Field Test of 5G New Radio (NR) UL-MIMO and UL-256QAM for HD Live-Streaming
Extracted source text · arXiv 2511.19868
%% bare_conf.tex
%% V1.4b
%% 2015/08/26
%% by Michael Shell
%% See:
%% http://www.michaelshell.org/
%% for current contact information.
%%
%% This is a skeleton file demonstrating the use of IEEEtran.cls
%% (requires IEEEtran.cls version 1.8b or later) with an IEEE
%% conference paper.
%%
%% Support sites:
%% http://www.michaelshell.org/tex/ieeetran/
%% http://www.ctan.org/pkg/ieeetran
%% and
%% http://www.ieee.org/
%%*************************************************************************
%% Legal Notice:
%% This code is offered as-is without any warranty either expressed or
%% implied; without even the implied warranty of MERCHANTABILITY or
%% FITNESS FOR A PARTICULAR PURPOSE!
%% User assumes all risk.
%% In no event shall the IEEE or any contributor to this code be liable for
%% any damages or losses, including, but not limited to, incidental,
%% consequential, or any other damages, resulting from the use or misuse
%% of any information contained here.
%%
%% All comments are the opinions of their respective authors and are not
%% necessarily endorsed by the IEEE.
%%
%% This work is distributed under the LaTeX Project Public License (LPPL)
%% ( http://www.latex-project.org/ ) version 1.3, and may be freely used,
%% distributed and modified. A copy of the LPPL, version 1.3, is included
%% in the base LaTeX documentation of all distributions of LaTeX released
%% 2003/12/01 or later.
%% Retain all contribution notices and credits.
%% ** Modified files should be clearly indicated as such, including **
%% ** renaming them and changing author support contact information. **
%%*************************************************************************
% *** Authors should verify (and, if needed, correct) their LaTeX system ***
% *** with the testflow diagnostic prior to trusting their LaTeX platform ***
% *** with production work. The IEEE's font choices and paper sizes can ***
% *** trigger bugs that do not appear when using other class files. *** ***
% The testflow support page is at:
% http://www.michaelshell.org/tex/testflow/
\documentclass[conference,a4paper]{IEEEtran}
% Some Computer Society conferences also require the compsoc mode option,
% but others use the standard conference format.
%
% If IEEEtran.cls has not been installed into the LaTeX system files,
% manually specify the path to it like:
% \documentclass[conference]{../sty/IEEEtran}
% Some very useful LaTeX packages include:
% (uncomment the ones you want to load)
% *** MISC UTILITY PACKAGES ***
%
%\usepackage{ifpdf}
% Heiko Oberdiek's ifpdf.sty is very useful if you need conditional
% compilation based on whether the output is pdf or dvi.
% usage:
% \ifpdf
% % pdf code
% \else
% % dvi code
% \fi
% The latest version of ifpdf.sty can be obtained from:
% http://www.ctan.org/pkg/ifpdf
% Also, note that IEEEtran.cls V1.7 and later provides a builtin
% \ifCLASSINFOpdf conditional that works the same way.
% When switching from latex to pdflatex and vice-versa, the compiler may
% have to be run twice to clear warning/error messages.
% *** CITATION PACKAGES ***
%
%\usepackage{cite}
% cite.sty was written by Donald Arseneau
% V1.6 and later of IEEEtran pre-defines the format of the cite.sty package
% \cite{} output to follow that of the IEEE. Loading the cite package will
% result in citation numbers being automatically sorted and properly
% "compressed/ranged". e.g., [1], [9], [2], [7], [5], [6] without using
% cite.sty will become [1], [2], [5]--[7], [9] using cite.sty. cite.sty's
% \cite will automatically add leading space, if needed. Use cite.sty's
% noadjust option (cite.sty V3.8 and later) if you want to turn this off
% such as if a citation ever needs to be enclosed in parenthesis.
% cite.sty is already installed on most LaTeX systems. Be sure and use
% version 5.0 (2009-03-20) and later if using hyperref.sty.
% The latest version can be obtained at:
% http://www.ctan.org/pkg/cite
% The documentation is contained in the cite.sty file itself.
% *** GRAPHICS RELATED PACKAGES ***
%
\ifCLASSINFOpdf
% \usepackage[pdftex]{graphicx}
% declare the path(s) where your graphic files are
% \graphicspath{{../pdf/}{../jpeg/}}
% and their extensions so you won't have to specify these with
% every instance of \includegraphics
% \DeclareGraphicsExtensions{.pdf,.jpeg,.png}
\else
% or other class option (dvipsone, dvipdf, if not using dvips). graphicx
% will default to the driver specified in the system graphics.cfg if no
% driver is specified.
% \usepackage[dvips]{graphicx}
% declare the path(s) where your graphic files are
% \graphicspath{{../eps/}}
% and their extensions so you won't have to specify these with
% every instance of \includegraphics
% \DeclareGraphicsExtensions{.eps}
\fi
% graphicx was written by David Carlisle and Sebastian Rahtz. It is
% required if you want graphics, photos, etc. graphicx.sty is already
% installed on most LaTeX systems. The latest version and documentation
% can be obtained at:
% http://www.ctan.org/pkg/graphicx
% Another good source of documentation is "Using Imported Graphics in
% LaTeX2e" by Keith Reckdahl which can be found at:
% http://www.ctan.org/pkg/epslatex
%
% latex, and pdflatex in dvi mode, support graphics in encapsulated
% postscript (.eps) format. pdflatex in pdf mode supports graphics
% in .pdf, .jpeg, .png and .mps (metapost) formats. Users should ensure
% that all non-photo figures use a vector format (.eps, .pdf, .mps) and
% not a bitmapped formats (.jpeg, .png). The IEEE frowns on bitmapped formats
% which can result in "jaggedy"/blurry rendering of lines and letters as
% well as large increases in file sizes.
%
% You can find documentation about the pdfTeX application at:
% http://www.tug.org/applications/pdftex
% *** MATH PACKAGES ***
%
%\usepackage{amsmath}
% A popular package from the American Mathematical Society that provides
% many useful and powerful commands for dealing with mathematics.
%
% Note that the amsmath package sets \interdisplaylinepenalty to 10000
% thus preventing page breaks from occurring within multiline equations. Use:
%\interdisplaylinepenalty=2500
% after loading amsmath to restore such page breaks as IEEEtran.cls normally
% does. amsmath.sty is already installed on most LaTeX systems. The latest
% version and documentation can be obtained at:
% http://www.ctan.org/pkg/amsmath
% *** SPECIALIZED LIST PACKAGES ***
%
%\usepackage{algorithmic}
% algorithmic.sty was written by Peter Williams and Rogerio Brito.
% This package provides an algorithmic environment fo describing algorithms.
% You can use the algorithmic environment in-text or within a figure
% environment to provide for a floating algorithm. Do NOT use the algorithm
% floating environment provided by algorithm.sty (by the same authors) or
% algorithm2e.sty (by Christophe Fiorio) as the IEEE does not use dedicated
% algorithm float types and packages that provide these will not provide
% correct IEEE style captions. The latest version and documentation of
% algorithmic.sty can be obtained at:
% http://www.ctan.org/pkg/algorithms
% Also of interest may be the (relatively newer and more customizable)
% algorithmicx.sty package by Szasz Janos:
% http://www.ctan.org/pkg/algorithmicx
% *** ALIGNMENT PACKAGES ***
%
%\usepackage{array}
% Frank Mittelbach's and David Carlisle's array.sty patches and improves
% the standard LaTeX2e array and tabular environments to provide better
% appearance and additional user controls. As the default LaTeX2e table
% generation code is lacking to the point of almost being broken with
% respect to the quality of the end results, all users are strongly
% advised to use an enhanced (at the very least that provided by array.sty)
% set of table tools. array.sty is already installed on most systems. The
% latest version and documentation can be obtained at:
% http://www.ctan.org/pkg/array
% IEEEtran contains the IEEEeqnarray family of commands that can be used to
% generate multiline equations as well as matrices, tables, etc., of high
% quality.
% *** SUBFIGURE PACKAGES ***
%\ifCLASSOPTIONcompsoc
% \usepackage[caption=false,font=normalsize,labelfont=sf,textfont=sf]{subfig}
%\else
% \usepackage[caption=false,font=footnotesize]{subfig}
%\fi
% subFiguresty, written by Steven Douglas Cochran, is the modern replacement
% for subfigure.sty, the latter of which is no longer maintained and is
% incompatible with some LaTeX packages including fixltx2e. However,
% subFiguresty requires and automatically loads Axel Sommerfeldt's caption.sty
% which will override IEEEtran.cls' handling of captions and this will result
% in non-IEEE style figure/table captions. To prevent this problem, be sure
% and invoke subFiguresty's "caption=false" package option (available since
% subFiguresty version 1.3, 2005/06/28) as this is will preserve IEEEtran.cls
% handling of captions.
% Note that the Computer Society format requires a larger sans serif font
% than the serif footnote size font used in traditional IEEE formatting
% and thus the need to invoke different subFiguresty package options depending
% on whether compsoc mode has been enabled.
%
% The latest version and documentation of subFiguresty can be obtained at:
% http://www.ctan.org/pkg/subfig
% *** FLOAT PACKAGES ***
%
%\usepackage{fixltx2e}
% fixltx2e, the successor to the earlier fix2col.sty, was written by
% Frank Mittelbach and David Carlisle. This package corrects a few problems
% in the LaTeX2e kernel, the most notable of which is that in current
% LaTeX2e releases, the ordering of single and double column floats is not
% guaranteed to be preserved. Thus, an unpatched LaTeX2e can allow a
% single column figure to be placed prior to an earlier double column
% figure.
% Be aware that LaTeX2e kernels dated 2015 and later have fixltx2e.sty's
% corrections already built into the system in which case a warning will
% be issued if an attempt is made to load fixltx2e.sty as it is no longer
% needed.
% The latest version and documentation can be found at:
% http://www.ctan.org/pkg/fixltx2e
%\usepackage{stfloats}
% stfloats.sty was written by Sigitas Tolusis. This package gives LaTeX2e
% the ability to do double column floats at the bottom of the page as well
% as the top. (e.g., "\begin{figure*}[!b]" is not normally possible in
% LaTeX2e). It also provides a command:
%\fnbelowfloat
% to enable the placement of footnotes below bottom floats (the standard
% LaTeX2e kernel puts them above bottom floats). This is an invasive package
% which rewrites many portions of the LaTeX2e float routines. It may not work
% with other packages that modify the LaTeX2e float routines. The latest
% version and documentation can be obtained at:
% http://www.ctan.org/pkg/stfloats
% Do not use the stfloats baselinefloat ability as the IEEE does not allow
% \baselineskip to stretch. Authors submitting work to the IEEE should note
% that the IEEE rarely uses double column equations and that authors should try
% to avoid such use. Do not be tempted to use the cuted.sty or midfloat.sty
% packages (also by Sigitas Tolusis) as the IEEE does not format its papers in
% such ways.
% Do not attempt to use stfloats with fixltx2e as they are incompatible.
% Instead, use Morten Hogholm'a dblfloatfix which combines the features
% of both fixltx2e and stfloats:
%
% \usepackage{dblfloatfix}
% The latest version can be found at:
% http://www.ctan.org/pkg/dblfloatfix
% *** PDF, URL AND HYPERLINK PACKAGES ***
%
%\usepackage{url}
% url.sty was written by Donald Arseneau. It provides better support for
% handling and breaking URLs. url.sty is already installed on most LaTeX
% systems. The latest version and documentation can be obtained at:
% http://www.ctan.org/pkg/url
% Basically, \url{my_url_here}.
% *** Do not adjust lengths that control margins, column widths, etc. ***
% *** Do not use packages that alter fonts (such as pslatex). ***
% There should be no need to do such things with IEEEtran.cls V1.6 and later.
% (Unless specifically asked to do so by the journal or conference you plan
% to submit to, of course. )
% correct bad hyphenation here
\hyphenation{op-tical net-works semi-conduc-tor}
\usepackage[left=1.57cm,right=1.57cm,top=0.95cm,bottom=2.54cm]{geometry}
\usepackage{listings}
\usepackage{setspace}
\usepackage{graphicx}
\usepackage{tabularx,booktabs}
\usepackage{dingbat}
\usepackage{diagbox}
\usepackage{multirow}
\usepackage[hyphens]{url}
\usepackage[hidelinks,breaklinks]{hyperref}
% \usepackage{hyperref}
\usepackage{xcolor}
\usepackage{amsfonts, amsmath, amsthm, amssymb}
\usepackage{subcaption}
\hypersetup{breaklinks=true}
\urlstyle{same}
\usepackage{tikz}
\usepackage{textcomp}
\usepackage{lipsum}
\usepackage{svg}
\IEEEoverridecommandlockouts
\newcommand\copyrighttext{%
\footnotesize \textcopyright 2026 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.}
\newcommand\copyrightnotice{%
\begin{tikzpicture}[remember picture,overlay]
\node[anchor=south,yshift=10pt] at (current page.south) {\fbox{\parbox{\dimexpr\textwidth-\fboxsep-\fboxrule\relax}{\copyrighttext}}};
\end{tikzpicture}%
}
\lstset{frame=tb,
language=Java,
aboveskip=3mm,
belowskip=3mm,
showstringspaces=false,
columns=flexible,
basicstyle={\tiny\ttfamily},
numbers=none,
numberstyle=\tiny\color{gray},
keywordstyle=\color{blue},
commentstyle=\color{dkgreen},
stringstyle=\color{mauve},
breaklines=true,
breakatwhitespace=true,
tabsize=3
}
\usepackage{graphicx}
\graphicspath{{figures/}}
\begin{document}
%
% paper title
% Titles are generally capitalized except for words such as a, an, and, as,
% at, but, by, for, in, nor, of, on, or, the, to and up, which are usually
% not capitalized unless they are the first or last word of the title.
% Linebreaks \\ can be used within to get better formatting as desired.
% Do not put math or special symbols in the title.
\title{Field Test of 5G New Radio (NR) UL-MIMO and UL-256QAM for HD Live-Streaming}
% author names and affiliations
% use a multiple column layout for up to three different
% affiliations
\author{
\IEEEauthorblockN{Kasidis Arunruangsirilert}
\IEEEauthorblockA{Department of Computer Science and Communications Engineering, Waseda University, Tokyo, Japan
\\kasidis@katto.comm.waseda.ac.jp}
}
% conference papers do not typically use \thanks and this command
% is locked out in conference mode. If really needed, such as for
% the acknowledgment of grants, issue a \IEEEoverridecommandlockouts
% after \documentclass
% for over three affiliations, or if they all won't fit within the width
% of the page, use this alternative format:
%
%\author{\IEEEauthorblockN{Michael Shell\IEEEauthorrefmark{1},
%Homer Simpson\IEEEauthorrefmark{2},
%James Kirk\IEEEauthorrefmark{3},
%Montgomery Scott\IEEEauthorrefmark{3} and
%Eldon Tyrell\IEEEauthorrefmark{4}}
%\IEEEauthorblockA{\IEEEauthorrefmark{1}School of Electrical and Computer Engineering\\
%Georgia Institute of Technology,
%Atlanta, Georgia 30332--0250\\ Email: see http://www.michaelshell.org/contact.html}
%\IEEEauthorblockA{\IEEEauthorrefmark{2}Twentieth Century Fox, Springfield, USA\\
%Email: homer@thesimpsons.com}
%\IEEEauthorblockA{\IEEEauthorrefmark{3}Starfleet Academy, San Francisco, California 96678-2391\\
%Telephone: (800) 555--1212, Fax: (888) 555--1212}
%\IEEEauthorblockA{\IEEEauthorrefmark{4}Tyrell Inc., 123 Replicant Street, Los Angeles, California 90210--4321}}
% use for special paper notices
%\IEEEspecialpapernotice{(Invited Paper)}
% make the title area
\maketitle
\copyrightnotice
% As a general rule, do not put math, special symbols or citations
% in the abstract
\setstretch{0.95}
\begin{abstract}
% Over the years, the amount of User Generated Content (UGC) has skyrocketed, particularly HD live video streaming, which poses a growing demand for high-performance uplink capabilities in mobile networks. To cope with the demand, the 5G New Radio (NR) standard introduced many key uplink enhancement features such as Uplink Multi-Input Multi-Output (UL-MIMO) and Uplink 256QAM to improve uplink throughput and spectral efficiency. However, while the benefits of these features on raw data rates have been studied, their practical impact on real-time applications such as live-streaming remains unknown. In this paper, the performance of UL-MIMO and UL-256QAM are investigated from the perspctive of HD live-streaming over a commercial 5G network via RTMP protocol. Through modem firmware modification of commercial UEs, the study provides a comparative analysis by enabling and disabling these features on the same user equipment to ensure a fair assessment. The performance is evaluated in terms of dropped video frames and connection stability. Furthermore, the study analyzes 5G RF parameters to understand the spectral efficiency impact of each feature, offering crucial insights for the planning and optimization of next-generation radio access networks.
The exponential growth of User-Generated Content (UGC), especially High-Definition (HD) live video streaming, places a significant demand on the uplink capabilities of mobile networks. To address this, the 5G New Radio (NR) standard introduced key uplink enhancements, including Uplink Multi-Input Multi-Output (UL-MIMO) and Uplink 256QAM, to improve throughput and spectral efficiency. However, while the benefits of these features for raw data rates are well-documented, their practical impact on real-time applications like live-streaming is not yet well understood. This paper investigates the performance of UL-MIMO and UL-256QAM for HD live-streaming over a commercial 5G network using the Real-Time Messaging Protocol (RTMP). To ensure a fair assessment, we conduct a comparative analysis by modifying the modem firmware of commercial User Equipment (UE), allowing these features to be selectively enabled and disabled on the same device. Performance is evaluated based on key metrics, including dropped video frames and connection stability. Furthermore, this study analyzes 5G Radio Frequency (RF) parameters to quantify the spectral efficiency impact, specifically examining metrics derived from the Channel State Information (CSI) framework, including Reference Signal Received Power (CSI-RSRP), Reference Signal Received Quality (CSI-RSRQ), and Signal-to-Interference-plus-Noise Ratio (CSI-SINR).
\end{abstract}
\begin{IEEEkeywords}
5G New Radio (NR), Uplink 256QAM, Uplink MIMO, Live-Streaming, Radio Access Network
\end{IEEEkeywords}
% no keywords
\setstretch{0.947}
% For peer review papers, you can put extra information on the cover
% page as needed:
% \ifCLASSOPTIONpeerreview
% \begin{center} \bfseries EDICS Category: 3-BBND \end{center}
% \fi
%
% For peerreview papers, this IEEEtran command inserts a page break and
% creates the second title. It will be ignored for other modes.
\IEEEpeerreviewmaketitle
\section{Introduction}
% Since the introduction of 5G New Radio (NR), many new emerging use cases has been introduced including massive IoT, Smart City, cloud gaming, Autonomous Vehicle, Virtual Reality (VR), and Augmented Reality (AR) \cite{10373900, 10574348}. One of the practical key use case that often observed today is the transmission of User Generated Content (UGC), a trend that has rapidly grown since the early 2010s \cite{4801529}. Moreover, the increasing popularity of social media platforms, where user are encourage to create and share the multimedia content, especially High-Definition live video \cite{ericsson_2016}, created a new challenges for mobile network, which usually prioritize downlink traffic. This is because conventionally, during the 4G Long-Term Evolution (LTE) era, the traffic on mobile network around the world is usually Downlink-heavy, where most mobile users consume the existing contents on the internet, instead of producing them. Furthermore, the rise of Generative AI (GenAI) with Multimodal AI capable of understanding multimedia contents, including images, audio and video, resulted in more UGC being transmit over the mobile network as user attach these multimedia payload with their prompt \cite{ericsson_2024}. This changing trend resulted in mobile traffic becoming more balanced as we progressed toward the 6G era.
The emergence of 5G New Radio (NR) has enabled new applications such as massive IoT, Smart City, cloud gaming, Autonomous Vehicles, and Virtual/Augmented Reality (VR/AR) \cite{10373900, 10574348}. A key use case is transmitting of User Generated Content (UGC), a trend that has grown since the early 2010s \cite{4801529}. The increasing popularity of social media platforms, which encourage users to create and share multimedia content like HD live video \cite{ericsson_2016, 10951237}, has presented new challenges for mobile networks traditionally optimized for downlink traffic. During the 4G Long-Term Evolution (LTE) era, mobile network traffic was predominantly downlink-heavy, with most users consuming existing internet content rather than producing them. The recent rise of multimodal Generative AI (GenAI), capable of processing images, audio, and video, has further increased uplink transmissions as users attach multimedia payloads to their prompts \cite{ericsson_2024}. This evolving usage pattern is causing mobile traffic to become more balanced as we progress toward the 6G era.
% To cope with the growing demand for high-performance uplink, 3GPP has introduced several new features as a part of 5G Standard \cite{3GPP_38-101-1} to improve the uplink performance and spectral efficiency. The two key features that has been widely adopted around the world is Uplink MIMO (UL-MIMO) and Uplink 256QAM (UL-256QAM). Unlike the conventional 4G LTE User Equipment (UE), which has multiple antenna for downlink data reception, but only one uplink antenna to transmit the user data, 5G UE may posess more than one uplink antenna, allowing the adoption of Uplink MIMO, where UE with two antenna may use spatial multiplexing technique to double their maximum uplink throughput in ideal signal conditions, whereas two antenna may also be operated together in transmit diversity mode in the weaker signal condition to improve the reliability of the uplink connection. Previous study has found that in real-world commercial network, UL-MIMO-enabled UE delivers an average of 19.8\% higher performance on moving train in urban area, whereas up to 33.5\% improvement in throughput was observed in more favorable signal conditions on C-Band spectrum \cite{10118777}. However, it should be noted that dual transmission antenna implementation is not mandatory according to the 3GPP standard. Therefore, many UE manufacturers choose to omit the optional diversity transmssion antenna for smaller design and cost-saving.
To address the growing demand for high-performance uplink, 3GPP has introduced several new features in the 5G Standard \cite{3GPP_38-101-1}, including Uplink MIMO (UL-MIMO) and Uplink 256QAM (UL-256QAM). Unlike conventional 4G LTE User Equipment (UE), which typically has only one uplink antenna, 5G UEs may possess more than one. This allows for the adoption of UL-MIMO, where a UE with two antennas can use spatial multiplexing to double its maximum uplink throughput in ideal signal conditions or operate in transmit diversity mode to enhance connection reliability in weaker signal conditions. A previous study found that in a real-world commercial network, UL-MIMO-enabled UEs deliver an average of 19.8\% higher throughput on moving trains in urban areas, with up to a 33.5\% improvement observed in more favorable signal conditions on the C-Band spectrum \cite{10118777}. However, since dual transmission antenna implementation is not mandatory under the 3GPP standard, many UE manufacturers omit this optional feature for design simplicity and cost-saving.
% On the other hand, Uplink 256QAM is simpler to implement due no additional antenna needed to be implemented on the UE side, but likely to be less effective. Uplink 256QAM works by assigning Table 5.1.3.1-2 or MCS Index Table 2 (\textit{qam256}), which is standardized under \cite{3GPP_38-214}, for uplink transmission over the Table 5.1.3.1-1 or MCS Index Table 1 (\textit{qam64}) Table, which can improve the spectral efficiency by the maximum of 25\% as each symbol may now carry 8 bit instead of 6 bit permits by the 64QAM modulation. However, the effectiveness of Uplink 256QAM strongly depend on the signal condition as 256QAM is much less resilient to noise and interference, requiring Signal-to-Noise Ratio of more than 30 dB to maintain stable communication, which is difficult to achieve in real-world commercial network environment. Real-world study conducted by Kasidis et al. \cite{10570635} found that this is indeed the case, as enabling UL-256QAM only resulted in 20\% of resource block being modulated using 256QAM on passive antenna network, which yield an uplink throughput of only 8.22\%.
Alternatively, Uplink 256QAM is simpler to implement as it requires no additional antennas on the UE side. It works by assigning Modulation Coding Scheme (MCS) Index Table 2 (\textit{qam256}) for uplink transmission instead of MCS Index Table 1 (\textit{qam64}), as standardized in \cite{3GPP_38-214}. This can improve spectral efficiency by a maximum of 33.3\% by allowing each symbol to carry 8 bits instead of the 6 bits permitted by 64QAM. The effectiveness of Uplink 256QAM, however, is highly dependent on the signal condition, requiring a Signal-to-Noise Ratio of over 30 dB to maintain stable communication, which is difficult to achieve in a real-world commercial network environment. In our previous work \cite{10570635}, we confirmed this, finding that enabling UL-256QAM resulted in only 20\% of resource blocks being modulated using 256QAM on a passive antenna network, yielding an uplink throughput gain of just 8.22\%. \looseness=-1
% While both study highlighted the advantages of UL-MIMO and UL-256QAM in terms of uplink throughput and spectral efficiency enhancement, their impacts on real-world applications such as real-time live-streaming remain unknown. Since, it is expected that these 5G feature will be the key enabler for Ultra High-Definition (UHD) live-streaming in the future, understanding their impact is crucial for planning the Next-Generation Radio Access Network (NG-RAN) to support the emerging use cases. In this paper, the performance impact of the UL-MIMO and UL-256QAM on live-streaming of HD video is investigated. By modifying the modem firmware, both of the feature can be turned on and off on the same User Equipment, allowing a fair comparison. The video were then transmitted to an RTMP server via commercial 5G UE and network, in which the performance is being evaluated in terms of dropped video frame and number of disconnection. Furthermore, 5G RF parameters were analyzed to understand the spectral efficiency impact of each features, which is crucial for RAN capacity planning.
While both studies highlight the advantages of UL-MIMO and UL-256QAM in enhancing uplink throughput and spectral efficiency, their impacts on real-time applications such as live-streaming remain largely unknown. Since these 5G features are expected to be key enablers for Ultra High-Definition (UHD) live-streaming, understanding their impact is crucial for planning Next-Generation Radio Access Networks (NG-RANs) to support emerging use cases. In this paper, the performance impact of UL-MIMO and UL-256QAM on the live-streaming of HD video is investigated. By modifying the modem firmware, both features can be enabled and disabled on the same User Equipment, ensuring a fair comparison. Videos were transmitted to an RTMP server via a commercial 5G UE and network, where performance was evaluated in terms of dropped video frames and the number of disconnections. Furthermore, 5G RF parameters were analyzed to understand the spectral efficiency impact of each feature, including key metrics from the Channel State Information (CSI) framework like CSI-RSRP, CSI-RSRQ, and CSI-SINR, which is crucial for RAN capacity planning.
\section{Field Test Setup}
\subsection{Uplink MIMO Evaluation Environment}
% For the evaluation of Uplink MIMO, a UE with two transmission antenna is required. Therefore, KDDI au Speed Wi-Fi HOME 5G L13 (based on ZTE MC888 Ultra) equipped with Qualcomm Snapdragon X65 5G RF Modem was used \cite{qualcomm_x65}. The modem diagnostic port was exposed via the internal service mode, which allows the modem firmware to be modified to limit the capability to Single-Tx as needed. To collect RF parameters, the device was connected to a laptop running \textit{AirScreen}, a professional network drive test tool. Before each trial, the \textit{UECapabilityInformation} packets were verify to ensure that the UE capability is as expected. The UL-MIMO field test was conducted on a Japanese Mobile Network Operator (MNO), SoftBank, with a commercial plan and SIM card. Since the UE used for this experiment only support UL-MIMO on Time Division Duplexing (TDD) frequency band, the JR Yamanote line, which is the loop line located in the central area of Tokyo, was chosen for the field test, as the line has good coverage of SoftBank 5G Band n77 (3.4 GHz). Due to SoftBank having multiple carriers of n77, including 3.4 GHz (40 MHz Bandwidth - NR-ARFCN 627360), 3.5 GHz (40 MHz Bandwidth - NR-ARFCN 638112), and 3.9 GHz (100 MHz Bandwidth - NR-ARFCN 660768), each with different path loss, coverage and bandwidth, the result maybe affected if the UE randomly choose the carrier to connect. Therefore, the mitigate this problem, the modem firmware was modified to lock the UE on 5G Standalone Mode with the target NR-ARFCN of 627360, the lowest frequency n77 carrier, in which SoftBank has the most base stations. The collected RF data were then analyzed using \textit{AirScreen}, the same software that was used to collect the log.
For the evaluation of Uplink MIMO, a User Equipment (UE) with two transmission antennas is required. Therefore, a KDDI au Speed Wi-Fi HOME 5G L13, based on the ZTE MC888 Ultra, equipped with the Qualcomm Snapdragon X65 5G RF Modem, was used \cite{qualcomm_X65}. The modem's diagnostic port was exposed via an internal service mode, which permitted firmware modifications to limit the capability to a single transmission (Single-Tx) as needed. To collect RF parameters, the device was connected to a laptop running \textit{AirScreen}, a professional network drive test tool. Before each trial, \textit{UECapabilityInformation} packets were verified to ensure the UE's capability was configured as expected. The field test for UL-MIMO was conducted on the network of a Japanese Mobile Network Operator (MNO), SoftBank, using a commercial plan and SIM card.
Since the UE used for this experiment supports UL-MIMO only on Time Division Duplexing (TDD) frequency bands, the JR Yamanote line in central Tokyo was chosen for the test, as it has good coverage of SoftBank's 5G Band n77 (3.4 GHz). To address potential result variability from the UE randomly connecting to SoftBank's multiple n77 carriers—each with different path loss, coverage, and bandwidth, such as 3.4 GHz (40 MHz Bandwidth - NR-ARFCN 627360), 3.5 GHz (40 MHz Bandwidth - NR-ARFCN 638112), and 3.9 GHz (100 MHz Bandwidth - NR-ARFCN 660768)—the modem firmware was modified. This change locked the UE to 5G Standalone Mode on the target NR-ARFCN of 627360, which is SoftBank's lowest frequency n77 carrier with the most base stations. Each trial involved looping the Yamanote Line once on a moving train with an average speed of 28 km/h according to GPS data, taking about 70 minutes to complete. The RF conditions are summarized in Table \ref{tab-RFCondition} for reference purposes. It should be noted that UL-256QAM was also enabled for these test scenarios. The collected RF data were then analyzed using AirScreen, the same software that was utilized to collect the logs. \looseness=-1
\subsection{Uplink 256QAM Evaluation Environment}
To evaluate Uplink 256QAM, the Japanese Samsung Galaxy S22 (SC-52C), which is equipped with the same Qualcomm Snapdragon X65 5G RF Modem, served as the User Equipment (UE). Since this smartphone runs the Android operating system, \textit{Network Signal Guru (NSG)}, the Android version of \textit{AirScreen}, was installed on the phone to collect RF parameters. The same commercial plan and SIM card were used for this test. By accessing Samsung's service mode menu, the UL-256QAM capability could be turned off as needed for comparison. This study focused on two of SoftBank's Frequency Division Duplexing (FDD) spectrums, n3 (1800 MHz - 15 MHz Bandwidth) and n28 (700 MHz - 10 MHz Bandwidth), because MNOs globally are refarming FDD bands from LTE service to 5G NR, which enables UL-256QAM on such spectrum. It is important to note that the n28 band is configured purely as 5G NR, whereas the n3 band employs Dynamic Spectrum Sharing (DSS), permitting the simultaneous use of the spectrum for both LTE and 5G NR services. To ensure each band could be tested independently, the UE was locked to a specific frequency using the forcing function within the NSG application.
The use of these FDD bands facilitates an investigation into not only the impact of UL-256QAM but also the advantages of migrating from LTE to 5G NR, since both Radio Access Technologies use Single-carrier FDMA (SC-FDMA) for the uplink. The experiment was conducted on Tokyu Den-en-toshi Line, excluding the underground section (Futako-tamagawa to Minami-machida Grandberry Park) due to densely deployed SoftBank FDD 5G NR base stations along the line. Each trial took approximately 30 minutes to complete on a moving train with an average speed of 51.5 km/h based on GPS data. Similar to the UL-MIMO evaluation, RF conditions can be found in Table \ref{tab-RFCondition}. Since the smartphone has only a Single Tx antenna, UL-MIMO is not supported and hence not enabled in these test scenarios. Following the trials, the RF logs from \textit{Network Signal Guru (NSG)} were analyzed using \textit{AirScreen} on a PC.
\subsection{Live-Streaming Configuration}
% To understand the behavior of live-streaming applications that utilize RTMP protocol, an RTMP server was hosted on a server located in Tokyo, Japan using SRS Media Server version 6.0.155. Then, Open Broadcaster Software (OBS) Studio Version 30.1.1 was installed on a laptop with Intel Wi-Fi 6 AX200 160MHz Adapter. The laptop was connected to the UE via Wi-Fi. On the KDDI au Speed Wi-Fi HOME 5G L13, the device is already a 5G Wi-Fi router, so the laptop can be connected directly. As for the Samsung Galaxy S22 Ultra, the built-in tethering function was used to share the cellular connection to the laptop via Wi-Fi. The OBS was set to encode the live video using AMD Advanced Media Framework (AMF) hardware encoder with rate control mode set to Constant Bit Rate (CBR) mode. Keyframe interval and maximum consecutive B-frame were set to two seconds and two frames, respectively. The preliminary experiment was done to determine an appropriate target bitrate for each evaluation on different frequency band. From the result, the target bitrate of 5000 kbps and 3000 kbps were used for UL-MIMO and UL-256QAM evaluation, respectively. The video were then transmit to the RTMP server using RTMP protocol, then the OBS logs was examined to determine number of disconnection and dropped frame.
To understand the behavior of live-streaming applications utilizing the RTMP protocol, an RTMP server was hosted on a server in Tokyo, Japan, using SRS Media Server (v6.0.155). For the client, Open Broadcaster Software (OBS) Studio (v30.1.1) was installed on a laptop equipped with an Intel Wi-Fi 6 AX200 adapter. The laptop was connected to the network via Wi-Fi, either directly to a KDDI au Speed Wi-Fi HOME 5G L13 router or through the built-in tethering function of a Samsung Galaxy S22 Ultra.
In OBS, the live video was encoded using the H.264/AVC codec via the AMD Advanced Media Framework (AMF) hardware encoder, with the rate control mode set to Constant Bit Rate (CBR) to provide a stable uplink load for isolating network performance. Furthermore, the ``\textit{Dynamically change bitrate to manage congestion (Beta)}" feature was not enabled, as it is unsupported by the AMD AMF hardware encoder used in this study. Following live encoding recommendations from YouTube and Twitch \cite{google, twitch}, the keyframe interval was set to two seconds, and the maximum number of consecutive B-frames was set to two. Based on preliminary experiments, target bitrates of 5000 kbps and 3000 kbps were selected for the UL-MIMO and UL-256QAM evaluations, respectively. The video was transmitted to the server using the RTMP protocol, and performance was evaluated by examining the OBS logs for the number of disconnections and dropped frames. \looseness=-1
\begin{table}[!tbp]
\setstretch{0.75}
\caption{Summary of RF Conditions by Test Case}
\vspace{-1.7mm}
\centering
\label{tab-RFCondition}
\resizebox{8.5cm}{!}{\begin{tabular}{@{}cccccccccc@{}}
\toprule
Freq&DL&UL&SSB&BW&MCS&Tx&CSI&CSI&CSI \\
Band&Freq&Freq&ARFCN&(MHz)&Table&Conf.&RSRP&RSRQ&SINR\\
&(MHz)&(MHz)&&&&&(dBm)&(dB)&(dB)\\\midrule
\multicolumn{10}{c}{Uplink MIMO Evaluation}\\\midrule
\multirow{2}{*}{n77}&\multirow{2}{*}{3420}&\multirow{2}{*}{3420}&\multirow{2}{*}{627360}&\multirow{2}{*}{40}&\multirow{2}{*}{\textit{qam256}}&1-Tx&-92.28&-15.43&8.76\\
&&&&&&2-Tx&-91.71&-15.53&8.58\\\midrule
\multicolumn{10}{c}{Uplink 256QAM Evaluation}\\\midrule
\multirow{2}{*}{n28}&\multirow{2}{*}{798}&\multirow{2}{*}{743}&\multirow{2}{*}{159630}&\multirow{2}{*}{10}&\textit{qam64}&\multirow{2}{*}{1-Tx}&-86.79&-15.33&6.34\\
&&&&&\textit{qam256}&&-89.60&-15.33&6.22\\
\midrule
\multirow{2}{*}{n3}&\multirow{2}{*}{1852.5}&\multirow{2}{*}{1757.5}&\multirow{2}{*}{369870}&\multirow{2}{*}{15}&\textit{qam64}&\multirow{2}{*}{1-Tx}&-90.19&-15.46&8.96\\
&&&&&\textit{qam256}&&-90.79&-15.60&8.78\\
\bottomrule
\end{tabular}}
\vspace{-5mm}
\end{table}
\section{Result and Analysis}
\subsection{Uplink MIMO}
\begin{figure}[t!]
\vspace{2mm}
\centering\includesvg[width=0.95\linewidth,inkscapelatex=false]{OBS-ULMIMO.svg}\\[2mm]
\centering\includesvg[width=0.25\linewidth,inkscapelatex=false]{ULMIMO_Legend.svg}
\caption{\textbf{UL-MIMO}: Live Encoding Performance}
\label{fig:OBS-ULMIMO}
\vspace{-7mm}
\end{figure}
\begin{figure}[t!]
\centering\includesvg[width=0.95\linewidth,inkscapelatex=false]{RF-ULMIMO.svg}\\[2mm]
\centering\includesvg[width=0.25\linewidth,inkscapelatex=false]{ULMIMO_Legend.svg}
\caption{\textbf{UL-MIMO}: RF Parameters}
\vspace{-5mm}
\label{fig:RF-ULMIMO}
\end{figure}
\begin{figure}[t!]
\centering\includesvg[width=0.90\linewidth,inkscapelatex=false]{QAM-ULMIMO.svg}
\caption{\textbf{UL-MIMO}: Uplink Modulation Utilization (\%)}
\vspace{-6.5mm}
\label{fig:QAM-ULMIMO}
\end{figure}
% Figure \ref{fig:OBS-ULMIMO} shows dropped frame percentage and number of time that OBS was disconnected from the RTMP server. It was found the Uplink MIMO does slightly reduce the dropped frame percentage by 0.1\%. However, this came a cost of a significant increase in amount of disconnection, as it was observed that OBS re-established the connection to RTMP server eight times when UL-MIMO was enabled, compared to the Single Tx case. This is because standard UE has the maximum transmission power limit of 23 dBm, which is corresponding to the Power Class 3. By enabling UL-MIMO, this power budget has to be shared among two of the transmission antenna, resulting in a maximum transmission power of 20 dBm per MIMO chain. While UL-MIMO-enabled UE can compensate this reduction of per-MIMO power by using transmission diversity, when the channel quality is very poor, stronger power per antenna is more effective in overcoming the propagation loss, resulting in a more stable connectivity. However, 3GPP has standardized High Power User Equipment (HPUE), which permits certified UE with Power Class 2 (PC2) and Power Class 1.5 (PC1.5), to use a higher transmission power of 26 dBm and 29 dBm, respectively, restoring the maximum power per antenna to be equivalent to the Single Tx case, mitigating the negative effect of splitting power across multiple antenna chains \cite{3GPP_38-101-1}. However, the effect of HPUE on real-time application is out of scope for this work and will be evaluated in future work.
Figure \ref{fig:OBS-ULMIMO} presents the dropped frame percentage and the number of disconnections from the RTMP server. The results indicate that UL-MIMO reduces the dropped frame percentage by a marginal 0.1\%. However, this slight improvement is accompanied by a significant increase in connection instability. When UL-MIMO was enabled, the OBS client re-established its connection to the RTMP server eight times, a substantial increase compared to five times for the single-transmit configuration. This instability arises because a standard UE, operating under Power Class 3, is limited to a maximum transmission power of 23 dBm. With UL-MIMO enabled, this power budget is divided between two transmission antennas, effectively limiting each MIMO chain to a maximum of 20 dBm. Although a UL-MIMO-capable UE can utilize transmission diversity to compensate for this reduced power per chain, a single, more powerful transmission is more effective at overcoming propagation loss in poor channel conditions, thus providing a more stable connection. To address this limitation, 3GPP has standardized High Power User Equipment (HPUE), which allows certified UEs in Power Class 2 (PC2) and Power Class 1.5 (PC1.5) to transmit at higher powers of 26 dBm and 29 dBm, respectively. This restores the power per antenna to levels comparable to the single-transmit case, thereby mitigating the negative effects of power division across multiple antenna chains \cite{3GPP_38-101-1}. The performance impact of HPUE on real-time applications is beyond the scope of this study and will be investigated in future work. \looseness=-1
% Figure \ref{fig:RF-ULMIMO} shows the RF parameter obtained from \textit{AirScreen}. Because the Constant Bit Rate (CBR) rate control mode was used, there is no significant difference in uplink throughput between two trial, as illustrated by the MAC layer throughput and Physical (PHY) layer throughput result. Furthermore, the amount of Block Error Rate increased slightly by 0.49\% after enabling Uplink MIMO, which means that the UE has to re-transmit more of the user data, potentially leading to higher latency and less stable connection. However, Uplink-MIMO-enabled UE was able to utilize spatial multiplexing tranmission mode, allowing two independent stream of data to be transmitted to the base station in better condition with better channel quality. MIMO Rank results shows the effectiveness of MIMO operation by providing insight on number of stream of data being transmitted. On the other hand, resource block per second shows the number of resource block consumed by the UE to transmit Physical Uplink Shared Channel (PUSCH) or user data. It was found that Uplink MIMO UE has an average MIMO rank of 1.41, which is a huge improvement over the fixed MIMO rank of 1.00 for Single Tx UE. This resulted in a reduction of uplink resource block consumed by the UE by 17.05\%, allowing a much more efficient use of the frequency spectrum. On the other hand, the spectral efficiency average number of bit represented per symbol, where high utilization of high modulation order will resulted in higher value. It was found that there was only slight reduction in number of bit being modulated per symbol after enabling Uplink MIMO. Figure \ref{fig:QAM-ULMIMO} shows the percentage of each type of modulation used for PUSCH modulation, which provide more insight into spectral efficiency. It was found that while enabling Uplink MIMO resulted in resulted in 7.3\% lower 256QAM modulation, the availability of transmission diversity did compensate in condition with medium channel quality, by allowing about 3\% of the resource block to be modulated using 16QAM (4 bit/symbol) instead of QPSK (2 bit/symbol), compensating the spectral efficiency loss from reduction of 256QAM utilization in better channel quality scenario.
Figure \ref{fig:RF-ULMIMO} illustrates the RF parameters obtained from \textit{AirScreen}. Due to the use of a Constant Bit Rate (CBR) rate control mode, there was no significant difference in the MAC layer or Physical (PHY) layer throughput between the two trials. However, enabling UL-MIMO resulted in a slight increase in the Block Error Rate (BLER) by 0.49\%, indicating a higher re-transmission probability, which could potentially increase latency. Despite this, the primary advantage of UL-MIMO is its ability to utilize spatial multiplexing, allowing for the transmission of two independent data streams. The effectiveness of this operation is quantified by the MIMO Rank, which indicates the number of data streams being transmitted. The results show that the UL-MIMO configuration achieved an average MIMO Rank of 1.41, a significant improvement over the fixed rank of 1.00 in the single-transmit configuration. This enhancement led to a 17.05\% reduction in the number of uplink resource blocks consumed for the Physical Uplink Shared Channel (PUSCH), signifying a substantial improvement in spectral efficiency.
Further analysis of spectral efficiency is presented in Figure \ref{fig:QAM-ULMIMO}, which details the utilization percentage of each modulation scheme for PUSCH. While enabling UL-MIMO led to a 7.3\% decrease in the use of 256QAM, the availability of transmission diversity compensated for this in medium-quality channel conditions. Approximately 3\% of resource blocks were modulated using 16QAM (4 bits/symbol) instead of QPSK (2 bits/symbol), which partially offset the spectral efficiency loss from the reduced 256QAM utilization. In summary, although enabling UL-MIMO may lead to increased disconnections in challenging RF environments, it improves spectral efficiency by 17.05\%. This allows for more effective use of the frequency spectrum, which can enhance cell capacity in congested scenarios like mass gathering and live concerts. The adoption of HPUE on UL-MIMO-capable UEs is expected to mitigate connection stability issues and improve the overall Quality-of-Experience (QoE) for live video transmission applications.
% Hence, it was found that while enabling Uplink MIMO might resulted in more disconnection for the user, especially in challenging RF condition, the spectral efficiency was increased by 17.05\% allowing more effective use of frequency spectrum, which can help MNO boost cell capacity in congested scenarios. Since HPUE has been standardized in 3GPP standard and some countries such us the United States, Thailand, and Japan are now permit the use of them, using HPUE on Uplink-MIMO-enabled UE and network can increase overall Quality-of-Experience (QoE) in live video transmission applications.
\subsection{Uplink 256QAM}
\begin{figure}[t!]
\centering\includesvg[width=0.95\linewidth,inkscapelatex=false]{OBS-256QAM.svg}\\[2mm]
\centering\includesvg[width=0.35\linewidth,inkscapelatex=false]{256QAM_Legend.svg}
\caption{\textbf{UL-256QAM}: Live Encoding Performance}
\vspace{-7mm}
\label{fig:OBS-256QAM}
\end{figure}
\begin{figure}[t!]
\centering\includesvg[width=0.95\linewidth,inkscapelatex=false]{RF-256QAM.svg}\\[2mm]
\centering\includesvg[width=0.35\linewidth,inkscapelatex=false]{256QAM_Legend.svg}
\caption{\textbf{UL-256QAM}: RF Parameters}
\vspace{-5mm}
\label{fig:RF-256QAM}
\end{figure}
\begin{figure}[t!]
\centering\includesvg[width=0.85\linewidth,inkscapelatex=false]{QAM-256QAM.svg}
\caption{\textbf{UL-256QAM}: Uplink Modulation Utilization (\%)}
\label{fig:QAM-256QAM}
\vspace{-7.5mm}
\end{figure}
% Figure \ref{fig:OBS-256QAM} presents the dropped frame percentage and the number of disconnections from the RTMP server across two frequency band: n3 (1800 MHz) and n28 (700 MHz) when UL-256QAM is enabled (\textit{qam256} MCS Table is used) and disabled (\textit{qam64} MCS Table is used). When considering the percentage of dropped frame, a mixed result was observed as it was found that enabling UL-256QAM reduced the amount of dropped frame by 0.6\% for the higher frequency band n3, while increase the number of dropped frame by 0.8\% for the lower frequency band n28. However, it was found that enabling UL-256QAM can significantly reduce the number of connection re-establishment, as the results shows the number of disconnection reduced from three and five, for frequency band n3 and n28, respectively, to two across the board. This resulted in a more consistent experience for the users.
Figure \ref{fig:OBS-256QAM} presents the dropped frame percentage and the number of disconnections from the RTMP server on frequency bands n3 (1800 MHz) and n28 (700 MHz), comparing performance with UL-256QAM enabled (\textit{qam256} MCS Table) versus disabled (\textit{qam64} MCS Table). The results for dropped frames were mixed; enabling UL-256QAM decreased the dropped frame rate by 0.6\% on the higher frequency band n3 but increased it by 0.8\% on the lower frequency band n28. However, a notable improvement in connection stability was observed. Enabling UL-256QAM reduced the number of connection re-establishments from three on band n3 and five on band n28 to a consistent two across both bands, suggesting a more stable user experience.
% Figure \ref{fig:RF-256QAM} presents RF parameters logged by \textit{Network Signal Guru (NSG)} during the UL-256QAM experiment. Similar to Uplink MIMO cases, because Constant Bit Rate mode was used with the target bitrate of 3000 kbps, there is no significant difference between each scenario. However, it was found that while enabling UL-256QAM can slightly improve spectral efficiency, its benefit is mostly offset by a significant increase in Block Error Rate (BLER), which skyrocketed from 6.92\% to 8.27\% for frequency band n3, and from 8.62\% to 9.72\% for the frequency band n28. This resulted in slightly worse spectral efficiency as the number of uplink resource block consumed by the UE increased by 9.19\% and 8.43\% for frequency band n3 and n28, respectively.
Figure \ref{fig:RF-256QAM} shows the RF parameters logged by \textit{Network Signal Guru (NSG)}. As in the Uplink MIMO experiments, the use of a Constant Bit Rate (CBR) mode resulted in no significant throughput difference between scenarios. However, enabling UL-256QAM had a detrimental effect on link reliability at the physical layer, evidenced by a significant increase in the Block Error Rate (BLER). The BLER increased from 6.92\% to 8.27\% for band n3 and from 8.62\% to 9.72\% for band n28. This degradation led to a decrease in spectral efficiency, as the number of uplink resource blocks consumed by the UE increased by 9.19\% and 8.43\% for bands n3 and n28, respectively.\looseness=-1
% Further analysis of spectral efficiency is presented in Figure \ref{fig:QAM-256QAM}, which shows the utilization percentage of each modulation scheme for PUSCH. It was found that while enabling UL-256QAM resulted in 17.7\% and 30.8\% of resource block being modulated using 256QAM modulation, the amount of resource block being modulated using the lower order QPSK and 16QAM also increased slightly by 5\% each, offsetting the spctral efficiency from utilizing 256QAM modulation. This is likely due to the implementation MCS Table in 5G standard, only 32 MCS entries were permitted. Therefore, implementation of qam256 Table is done by eliminating the intermediate target code rate of lower order modulation. For example, in qam64 Table, there are ten, seven, and twelve target code rate for QPSK, 16QAM, and 64QAM modulation, respectively. While there are only five, six, and nine target code rate for QPSK, 16QAM, and 64QAM modulation, respectively, in the qam256 Table. This resulting in reduction in link-adaptation capability under less than ideal signal condition. Some RAN equipment manufacturers offers a proprietary solution to mitigate the issue by dynamically switch between two MCS table depending on the UE uplink signal quality. However, this is beyond the scope of this work.
As detailed in Figure \ref{fig:QAM-256QAM}, which illustrates the utilization of each PUSCH modulation scheme. When the \textit{qam256} MCS table was enabled, 17.7\% and 30.8\% of resource blocks were modulated using 256QAM on bands n3 and n28, respectively. However, the utilization of lower-order modulations like QPSK and 16QAM also increased by approximately 5\% each, which negated the potential spectral efficiency gains. This outcome is a consequence of the MCS table implementation in the 5G standard, which permits a maximum of 32 entries. The \textit{qam256} table accommodates the higher modulation by removing intermediate target code rates for lower-order modulations. For instance, the \textit{qam64} table defines ten, seven, and twelve target code rates for QPSK, 16QAM, and 64QAM, respectively, whereas the \textit{qam256} table reduces these to five, six, and nine. This reduced granularity impairs link-adaptation in suboptimal conditions. Although some RAN vendors offer proprietary solutions to dynamically switch between MCS tables based on signal quality, such features are beyond the scope of this work.
% Hence, while UL-256QAM can slightly increase the channel capacity, which resulted in more stable connection at the application layer, enabling them resulted in a negative impact at the physical layer as it was found that the UE consume significantly more resource block, while exhibiting higher Block Error Rate (BLER), which offsetting the spectral efficiency gain from enabling UL-256QAM in the first place. Therefore, the MNO should examine the channel quality and cell capacity carefully before considering enabling UL-256QAM. Furthermore, an improvement in MCS Table implementation with more than 32 entries should be considered toward the Next-Generation RAN (NG-RAN), which is expected to incorporate higher modulation index like 1024QAM.
In conclusion, while enabling UL-256QAM can improve channel capacity and result in a more stable application-layer connection, it negatively impacts physical layer performance. The UE consumes significantly more resource blocks and exhibits a higher BLER, offsetting the spectral efficiency gains from 256QAM. Therefore, MNOs should carefully evaluate channel quality and cell capacity before deploying this feature. Furthermore, future Next-Generation RAN (NG-RAN) standards should consider expanding the number of MCS table entries to better support higher modulation schemes such as 1024QAM.
\section{Conclusions and Future Work}
% In this paper, the performance of Uplink MIMO (UL-MIMO) and UL-256QAM (UL-256QAM) for live HD video streaming over a commercial 5G Standalone (SA) network was investigated. The results show a significant trade-off between improving spectral efficiency and maintaining connection stability for real-time application. Enabling UL-MIMO demonstrated a substantial 17.05\% improvement in spectral efficiency due to the utilization of spatial multiplex, which reduced the consumption of uplink resource block. This enhancement is highly beneficial for increasing overall cell capacity, particularly in congested environments. However, the gain in efficiency came at the cost of reduced connection stability, with the number of RTMP disconnections increasing from five to eight when UL-MIMO was active. This instability is attributed to the division of the UE's 23 dBm power budget across two antennas, which can impair performance in poor channel conditions.
In this paper, the performance of Uplink MIMO (UL-MIMO) and UL-256QAM for live HD video streaming over a commercial 5G Standalone (SA) network were investigated. The results show a significant trade-off between improving spectral efficiency and maintaining connection stability for real-time applications. Enabling UL-MIMO demonstrated a substantial improvement in spectral efficiency by reducing resource block consumption, which is highly beneficial for increasing cell capacity in congested environments. However, this gain came at the cost of reduced connection stability, evidenced by a significant increase in RTMP disconnections when UL-MIMO was active. This instability is attributed to the division of the User Equipment's (UE's) 23 dBm power budget across two antennas, which can impair performance in poor channel conditions.
% Conversely, enabling UL-256QAM improved application-layer stability by reducing the number of disconnections. However, this benefit was offset by degraded physical layer performance, as the results shows a significantly increased Block Error Rate (BLER) and 8-9\% increase in resource block consumption. The degradation in physical layer link stability is due to the limited number of entries in the standardized \textit{qam256} MCS table, which impairs link-adaptation capabilities, negating the potential spectral efficiency gains. These results indicate that while these 5G features can enhance network performance, their application to live streaming requires careful consideration of deployment environment, number of user, cell planning and overall channel quality.
Conversely, enabling UL-256QAM improved application-layer stability by reducing the number of disconnections. However, this benefit was offset by degraded physical layer performance, as the results show a significantly increased Block Error Rate (BLER) and an 8-9\% increase in resource block consumption. The degradation in physical layer link stability is due to the limited number of entries in the standardized \textit{qam256} Modulation and Coding Scheme (MCS) table, which impairs link-adaptation capabilities and negates potential spectral efficiency gains. These results indicate that while these 5G features can enhance network performance, their application to live streaming requires careful consideration of the deployment environment, number of users, cell planning, and overall channel quality.
% For future work, as identified in this study, the main drawback of UL-MIMO is the reduced transmission power per antenna chain. HPUE, which permits higher transmission power, is the standardized solution to mitigate this issue. Therefore, a future investigation is required to understand the affect of HPUE on real-time live-streaming application, especially on UL-MIMO-enabled UE, whether or not if HPUE can can mitigate the observed connection instability, allowing live-streaming applications to benefit from UL-MIMO's spectral efficiency gains without compromising reliability. Additionally, the current implementation of MCS Table in 5G standard prompt a future research for a better implementation of MCS table for the Next-Generation Radio Access Network (NG-RAN) standards to accomodate higher-order modulation like 1024QAM as we progressed toward the 6G era.
For future work, an investigation into High-Power User Equipment (HPUE) is needed to determine if its higher transmission power can mitigate the connection instability observed with UL-MIMO, thereby allowing applications to benefit from its spectral efficiency gains without compromising reliability . Additionally, future Next-Generation RAN (NG-RAN) standards should address the limitations of the current MCS Table design, which impairs link-adaptation capabilities, to better accommodate higher-order modulations such as 1024QAM as networks evolve toward 6G . Finally, to generalize the findings beyond the specific context of this study, urban commuter rail lines in Tokyo, subsequent research should encompass a wider variety of environments, including diverse geographies, mobility scenarios, network operators, and real-time protocols such as SRT or WebRTC.
\section*{Acknowledgement}
This paper is supported by the Japan Science and Technology Agency (JST) CRONOS Grant Number JPMJCS25N2. Additionally, the author would like to express their gratitude to \textbf{PEI Xiaohong} of \textit{Qtrun Technologies} for providing \textit{Network Signal Guru (NSG)} and \textit{AirScreen}, the cellular network drive test software used for result collection and analysis in this research.
% use section* for acknowledgment
% trigger a \newpage just before the given reference
% number - used to balance the columns on the last page
% adjust value as needed - may need to be readjusted if
% the document is modified later
%\IEEEtriggeratref{8}
% The "triggered" command can be changed if desired:
%\IEEEtriggercmd{\enlargethispage{-5in}}
% references section
% can use a bibliography generated by BibTeX as a .bbl file
% BibTeX documentation can be easily obtained at:
% http://mirror.ctan.org/biblio/bibtex/contrib/doc/
% The IEEEtran BibTeX style support page is at:
% http://www.michaelshell.org/tex/ieeetran/bibtex/
%\bibliographystyle{IEEEtran}
% argument is your BibTeX string definitions and bibliography database(s)
%\bibliography{IEEEabrv,../bib/paper}
%
% <OR> manually copy in the resultant .bbl file
% set second argument of \begin to the number of references
% (used to reserve space for the reference number labels box)
\setstretch{1}
\Urlmuskip=0mu plus 1mu\relax
\bibliographystyle{IEEEtran}
\bibliography{b_reference}
% that's all folks
\end{document}