer high-quality, high-stability, low-latency
and movable broadband networks, allow 5G network applications to further expand into the live
broadcast system of broadcast media, and be used in various fields including in-vehicle production,
remote production, cloud-based production and slow live broadcast of landscape.

Research statistics show, the above-mentioned 5G backpack and multi-network aggregation 5G

router are already applied to production systems at major television stations including China
Media Group (CMG), Xinhua News Agency, Shanghai Media Group (SMG), Zhejiang Television
Station, Fujian Television Station, Shandong Television Station, Hunan Television Station, Shanxi
Television Station and Guangdong Television Station.

4. Typical Transmission Cases

4.1 5G mmWave-based 8K Video Field End-to-End Transmission Demo

mmWave can offer Giga bps and even higher uplink transmission speeds by leveraging a
multitude of enhancement technologies such as high-bandwidth and MIMO. China Unicom, ZTE
Corporation and Qualcomm Technologies achieved the uplink peak rate of 920 Mbps in an
interoperability test in May 2021, making 8K video backhaul on mobile communication links
possible [2]. China Unicom, ZTE Corporation, Qualcomm Technologies Inc. and TVU Networks
achieved the world’s first DSUUU frame structure-based 5G mmWave 8K video backhaul
demonstration over the 26GHz (n258) band through lab testing in May 2021.

The demonstration was conducted by using 5G base stations from ZTE to offer a live network
environment and 5G Multi-Network Aggregation CPE form-factor test devices from TVU
Networks powered by the Qualcomm® Snapdragon™ X55 5G Modem-RF System and the
Qualcomm® QTM527 mmWave antenna module to offer 5G mmWave connections, enabling
robust backhaul for 8K video captured by 8K cameras in real time. In this demo, participants in
the industrial chain not only achieved validation of 5G mmWave UL-heavy peak rates of up to
920 Mbps in the live network on single devices, but also completed validation of the industry’s
first 8K video real-time routing and transmission between access network and core network on
multiple devices.

This test aggregates sub6 and mmWave spectrum, conducts 200MHz single-user uplink bandwidth,
2 spatial multiplexing layers, 64QAM modulation and 920Mbps uplink peak rate. Due to the use

18
of the DSUUU frame structure with uplink accounting for approximately 63%, the uplink
capability increases approximately 3 times, compared to the DDDSU frame structure. Meanwhile,
the air interface slot duration as low as 0.125ms is significantly reduced, compared to the latency
of Sub-6G air interface, so it can meet the requirement for 5G air interface latency less than 1ms
and more fit for a broad range of scenarios such as 5G industrial Internet of Things (IoT),
augmented reality/virtual reality (AR/VR), cloud gaming and event broadcast. 800MHz bandwidth
transmission is enabled with carrier aggregation technology in downlink, and the peak rate reaches
1.8Gbps. Both uplink and downlink capabilities can address high-bandwidth transmission for 8K
video.

Figure 11 5G mmWave field demo system

4.2 Extended Discussion over 5G mmWave Network Applications

The backhaul of today’s mainstream UHD and 4K UHD signals on 5G networks has become a
hotspot in HD and UHD video production since 5G commercialization. With growing
improvement in 8K coding and production services, 5G mmWave will be a service development
direction of 8K UHD+ 5G mmWave in the future after it is formally commercialized.

19
4.2.1. Various Mobile Live Broadcast and Coverage

In main service forms of commercial 5G applications, now the uplink mode of HD/4K camcorder
+ encoder + 5G aggregation router is used dominantly. The coding module and 5G aggregation
module will be integrated in the portable service, and the forms of 5G backpack and 5G phone are
used. However, due to the current limitations of coding technology and 5G module
industrialization, HD/4K+Sub-6 5G is dominant for 5G backpacks and 5G phones and can be used
as extended reference applications for future 8K+5G.

Typical case: CCTV World News PGC 4G/5G collection and transmission application

5G backpack and phone transmissions are widely applied on CCTV’s major channels, and in
particular play an important role during the pandemic.

20
4.2.2. Field Production / In-Vehicle Production

A high-speed, stable and low-latency production network is quickly built on the spot by leveraging
the 5G convergence network, which is offered for an in-vehicle/in-box system to receive and send
all kinds of IP video/audio signals and data signals, as well as for publishing and production on
various traditional media and convergence media.

Typical case: Zhejiang Daily Tianmu News 5G OBVan – participates in the “Along China’s
Expressways” event

21
4.2.3. Remote Production

A high-speed, stable and low-latency network is quickly built on the spot by leveraging the 5G
network to transmit all kinds of field signals to the back-end production system, reduce front-end
production pressure and costs and thus better use back-end production resources.

Typical case: Shichadao, 15-hour large New Year event in 8 cities (2020~2021) — full
livestream with 5G cameras

5G was used to transmit the 15-hour slow livestream service in 8 cities to the back end for
production.

4.2.4. Slow Livestream with Landscape Cameras

The slow livestream service was initially intended to cater to users who want to enjoy splendid
scenery and charisma of different locations at home. The slow livestream service culminated in
“Livestreaming the Construction of Huoshenshan and Leishenshan Emergency Hospitals” earlier
in 2020, with approximately 200 million views of hospital construction livestreaming on all types
of devices.

22
 Slow livestream transmitting landscape images captured with UHD cameras or VR360
panoramic camcorders on 5G networks.

Typical case: 5G+4K slow livestream of Shougang Park landscape captured at CIFTIS 2021

Transmitting landscape images captured with 4K cameras at Shougang Park to CIFTIS IBC News
Center

23
5. Conclusion
With the 5G commercial deployment under way, transmitting 4K and HD video on mobile links is
applied widely. In particular, video backhaul technology like TVU has already become a typical
case that combines video production and 5G mobile communication, which is widely applied to
various professional livestream fields such as mobile live broadcast and coverage, field production,
remote production and various professional livestream. 5G mmWave is expected to improve 4K
video quality in backhaul links and unlock advanced use cases such as 8K UHD backhaul and
remote production as it can offer Giga bps uplink bandwidths.

This white paper first analyzes growth potential of 8K UHD video services, then proposes
end-to-end enabling solutions for 8K video applications based on mmWave networks, and
describes transmission case expectations for current 4K UHD video applications. To drive UHD
livestream, remote production and VR/AR applications, dominated by 8K video applications, to
offer industries with more economic and entertainment value, further efforts need to be taken in
the following aspects.

Application deployment: Now 8K video livestream and broadcast applications via optical fibers
are relatively mature, and are deployed at major competition, evening party and concert platforms
in China and other countries. However, because locations are relatively fixed, 8K video
production is commonly undertaken on fixed stations. It is expected in the future to use 5G and 5G
mmWave technologies, leverage the innovation environment, actively explore 8K innovative
technologies and solutions, strive for 8K “wireless” production and stably push ahead with mature,
convenient application and deployment of 8K video end-to-end solutions.

Network: In a typical 8K video transmission case described in this white paper, the 5G mmWave
network configured with the DSUUU frame structure supports the maximum total bandwidth of
couples of 200MHz channels. The next step consideration is to further boost the UL data rate with
larger bandwidth, higher modulation and other uplink enhancement. Massive application of 8K
video will create a capacity shock on core network, transmission network and radio network, so
operators need to further increase network capacity and reduce end-to-end latency and jitters.
They need to consider private network scenario deployment for 8K video applications, and drive
standalone deployment of private mmWave networks and research, development and trial of
enhanced MIMO technology.

Technology: We need to drive flexible conversion and technology maturation of DDDSU/DSUUU

frame structures; improve UL MIMO technology maturity and application; increase mobility
enhancement of key 8K video applications on CPEs, mobile phones, camcorders and VR/AR
devices, while developing high-power user equipment (HPUE).

Protocols: H.265 and AVS3 heavy compression encode/decode protocols are used in existing 8K
video end-to-end application solutions, which have relatively long coding latency. In the next step,
we may further drive application of light weight compression protocols, reduce “screen-to-screen”
latency and propel 5G application in the UHD video production field.

24
Now the 8K UHD video industry is still in the early days of product research and development
and in the small-scale trial phase, so there are no clear business demands generated at scale. 8K
application examples have been dominated by demonstration cases in the recent two years, and it
will take time for 8K to touch home users. There are a lot of issues such as the content production
cost for 8K UHD applications remaining high, a small number of 8K home devices available and
narrow channels for monetization in the downstream. This further leads to a weak passion of
producers and communication agencies in investment. From the perspective of general consensus
reached by participants in the industrial chain, business opportunities arising from short-path
livestream of large events will become mainstream in the early stage, such as 8K live theater, 8K
live sports and 8K live variety. Moreover, the top sports games like Olympic Games and Asian
Games will also further accelerate rolling out 8K UHD video applications. Nevertheless, overall,
all players still need to explore relatively reliable business models and profit closed loops on the
basis of relatively high costs along the industrial chain in the short term.

It can be judged from the history of the audiovisual industry that 4K and 8K UHD will co-exist in
the 5G era for quite a while as these two video modes have their own merits, rather than undergo a
simple iteration relationship characterized by replacement, and gradually generate rich application
scenarios in their own fields. But in comparison, 8K has higher requirements for viewing on
large-format screens, presence and immersion, and boasts unrivalled advantages in pursuing
ultimate watching and listening experiences and seeking more precise image data. In addition,
now VR content production is relatively mature, and VR application scenarios include VR gaming,
VR shopping, VR education, VR social, VR livestream, VR tour, VR healthcare, VR engineering
and VR house; it is foreseeable that 8K VR content will first be commercially available at scale,
4K or 1080P content will be shown on VR devices via FOV in the early days, true 8K VR can be
realized only after 8K content capture and presentation devices are both 8K-capable.

Similarly, in the ToB industry sectors, application scenarios such as 8K video surveillance, 8K
remote healthcare, 8K videoconference and 8K machine vision inspection will be constantly
mentioned. By offering ultra-high precision display, multi-perspective/paranoiac presentation and
combining with professional industry-specific technology, the UHD technology can greatly
improve the capability level in all industry sectors and thus drive commercialization at scale in the
ToB field.

25
References
[1] China Media Group (CMG), Technical Requirements for 8K UHD TV Program Production
and Broadcast (Interim Provisions), January 21, 2021.
[2] Milestone Achieved In 5G mmWave Testing Plan of the IMT-2020 (5G) Promotion Group by
Completing 5G mmWave DSUUU frame structure-based 8K Video Backhaul Demonstration on
May 21, 2021, https://www.qualcomm.cn/news/releases-2021-05-21-0
[3] mmWave Bands 24.25 GHz: May 2021,
https://gsacom.com/paper/mmwave-bands-24-25-ghz-may-2021-executive-summary/
[4] Yiqing Cao: "Live 8K production using 5G mmWave",
https://www.5g-mag.com/post/follow-up-workshop-media-production-over-5g-npn-deep-dive-into
-protocols

26
Acknowledgements
Strong support and contribution received from experts at the following universities, research
institutes and companies when the report was written were greatly appreciated.

4K Garden Lu Yu
ABS - Academy of Broadcasting Yu Zhang
Science
Arcvideo –Arcvideo Tech Jiaxing Chen
Xiaoming Shi
Kedi Zhu
B&M –B&M Modern Media Inc. Violet Wang
CU – China Unicom Zhonghao Zhang
Shuai Gao
Qualcomm Yiqing Cao
Yan Li
Richard Li
Zhenyu Li
Wentao Zhang
Zhimin Du
Xiangkai Guan
Robin Zhu
Bo Chen
TVU Networks – Shanghai Tongwei Zhi Wu
Jing Wang
Pokemon Li
Xin Shen
ZTE Corporation Ruijing Hao
Yaodong Liu
Junjie Shao
Qin Yin
Jindong Wang

27