The Broadcast Engineering and IT (BEIT) Conference is designed for broadcast engineers and technicians, media technology managers, broadcast equipment manufacturers, and R&D engineers.

This BEIT Conference Opening will feature a keynote, as well as the presentation of the Best Paper Award and Best Student Paper Award honoring the author(s) of the best paper and student paper published in the NAB Broadcast Engineering and IT Conference Proceedings.

While analog FM and hybrid HD Radio boosters in Single Frequency Networks (SFNs) have been widely studied, deployed, and discussed, this paper will explore experimental observations of the first fielded high-power IBOC digital-only booster using off-the-shelf Gen4 HD Radio™ hardware.  What makes this different from existing installations is that the booster transmits a digital-only MP1 signal with no host analog FM component.  Topics include potential applications, appropriate timing of implementations, and impact on analog and digital signal performance in the field.

The evolution of live radio broadcasting is accelerating, driven by advancements in cloud technology and AI. As radio professionals face increasingly complex demands in delivering real-time, high-quality broadcasts, Super Hi-Fi, in collaboration with key equipment and ecosystem partners, is developing a next-generation, cloud-native live radio broadcasting solution that reshapes the traditional workflow from the studio to the airwaves. This technical presentation will delve into the architecture and capabilities of an innovative, IP-delivered signal path and operational model for live radio, designed for maximum flexibility, efficiency, and quality.

Imagine a broadcast workflow that functions seamlessly from any location, including mobile devices like an iPad, with no drop in quality. Picture real-time video collaboration among show talent and production teams from anywhere globally, merging seamlessly with both in-studio and remote participants. This system transcends traditional barriers by integrating fully cloud-based audio processing—mixing, mastering, and production—powered by advanced AI algorithms. This session will provide a detailed exploration of the technology and vision behind this approach, its implications for talent, and the evolving role of broadcast engineers and IT managers in supporting it.

Key Areas of Focus:

1. Architecture of a Cloud-Native Broadcast Workflow

This discussion will cover the technical architecture behind an end-to-end cloud-native live radio IP-delivered signal path, including integration with studio hardware and transmitter-based playout devices with proper metadata support. We’ll explore how the system leverages cloud-based AI to manage audio production in real time, dynamically adapting between fully-automated song selection, sequencing, segues and imaging with integrated live breaks segments, ensuring broadcast-grade quality with unparalleled real-time control.

2. Revolutionizing Talent Collaboration

With real-time video and audio connectivity built into the platform, this model allows for unprecedented collaboration flexibility. Attendees will learn how remote talent can join studio-based colleagues without latency or quality issues – leveraging cloud infrastructure to create a unified, live production environment.

3. Operational Efficiencies for Engineers and IT Managers

This cloud-based model significantly reduces the physical equipment and maintenance traditionally required for radio broadcasting. Engineers and IT managers will gain insights into a streamlined deployment and operational model with centralized monitoring and control, empowering technical teams to deliver broadcasts with a fraction of the complexity and cost.

4. AI-Driven Audio Processing

By leveraging AI for automatic song selection, sequencing, segues, mixing, mastering, and content processing, this approach enables talent to focus on being fully-present with their colleagues to connect with their audience, without worrying about managing the minute-by-minute minutiae. This segment will describe the specific AI models and processing techniques employed in achieving professional-quality output in a fully automated environment.

6. Case Study: A New Era of Remote and Hybrid Radio Broadcasting

Super Hi-Fi will present a practical case study on the deployment of this technology, highlighting the technical challenges, solutions, and results from early-stage testing with partner radio companies and equipment manufacturers. Attendees will gain a real-world perspective on the future of radio broadcasting, with actionable insights into the implementation and integration of cloud-based workflows.

This session will guide broadcast engineers, technology managers, and IT professionals through the practical and technical aspects of implementing the world’s first fully AI-powered, cloud-native, live radio broadcasting solution. Attendees will leave with an understanding of how to support talent and operational needs with reduced infrastructure and cost requirements, enhanced flexibility, and a forward-thinking approach that positions radio at the cutting edge of media technology.

GNSS vulnerability is well-known, and search continues to find another independent system that can provide PNT services when GNSS services are denied or unavailable. NAB has invented and has been deploying, with collaboration with TV broadcasters, a terrestrial PNT solution known as Broadcast Positioning System (BPS) that uses the ATSC 3.0 enabled TV facilities to deliver time. BPS can provide timing to national critical infrastructure that depends on GNSS time and thus is vulnerable. When fully deployed, this free-to-use service can also be used to complement GNSS service for the public.

BPS has been deployed at WHUT in Washington, DC and in KWGN at Boulder, Colorado. BPS is also being deployed at WNUV at Baltimore, Maryland. NAB has been working with NIST, DOT, and DHS on BPS testing and deployment. With collaboration with ATSC 3.0 equipment vendors, NAB is also in the process of developing BPS network operations center (NOC) software and a new BPS receiver.

This presentation will be a progress report and current development status of BPS in the following areas: (1) installation and system calibration; (2) deployment status; (3) time delivery performance of the deployed systems; (4) new receiver development; and (5) monitoring health of the system with NOC.

In today’s multimedia landscape, radio has transformed dramatically, adapting to the demands of a more interconnected audience. Advancements in broadcasting technologies now enable you to expand your audience reach by delivering live radio programming seamlessly across both radio and video platforms. This innovation opens up new avenues to connect with listeners, all while significantly reducing the traditional costs and efforts associated with TV production

At the forefront of this transformation is the implementation of an effective visual radio management software. This innovative tool allows radio stations to produce captivating live broadcasts without the necessity of a specialised video operator. Featuring advanced camera-switching technology that employs automatic voice detection, the software can effortlessly shift focus to the active speaker, ensuring that viewers remain engaged with dynamic content. Smart automation handles the entire live switching process, eliminating the need for a camera director in the studio and allowing broadcasters to focus on creating great content rather than managing technology.

Moreover, by leveraging on the flexibility of IP and NDI technology, broadcasters can manage live video broadcasts with pre-set playlists and unlimited audio and video previews. The system interfaces smoothly with GPI, control surfaces, and serial controllers, providing a comprehensive solution customisable to diverse broadcasting needs. This seamless integration empowers broadcasters to efficiently organize tracks, build playlists, and schedule multiple channels across Radio, TV, Web, and Mobile platforms.

One of the standout features of an effective visual radio system is its ability to enhance visual streams with stunning 2D and animated graphics. Broadcasters can incorporate logos, crawls, and custom animations, enriching the viewer experience. For an even more dynamic presentation, Picture-in-Picture effects can be created, allowing for up to eight layers of video, thereby providing a visually engaging environment that captivates audiences.

Furthermore, a centralized dashboard simplifies radio playout management, offering real-time access to essential information such as intros, outros, duration, thumbnails, and artist details. This streamlined command centre enhances operational efficiency and provides complete visibility within a software interface, making it easier for broadcasters to manage content.  

Additionally, integration with Media Asset Management systems streamlines the organization of audio assets.  Advanced metadata fields enhance sorting capabilities across scheduling and logging applications, making it simpler for users to find and utilize their media effectively.

With these power-packed features, broadcasters can elevate their capabilities and reach a wider audience than ever before. Embracing innovative tools such as visual radio software enables radio stations to engage their audiences in new ways, transforming the listening experience. As the broadcasting landscape continues to evolve, these advancements represent an important step forward, offering the potential to redefine how radio content is created and consumed.

In conclusion, the future of broadcasting lies in the integration of advanced technologies that enhance both the quality and accessibility of content. By leveraging visual radio software, broadcasters can connect with their audiences on multiple platforms, making the radio experience more interactive and compelling than ever.

Department of Homeland Security (DHS) conducted a study of PNT survivability if GPS satellites are disabled or lost. The committee that conducted the study consisted of PNT experts and consultants. The committee identified that a combination of eLORAN and BPS could meet the PNT survivability needs.mThis panel will highlight the strength and weaknesses of a joint eLORAN and BPS solution.

The broadcast industry is undergoing a major shift as broadcasters increasingly integrate cloud-based solutions to enhance their operations’ efficiency, reliability, and scalability. The advent of ATSC 3.0—the international ‘state-of-the-art’ broadcast and data distribution standard—enables the entire broadcast operations chain to become ‘cloud-native,’ opening unprecedented opportunities for flexible and resilient deployment strategies. While a broadcast facility’s operations can be fully migrated to the cloud with cloud-native operations, this paper proposes a unified approach for cloud-based broadcast integration. We will examine how ATSC 3.0’s capabilities can support fully integrated broadcast infrastructures.

By analyzing current and potential implementations, we highlight the transformative potential of cloud-based broadcasting for optimizing playout, signal distribution, and emission. The goal is to provide a blueprint for broadcasters navigating the shift to a future-ready, cloud-centric operational paradigm.

This paper presents a reference architecture for implementing ATSC 3.0/TV 3.0 workflows in the cloud, facilitating Studio-to-Transmitter Link (STL) over public internet and offering a practical framework for researchers and practitioners. Key features of ATSC and TV 3.0 are highlighted, emphasizing its scalable and flexible cloud-based deployment capabilities. With all components of an ATSC/TV 3.0 system, except the exciter and transmitter, implemented in the cloud, the paper details the architecture’s handling of ROUTE/DASH streams, Physical Layer Pipes (PLPs), and STLTP. It explores dual reception capabilities, allowing receivers to access content from both broadcast and broadband sources simultaneously, and addresses challenges such as CDN overflow and STLTP distribution.

Leveraging AI tools can dramatically reduce time spent on back-office tasks, allowing creative teams to focus more on storytelling and artistry. This panel will explore how Gen AI is being used to enhance, rather than replace, key elements of the filmmaking process—from initial scriptwriting through to post-production. 

Through real-world examples, this session will offer deep insights into how the new AI-powered tools are reshaping the production process. Featuring leading innovators from various AI-driven filmmaking platforms and professional cinematographers.

Attendees will hear about real-world examples of AI augmenting creativity in filmmaking, gain insights on adopting AI tools to optimize workflows, and come away with practical tips on integrating AI without losing creative control.

NextGen TV, enabled by the ATSC 3.0 standard, marks a transformative shift in broadcasting, using IP-based delivery to offer enhanced, interactive, and data-driven experiences. Leveraging cloud infrastructure for managing and delivering the ATSC3.0 content presents broadcasters with opportunities and challenges alike. Key advantages of using the cloud include scalability, flexibility, and reduced need for on-premise infrastructure. By moving content delivery into the cloud, broadcasters can dynamically scale resources, reach wider geographic audiences with low latency, and shift from capital expenditure (CAPEX) to operational expenditure (OPEX) models.

However, cloud-based ATSC 3.0 workflows also face notable challenges. One significant issue is using the open internet for both content ingest (upload) and egress (distribution). The open internet introduces risks related to bandwidth limitations, packet loss, and potential service interruptions, impacting the reliability of content delivery. Moreover, cloud providers charge for outgoing data, which can lead to substantial egress costs, particularly given the high-bitrate video typical of ATSC 3.0. Additionally, the cloud’s lack of native multicast support complicates content delivery. While ATSC 3.0 relies on multicast for efficient broadcasting to many receivers, most cloud infrastructures are built around unicast. 

Reliability is also a critical consideration for broadcasters. The cloud enables high availability through multi-region deployments and load balancing, but is not necessarily adapted for the traditional one-plus-one (1+1) redundancy where a backup system mirrors the primary.

This paper study all these problematics, providing proven technical solutions and recommendation – like the usage of SRT, the main and backup architecture with fail over witch… 

In an era where traditional studio infrastructure is being reimagined, this panel explores the revolutionary journey of building and scaling the world’s first completely cloud-native creative studio. Through the lens of Untold Studios’ groundbreaking experience, we’ll dive deep into the technical architecture, operational challenges, and transformative solutions that enable a modern creative powerhouse supporting 300+ artists across four global locations.

The session will examine critical technical decisions and implementations across several key areas:

1. Cloud Infrastructure Evolution

• Architectural decisions that enabled seamless scaling from startup to global studio
• Real-world performance optimization strategies for high-bandwidth creative workflows
• Implementation of secure, low-latency remote collaboration systems
• Practical solutions for managing cloud costs while maintaining creative flexibility

2. Generative AI Integration

• Technical framework for evaluating and deploying AI tools in production
• Infrastructure considerations for running compute-intensive AI workloads
• Real-world examples of AI implementation in creative pipelines
• Strategies for maintaining content provenance and managing ethical considerations

3. Global Workflow Architecture

• Technical solutions for managing distributed creative teams
• Implementation of cloud-native render management systems
• Storage and data management strategies across multiple regions
• Performance optimization for international collaboration

4. Future-Proofing Technical Infrastructure

• Scalable approaches to new technology adoption
• Strategic planning for emerging creative technologies
• Building flexible systems that adapt to rapid industry changes
• Maintaining technical agility while ensuring operational stability

The panel will provide attendees with practical insights into:

• Key architectural decisions that enable cloud-native creative workflows
• Technical requirements for successfully implementing hybrid AI/human creative pipelines
• Real-world solutions to common challenges in distributed creative production
• Strategies for evaluating and adopting emerging technologies
• Approaches to building and managing technical teams in a cloud-first environment

Drawing from Untold Studios’ experience of delivering award-winning work for global brands and entertainment projects, this session will offer valuable insights for broadcast engineers, technology managers, and decision-makers looking to modernize their creative infrastructure. Attendees will gain practical knowledge about cloud migration strategies, technical team development, and the integration of emerging technologies in creative workflows.

The discussion will conclude with a forward-looking examination of how cloud-native infrastructure and AI tools are reshaping the future of creative production, offering attendees actionable insights for their own technology transformation initiatives.

The spectrum allocated to a broadcaster is their most valuable resource, making it essential to find methods and technologies that can optimize its use. Modern standards like ATSC 3.0 include technologies that enhance spectrum efficiency, such as advanced video codecs, an adaptable physical layer, and support for diverse data transmission. However, fully realizing the potential of these innovations remains an ongoing challenge. This presentation examines the infrastructure, implementation techniques, and additional technology necessary to most efficiently utilize broadcast spectrum. Key use cases will be considered, including complete channel utilization to avoid unused (null) packets, dynamic data transfer adjustment to support critical short-term events, and scheduled physical layer adaptation to fit different applications. Additionally, an analysis will compare the spectral efficiency of typical ATSC 3.0 deployments with optimized configurations, highlighting potential improvements in spectral efficiency that can lead to improved channel monetization through increased service adaptability. 

With cloud capabilities, broadcasters can unify their content production workflows, simplify management, and deploy infrastructure that supports both traditional broadcast and broadband (OTT) networks.

The recent adoption of the ATSC 3.0 standard is further enhancing these possibilities by enabling next-generation broadcasting that combines OTA and OTT pathways. ATSC 3.0 enables a convergence of broadcast and broadband. This potential dual-reception capability, where devices can receive content both over-the-air and over-the-top, opens up unprecedented opportunities for broadcasters. It allows for new use cases such as seamless service continuity, where broadband can act as a backup for broadcast delivery, and expanded business models that leverage targeted advertising, personalized content, and interactive features.

This paper explores the technical challenges associated with implementing a seamless, integrated content delivery system that combines OTA and OTT. Key technical hurdles include synchronization of broadcast and broadband streams, managing latency between the two delivery methods, and ensuring a smooth, uninterrupted viewer experience even during signal disruptions.

This paper examines specific use cases in detail, such as:

• Seamless Service Delivery: Ensuring content continuity by automatically switching to broadband delivery when broadcast signals are weak or interrupted.
• CDN Load Management: Addressing CDN overflow challenges during times of increased demand or broadcast disruptions, including strategies for managing server load and maintaining quality of service.
• Additionally, the paper will analyze infrastructure requirements, such as scalable cloud resources and effective CDN configurations, needed to support this hybrid delivery model. 

In conclusion, this paper aims to provide broadcasters with a comprehensive guide to adopting and deploying hybrid content delivery models that leverage the strengths of both broadcast and broadband networks.

C-band broadcast distribution links are susceptible to various kinds of disturbance like 5G interference, sun transit or snow accumulation.

This presentation explores an alternative approach for interference mitigation based on SRT-based backup links over the public internet. SRT is an open, industry-wide protocol which can be used to reliably deliver MPEG-TS-based signals over the public internet. The presentation explains in detail how SRT can be used for backup transmission, which tools the protocol provides and the necessary infrastructure requirements.

Datacasting over ATSC 3.0/NextGen TV offers a variety of new opportunities, especially in public safety and incident response scenarios. Mosaic ATM, alongside business partner Device Solutions, Inc., has received NASA Small Business Innovation Research (SBIR) funding to investigate the feasibility of using ATSC 3.0 to help support wildland fire management operations.

In this presentation, we will discuss a proposed innovation called the NextGen Incident Response Communication System (NIRCS). NIRCS is a rapidly deployable, mobile, long-range broadcast communications system. NIRCS will make use of ATSC 3.0 technology – the digital terrestrial broadcast system built on the internet protocol (IP) – to enable one-way datacasting of IP-compatible data, including ultra-high-definition video, high-fidelity audio, and other types of data packets (e.g., aircraft position messages). One can think of NIRCS as an ATSC 3.0 broadcast station “on the go,” ready to move to and support emergencies on demand.

The intent behind NIRCS is to address one of the NASA ACERO project’s critical technological needs: a reliable, resilient, and secure data communication system for quick data dissemination to support effective decision-making. NASA’s Advanced Capabilities for Emergency Response Operations (ACERO) project focuses on the use of uncrewed aircraft systems (UAS) and other advanced aviation technologies to improve wildland fire coordination and operations. In today’s wildland firefighting operation, UAS, which have many practical and life-saving applications, are considerably separated (in time or space) from the firefight because crewed aircraft (e.g., slurry bombers) are unaware of UAS positions during operations. This uncertainty in UAS position results in an unacceptable collision risk between crewed and uncrewed aircraft, thus the need for expansive separation. With the proposed innovation, NIRCS, collocated with UAS wildland firefighting operations, would datacast the real-time position of UAS to crewed aircraft (ground to air), allowing for the safe and more efficient integration of UAS in the operational airspace. Furthermore, owing to the flexibility of the ATSC 3.0 technology, NIRCS would also datacast UAS operational volumes and UAS surveillance video.

The presentation will cover a variety of topics regarding NIRCS, including details on:

• The need for such a system within the nation’s emergency communication infrastructure,
• The design of the mobile broadcast station, its corresponding ATSC 3.0 receiver, and a wildland firefighting software application for reviewing broadcast content,
• The envisioned use case for wildland firefighting operations and how it will improve air traffic coordination and aviation safety in the operational airspace
• Other important technical considerations, such as RF propagation and anticipated broadcast range in representative operational environments, data sources, data encryption and decoding strategies, and interoperability with other emergency software applications

Digital Broadcast to Handhelds has long been anticipated to revolutionize the way we consume media on the go. However, despite high expectations, the reality has yet to meet these promises. This presentation provides an overview of the various Broadcast to Handhelds solutions, examining both their technical capabilities and limitations. In addition, the presentation delves into the commercial challenges faced by these systems and explores the reasons behind their underwhelming success. Finally, insights into the potential future of digital broadcast to handhelds will be discussed.

The design and implementation of a multi-platform and multi-target advertising campaign currently requires many steps performed by experienced media staff.  This presentation describes providing media knowledge to an AI agent so that a partnership between the campaign planner and the AI agent can efficiently and flexibly create and adjust complex advertising campaigns.  The AI agent can be provided with the arriving RFP documents by the planner, and then it can validate attributes and assign missing values to enable efficient and validated campaign creation.  The AI agent is provided with OpenAPI documentation of the API into a production campaign repository to create campaigns, validate attributes, and manage the campaigns with attribute updates.  Campaign analysis and graphical reporting can be flexibly requested to monitor the detailed campaign assignments.  What-if analysis can be requested by the planner to model the campaign outcome metrics so that attributes can be selected that optimize the campaign.  This partnership between the user and the AI agent brings the user’s industry experience together with the modelling and analysis capability of the AI agent to efficiently perform the workflow necessary for a valid and optimized campaign. 

In today’s ultra-competitive media and entertainment industry, captioning demands exceptional precision. However, when quality control (QC) is conducted manually, it is very labor-intensive and prone to errors, which can lead to compromises in quality. For example, the QC process of ensuring that Federal Communications Commission (FCC) guidelines —relating to sync, accuracy, and completeness — are met requires multiple reviews. Furthermore, it involves verification of segmentation, reading speed, display duration, and layout metrics like row and column count. Caption placement also needs careful adjustment to avoid obstructing important visual elements, while global deliveries necessitate multilingual quality checks to meet diverse audience standards. Additionally, profanity censoring is critical. With all these requirements, performing general checks — such as ensuring captions aren’t delayed during critical or suspenseful moments in a scene — can often be overlooked.

This presentation will explore how AI is streamlining these complex QC tasks and freeing up human resources, enabling media companies to focus on the more creative aspects of their workflows. Conference attendees will gain insight into how advanced NLP and language models are addressing segmentation and statistical metrics such as reading speed, while machine learning algorithms for burnt-in text detection, face recognition, and shot boundary detection enable more accurate caption placement and timing adjustments. In addition, the presentation will discuss how media companies can leverage AI to easily check for text accuracy and sync. Machine translation capabilities now support multilingual accuracy checking, which is crucial for global content delivery. Using Large Language Models (LLMs) the translation can be adjusted to fit into desired audio intervals and number of words. Furthermore, the presentation will also highlight the latest advancements enabling LLMs to transcribe and interpret non-speech sounds. These models, often termed Audio-Language Models, can help media companies identify missing audio descriptions, which is a requirement for accessibility.

To conclude, the presentation will discuss the integration of AI to optimize and enhance QC workflows for captioning, emphasizing automation’s growing role in improving quality and efficiency in modern media production.

Live Ultra HD broadcasts require overcoming complex challenges in HDR and SDR infrastructures, from content creation to distribution and device compatibility. This panel will explore the technical, operational, and creative solutions needed to drive global adoption of real-time Ultra HD services.

Experts from content production, network distribution, and device development will share their perspectives, addressing issues such as multi-camera HDR/SDR workflows, maintaining consistent quality across formats, and future-proofing archives for Ultra HD content.

Attendees will also gain insights into managing the transition to Ultra HD through lessons learned from past format shifts, such as SD to HD and B&W to color. The discussion will include questions like:

• How can HDR/SDR workflows optimize live events with multiple camera angles?
• What strategies ensure a consistent HDR/SDR experience without compromising quality?
• How can Ultra HD workflows engage global audiences with distinct visual “looks”?
• What are the best practices for future-proofing live content archives?
• How do broadcasters ensure proper display on devices with varying Ultra HD capabilities?

The panel features leaders from the Ultra HD ecosystem, including broadcasters, device manufacturers, and content distributors.

Attendees will leave with actionable insights into building scalable Ultra HD workflows that balance HDR/SDR requirements while engaging audiences worldwide.

In the rapidly evolving broadcast, OTT, and VOD landscape, interactive and shoppable media experiences are transforming viewer engagement and commercial possibilities. This paper presents a framework for real-time brand and product detection within broadcast, OTT, and VOD content. Leveraging multi-model AI detection engines, SCTE-35, SEI, and VMAP metadata, the system seamlessly integrates product detection with HTML applications, enabling real-time, commerce-enabled media experiences.

OTT and VOD platforms are increasingly popular, shifting media consumption toward personalized, interactive viewing. Users demand seamless, on-demand product information, creating a challenge in synchronizing content playback with real-time shoppable features. This solution bridges content, product detection, and interactivity through a multi-model object detection system. The system integrates machine learning models to recognize brands and products in live streams, dynamically synchronizing this information through SCTE-35, SEI, and VMAP metadata to trigger HTML-based interactions. Real-time synchronization enables personalized shopping experiences across IP, DVB, ATSC, HLS, DASH, and CDN channels, elevating broadcast, OTT, and VOD ecosystems.

The architecture includes workflows for live and VOD content. The Live Content Workflow handles real-time streaming and immediate interactivity. Advertisers input product data and promotional assets, ensuring detection aligns with live content. A flexible Bring Your Own Model (BYOM) approach allows machine learning models to detect brands and products with precision. Detected products trigger SCTE-35 or SEI signals, activating clickable links, QR codes, or overlays in the HTML app for instant viewer interaction.

The VOD Workflow offers thorough analysis through offline media processing for pre-recorded content. Models like YOLO and ResNet ensure high detection accuracy in VOD assets, with metadata aligned before playback. VMAP standards cue interactions at precise moments, enriching playback with shoppable options that resemble live streaming. The asynchronous nature of VOD allows users to interact flexibly—pausing, rewinding, or fast-forwarding—while the HTML application responds to triggers, presenting product information and shopping links without disrupting the experience.

The system’s scalability and adaptability are key. The multi-model detection engine integrates new models and data, ensuring ongoing relevance without needing extensive updates. Models operate in parallel or sequentially, optimizing for performance and accuracy across diverse content, making this solution powerful for high-volume streams and varied business needs.

The framework enhances viewer experience by embedding real-time brand and product interactivity directly into content, offering non-intrusive engagement. Using a shopping bar, overlay, or popup, the HTML app provides smooth, personalized product engagement that complements content without interruptions.

Integration with programmatic advertising platforms also bridges AI-driven detection and ad targeting. SCTE-35 and VMAP metadata ensure detected products trigger relevant ads through real-time bidding (RTB) and direct campaigns, enhancing ad relevance and engagement. This alignment between content analysis and ad-serving allows advertisers to display precise offers for detected brands, optimizing viewer engagement and ad effectiveness.

In summary, this framework enhances interactive commerce within media platforms by combining AI-powered content analysis with real-time shopping integration. By employing multi-model AI detection, metadata signaling, and HTML-based interactivity, this solution aligns with IAB standards and meets the expectations of today’s digital audience. It offers significant potential for expanding interactive commerce in media, providing a powerful tool for content providers, advertisers, and viewers alike. This approach not only enriches the viewer experience but also introduces new commercial opportunities in the ever-advancing broadcast, OTT, and VOD landscape.

The BEIT 2024 paper, ATSC 3.0 Broadcast Positioning System (BPS) Mesh Network, provided a description of a notional mesh network intended to manage and monitor a collection of ATSC 3.0 television transmissions providing high precision traceable time. A preliminary network of these Broadcast Positioning System (BPS) transmissions has been constructed in the Baltimore and Washington DC region using two transmitters and a third simulated transmitter in the NAB 1M laboratory. Another BPS transmitter has been deployed in the Denver area to reach the NIST facility in Boulder, Colorado where the signal is being analyzed.

This paper will describe the actual systems deployed, how they are operating, and the results of the various experiments being carried out using the first installation of a BPS leader / follower architecture in preparation for a fully operational network deployment.

In addition, the paper will describe the initial implementation of a network operating software system which has been developed with the intention of supporting a large, perhaps nation-wide deployment of the BPS Mesh Network.

ATSC 3.0 specifies synchronization requirements in both physical- and data layers which refer to International Atomic Time (TAI) or Coordinated Universal Time (UTC; differing from TAI by an integer number of seconds). As examples, synchronization in the physical layer enables single-frequency networks (SFN); synchronization in the data layer enables coordinated interaction with IP media and other out-of-band services.

As such, ATSC 3.0 transmitters may satisfy an important need for a national-scale high-precision time distribution network, complementary to the Global Positioning System (GPS) and other Global Navigation Satellite Systems (GNSS). Robust, accurate time sources are demanded in telecommunications, power distribution, and financial markets; nanosecond-level precision is required for meter-scale positioning systems.

The optional Broadcaster Positioning System (BPS) protocol inserts metadata into the ATSC 3.0 frame allowing receivers to estimate signal time-of-flight and to identify the time-of-transmission of frame symbols with high precision and accuracy. In principle, BPS enables an independent radio positioning service and may derive its time reference with complete independence from GPS/GNSS, with no constraints on air-chain implementation.

In collaboration with NAB and KWGN (Nexstar), NIST tested several aspects of time transfer stability in a BPS field deployment. Station KWGN transmitted ATSC 3.0/BPS from a position west of Denver, Colorado. Signals are received in three locations: the KWGN studio (22 km east), NIST in Boulder, Colorado (30 km north) and the NIST WWVB radio station near Ft. Collins, Colorado (106 km north). The propagation is believed to be line-of-sight except to NIST in Boulder, located in a slight valley.

At the BPS transmitter and KWGN studio receiver, local time references are derived from a dual-frequency GNSS-disciplined oscillator. At NIST, the time reference is the steered atomic clock ensemble, UTC(NIST). At WWVB, the time reference is an independent atomic clock ensemble loosely steered to follow UTC(NIST), and whose offset is independently and continuously measured by two different GNSS-common-view techniques. At NIST and WWVB, time interval counters (TICs) measured the time difference of pulse-per-second signals from each local time reference and a prototype BPS receiver for more than three weeks.

The measurements of BPS received at NIST show deviations typical of UTC(NIST) – GPS, dominated by a diurnal modulation with a maximum swing of about 35 ns over a month-long interval. Measurements of BPS at the KWGN studio show smaller deviations as the GNSS-derived local time references are highly common.

Fluctuations due to signal propagation, un-compensated multipath interference, and uncorrelated fluctuations in BPS receivers likely bound the performance of BPS time transfer. Measurements of BPS at WWVB, when subtracted from the measurements of BPS at NIST remove the influences of the BPS transmitter time reference and BPS synchronizer element, allowing tests of the propagation and remote receiver instability. Another subtraction of UTC(NIST) – WWVB time scales (obtained by common-view GNSS) removes the remaining time reference fluctuations but inserts noise/biases of the common-view GNSS measurement. In such a “double-difference,” the stability of BPS time transfer over the tested baseline exhibited a peak fluctuation amplitude of about 15 ns, or a peak time-deviation (TDEV) statistic of 2 ns over one-day averaging intervals. This level of noise is comparable with the GNSS common-view technique over similar baselines.

We note other BPS test activity at NAB studios (Washington, DC) involving stations WHUT and WNUV. We will also comment about BPS time reference implementations robust against transient GNSS outages or with complete independence from GNSS.

In today’s competitive media landscape, cost-efficiency and high-quality production are paramount. Media service providers are tasked with attracting viewers while managing tight budgets and industry demands. And for content creators, they face increasingly challenging and complex workflows.

For Paramount, production of their entertainment brand MTV was faced with schedule and resource issues when performing and formatting fixes to master files. The solution an advanced integrated delivery platform built with Paramount designed to streamline content creation, production and delivery workflows for media and entertainment professionals.

Redelivery of long-form content is a common issue that affects production schedules and resource allocation. When master files need corrections, delays in the content supply chain and quality control (QC) processes create inefficiencies and additional costs. To address this, organizations need a way to quickly fix content stored in the cloud without resorting to costly mastering sessions.

Previously to make changes to MTV’s programming content the entire master file would need to be downloaded, which was costly affecting schedules and resources for Paramount. Integration of the Adobe Frame.io and ConneX interfaces allows patches to be easily applied and dropped in where they are needed without the re-delivery of long-form content. Production and operations teams can address QC notes in the content supply chain. This approach enables teams to upload only the necessary changes to Frame.io, review them, and directly insert them into existing cloud-based master files, without the time and bandwidth required for large file redeliveries. This integration provides a powerful production-focused workflow, reducing the need for additional mastering or complex backend conversions to work with existing workflows.

The main benefits for media service providers include:

• Reduced Supply Chain Delays: The ability to fix content in the cloud streamlines QC processes and reduces delays.
• Lower Production Costs: Minimizing rework and avoiding costly mastering sessions lead to cost savings.
• Enhanced Collaboration: A shared platform for production and operations to address QC notes efficiently.
• Enhanced Efficiency: Reduces production delays and costs associated with traditional mastering sessions, ensuring a faster and more efficient workflow.
• Seamless Integration: Allows content creators to make precise production and formatting fixes directly to cloud-hosted master files.

This session will detail the common challenges in content delivery and how the application of advanced technology enabled quick fixes to cloud-stored content without costly mastering sessions. It will share the underlying technology now used to broadcast Paramount MTV and the cost and efficiency savings of the new production-focused workflow.

Broadcast Positioning Systems (BPS) necessitate precise time synchronization across base stations, utilizing a shared time reference and their fixed locations for triangulation and reliable positioning, with time over fiber solutions connected to trusted timing sources serving as a viable method for maintaining accurate synchronization. In this paper, we present a concept to link BPS leader stations to timescales maintained by National Metrology Institutes (NMIs) using High Accuracy White Rabbit synchronization through optical networks.

This panel highlights how cloud-native solutions are revolutionizing live sports broadcasting. Industry leaders will discuss real-time editing, seamless global collaboration, and rapid delivery of high-quality content, all while reducing costs and environmental impacts. Topics include the power of cloud-native workflows, the integration of familiar tools like Adobe Premiere, and the role of social media teams in enhancing audience engagement. Attendees will gain actionable strategies for achieving faster turnarounds, optimizing global collaboration, and driving fan interaction through innovative cloud-based approaches.

The digital advertising landscape is evolving, and with it, the lines between traditional broadcast and Digital Out-of-Home (DOOH) are blurring as both platforms are experiencing similar needs and similar challenges. Even though broadcast is a “Lean Back” experience and DOOH is a “Passing By” experience, both need to get messages communicated effectively, and with minimal interruption to a person’s day. Delivering the right message to the right person at the right time is timeless and the technology available today is making this easier to do than ever before. This paper explores the concept of using an end-to-end datacasting topology to converge these two mediums, enabling the delivery of contextually relevant content in near real-time without relying on first-party data. These insights were gained from decades of experience in both Digital Signage and Broadcast, along with a three-year journey supporting ATSC 3.0 “datacasting to digital signage” trials throughout the country.

The paper will examine several factors of the datacasting topology that was developed throughout these trials and demonstrations. For example, pinpointing what non-personally identifiable information (PII) can be leveraged across the different data types and sources available. For example, what is the impact of second- and third-party data on relevance and how does it compare to the speed of targeted advertising with first-party data? When it comes to personalized marketing, there is strong value associated with first-party data because it is collected from direct interactions, such as web searches. Comparatively, second-party data is collected through trusted business partnerships. While perhaps less viable for targeted advertising, it can enhance audience reach. Third-party data, collected through indirect relationships such as mobile apps, is helpful for audience expansion and new market exploration. We will explore how each data set serves a unique purpose in a DOOH marketing strategy.

The level of trust that exists in this data, content, and messaging will also be explored as critical success factors as they can be differentiators in a world of growing misinformation and Deep-Fakes. We will review sources and methods of generating relevant content and adapting existing content, including advertising, to become more relevant with the use of AI. We will also discuss how the convergence of live linear and file-based content can positively affect public safety, public communications, and smart city support.

Once the data and the content are understood we will review the technologies which bridge the gap between broadcast and DOOH in a combined linear and non-linear model, AKA datacasting. Once the content is delivered, we will review how leveraging local storage, various sensors, AI and Edge-computing can construct a unique and engaging experience at each edge point for DOOH. This same topology of a mix of file-based and live-linear media could be used for local broadcast delivering to their audiences at home to create personalized experiences. Finally, methods of measuring and closing a feedback loop for continuous improvement and automation.

The ultimate goal is to understand a new approach to create, deliver, measure, and monetize content effectively in this new converged space.

As media-workflows migrate to open-source software-defined frameworks consisting of a collection of containerized applications running as processing nodes on a cluster of common-of-the-shelf (COTS) servers, there is a need for these distributed application processes to asynchronously share media essence data in the most efficient and secure way possible without going back to the baseband transport.   This sharing of media essence data is complicated by the fact that some applications perform GPU processing requiring the essence data to be in GPU memory while others perform media processing on the CPU requiring the essence data to be in system memory.   This requirement for asynchronous media exchange between media functions is also specified as part of the EBU Dynamic Media Facility reference architecture.  Several technologies exist for such sharing of media essence data between processes on different compute nodes of a cluster so the ultimate solution needs to support multiple transports with an application layer API that also facilitates initiating and terminating communications, in the data center, on the edge or in the cloud.  Our paper will compare the various techniques and demonstrate how they can be deployed by applications on a Kubernetes-based platform for the easy development and deployment of distributed media process pipelines for transcoding and AI applications.

Spectrum for wireless microphones has been disappearing over the past decade. What’s left is becoming more and more crowded as the use of wireless mics proliferates. WMAS (Wireless Multichannel Audio System) is a new technology that uses spectrum more efficiently, but does it have all the answers? This panel, presented by the North American Broadcasters Association, will explore the pros and cons of WMAS and lay out the issues you need to consider when choosing your next audio upgrade. 

The term “virtualization” is the current big buzzword in the broadcast industry. Virtualization is seen as a path to help reduce infrastructure and maintenance costs and system complexity. Virtualization technology opens new possibilities that could provide cost savings and new functionalities, but virtualization has its limits. Not every application is well-suited to run on a virtual machine. This is particularly the case when there is a complex matrix of regulatory, security, and operational considerations with which to contend. The Emergency Alert System (EAS) is just such an environment. 
 
In this paper, we present the solution to integrating EAS operations into the modern virtual environment while maintaining the integrity of EAS message delivery and adhering to the range of FCC regulatory requirements.
 
First, we review the current configurations of EAS message presentation in various broadcast environments, representing radio and TV transmission chains. By comparing these workflows, one can better understand the issues and solutions available between these two examples. To wit, the modern television air chain has been using IP messaging elements for years. Radio is just now catching up but still faces unique challenges to ensure EAS messages are correctly presented.
Second, we outline an architectural approach supporting virtualization, which involves distributing tasks between advanced EAS edge devices, virtualized applications, and centralized cloud servers. Dedicated edge appliances in this virtualized architecture serve as specialized hardware platforms explicitly designed for the required emergency alerting tasks while seamlessly interfacing with the virtualized broadcast system and supporting other cloud-based services. Overall, dedicated EAS edge appliances continue to play a critical role in enabling the deployment of virtualized workloads at the edge.
We outline and discuss many key benefits and pitfalls, such as:
Streamlined Infrastructure

• More efficiently integrating EAS functionality into existing systems.
• Demonstrate how a single network connection links the EAS system to automation systems and encoding/transmission equipment, reduces equipment footprint, and simplifies system design.

Flexibility and Scalability

• IP-based workflows allow more options for device placement to modernize operations.
• Working across multiple encoding systems eliminates separate units for each program stream.

Improved Efficiency

• Streamlining air chains reduces points of potential failure.
• Configuration is simplified, requiring only a network connection between systems.

Enhanced Alert Distribution

• Improved efficiency increases the reliability of alert distribution, which is critical to viewers and listeners.
• Geotargeting of alerts ensures they are correctly routed to specific areas.

Costs — Savings or Spending

• Are these implementations a lower-cost alternative, or will they ultimately cost more over the long term?

Future-Proofing

• How this approach allows for easier integration of newer, higher-performance equipment while maintaining emergency alert capabilities.
• The flexible nature of IP-based systems ensures compliance with current and potential future FCC regulations.

Finally, we will summarize with real-world examples showcasing various facilities that have deployed these workflows.

As productions continue to move towards software-defined workflows and delivery, we continue to add traditional capabilities to our cloud and cloud delivery options.

This presentation will examine the realities of preparing for production with high resolution (4K and beyond) realtime workflows, and adding HDR capabilities to both spatial and non-spatial delivery formats.

With the increasing demand for rapid and reliable emergency delivery, this paper introduces a cloud-based Advanced Emergency Information (AEI) system designed to enhance public safety by integrating emergency alerts from the Integrated Public Alert and Warning System (IPAWS). This system captures and updates alerts in real-time within the cloud, enabling prompt, synchronized delivery to ATSC 3.0 broadcasters for over-the-air (OTA) transmission and various service operators for IP-based distribution. The proposed AEI solution addresses interoperability challenges across different broadcast and service platforms, ensuring that emergency information reaches diverse audiences effectively. The system architecture, alert processing workflow, and delivery methods for both OTA and IP networks are examined in detail, demonstrating the potential to transform how emergency information is managed and shared across multi-channel delivery platforms.

The advent of NextGen TV, based on the ATSC 3.0 standard, marks a transformative shift in television broadcasting– from a traditional one-way transmission model to an interactive IP-based ecosystem. ATSC 3.0 not only enhances picture and sound quality but also fundamentally redefines interactivity and dynamic content delivery in broadcast television. Leveraging IP-based technologies, this standard enables two-way communication, enabling unprecedented user engagement. Viewers can now interact directly with content by participating in live polls, playing games, buying items, exploring additional contextual media and accessing on-demand programming. These interactive features represent a significant evolution in user experience, allowing audiences to engage more deeply and autonomously with content.

A leading example of ATSC 3.0’s potential is the NBCUniversal’s NextGen TV broadcast app developed in collaboration with Fincons. The app utilizes this standard along with the RUN3TV framework as an accelerator to access features and API’s provided by ATSC 3.0, to avoid duplicative developments of non-differentiating features and enabling full user experience customization, including navigation controls, branding, and other design and interactive elements.

This innovative app leverages NextGen TV’s full potential, introducing personalization, hyper-localization and enhanced content features to broadcast television, to improve viewer experience and enable advanced engagement and measurement. Main features include:

• restart any program from the beginning;
• VOD content access;
• local weather forecasts;
• advanced public safety alerts;
• real-time measurement.

Latest evolutions of NextGen TV applications seen in the market are also enabling new monetization opportunities through the delivery of interactive advertising on broadcast channels and dynamic advertising insertion (DAI) within VOD content, exposing each viewer to advertisement spots tailored to their profiles.

ATSC 3.0 broadcast applications really proved it’s possible to deliver deeply personalized, engaging and immersive experiences on broadcast TV. This trend also includes the possibility to interact with XR content, extending experiences through companion devices that allow the user to explore and navigate virtual environments viewable on connected TV. Fincons is gaining relevant experience in this field taking part in the XReco project, an initiative co-funded by the European Commission under the Horizon Europe program, that promotes the creation and management of new type of content through state-of-the-art XR services, paving the way for their regular distribution and use.

Advanced technologies leveraged to simplify the generation of XR content include:

• Neural Radiance Fields (NeRF): enables 3D scene reconstruction from 2D images;
• Holoportation: offers real-time volumetric capture for immersive remote experiences;
• Free Viewpoint Video (FVV): allows immersive scene navigation and real-time rendering from any virtual camera angle.

State of the art technologies are significantly simplifying XR content generation, that now can be effectively distributed through NextGen TV, creating the ideal conditions to push broadcast television interactivity to a whole new level.

The VVC video coding standard was finalized in July 2020. As the five-year anniversary approaches in 2025, it is a good time to review VVC’s deployment status.  This presentation details major milestones in VVC deployment and offers a comparison with HEVC at the equivalent post-finalization point in time.

Deployment of a video codec involves orchestrating many ecosystem participants and their systems. Adoption of a codec into a broadcast standard is one of the first, rather than one of the last, steps in the deployment process. Thus, after becoming part of broadcast standards more than a decade ago, HEVC had to be implemented by encoder vendors and component manufacturers, then integrated into devices and workflows. We now see HEVC services being activated, although deployment status varies greatly by geography. For OTT the situation is different, with HEVC uptake occurring much more rapidly.

This holds some lessons when it comes to HEVC’s successor, VVC. Most obviously, although VVC is included in (or on the roadmap for) major broadcast standards, services won’t commence overnight. We can nonetheless identify key milestones in terms of product availability that indicate progress.

In this presentation, we divide ‘deployment’ into three categories. First, we consider hardware products that are either available or publicly announced with an imminent release date. This presentation excludes non-public information and vague ‘it’s on the roadmap’ statements. It is seen that VVC is making its way into consumer electronics, ultimately facilitating broadcast services. In terms of OTT services, Intel announced VVC decode support in June 2024. These decoders are found in PCs sold by Dell, HP, Lenovo, LG and others, and have been available since September 2024. We note that in the past, the chicken-and-egg issue for codecs is generally resolved by decoder deployment first; as an important manufacturer with a large deployment base, this release could have a meaningful impact on the adoption of VVC for OTT.

Second, we consider software implementations of VVC that are currently available. While more related to OTT services, software implementations facilitate interchange of VVC content and in a sense ‘prime’ the ecosystem. A highlight of the past few months was VVC’s inclusion in the widely used FFMPEG software, which is a ‘Swiss army knife’ for media coding. Generating a VVC bit stream involves changing a few parameters – not the software itself – and therefore lowers the barrier to unlocking the coding efficiency gains that VVC provides.  While the performance of a software decoder centres on speed and power efficiency, the encoder design has a great bearing on the coding efficiency that can be achieved. FFMPEG’s VVC encoder is based on the already well-performing VVenC from Fraunhofer HHI, and as with any software implementation, there is the potential for further optimization.

Third, we review the ‘adoption’ of VVC by standards bodies since their doing so provides a good indicator of likely future deployment such as the expected launch of TV 3.0 system across Brazil in 2025.

As streaming services strive to deliver highly personalized experiences, the need for user data has never been greater. However, with growing concerns over privacy and data misuse, obtaining explicit subscriber consent has become paramount.

In this session, Sebastian will explore the delicate balance between leveraging user data for enhanced personalization and respecting consumer privacy preferences. Drawing from 24i’s recent work integrating FAST channels into set-top boxes, he will share insights on crafting seamless consent-gathering experiences that empower viewers to control their data sharing settings.

Combining privacy control with the best Cost Per Mille (CPM) and highest revenue requires some level of ad targeting. This means there is a requirement to share some user data with the FAST channel content owners and their technology suppliers, and this requires consumers to give their consent. This process is more complicated than a simple, one-off yes or no event because ad inventory can come from many vendors. Effective ad targeting, which includes refining audience segments based on demographics, interests, and behaviors, is crucial for maximizing revenue and fill rates.

Sebastian will discuss how deploying a robust consent management platform (CMP) quickly and easily enables video service providers to better monetize their channels with higher Cost Per Mille (CPM) and fill rates, while delivering a more personalized TV experience. It also allows customers to select what data they wish to share and with whom, providing advertisers with deeper insights to deliver more relevant ads.

Sebastian will also share best practices for transparent data usage policies, ensuring compliance with evolving GDPR, CCPA, and hundreds of other regulations while fostering trust among subscribers. Attendees will gain a deeper understanding of how to navigate the complex landscape of user consent, enabling them to unlock the full potential of personalization while upholding the highest standards of privacy protection.

IP network is a dark place regardless if you use private , public, cloud network. The It industry does have monitoring tools to monitor the network availability and general bandwidth use, but this information is not relevant for video delivery. A broadcast organization needs to be on top of the network characteristic ; Error rate, Jitter, Latency, RTT and additional performance KPI’s. And the IT team must have the right tools to find and correlate problems to Alarm reports by the broadcast team. This session will show a novel new approach to test the network, cloud and any IP based network to discover the available routes for the Video to take, continuous discovery of the network elements that are he key building blocks of the route. Each route then is probed to discover the important network characteristic in realtime and historical data. Using this method the user can pre test a network before going live, evaluate the network during a live session and have new visibility that old techniques like ETR 290 can’t fulfill.

The Common Media Application Format (CMAF) is a modern and versatile format with the potential to be used across all stages of video delivery—from contribution and distribution to final client delivery. However, in practice, CMAF is primarily used at the end-user stage. Expanding CMAF’s use throughout the entire video workflow required a reimagining of traditional video delivery processes. A significant milestone was reached in 2021 with the introduction of the Live Media Ingest Protocol, which positioned CMAF as a B2B mezzanine format between encoders and packagers, supported by a universal TCP-based communication framework.

Using CMAF for contribution or primary distribution, however, presented challenges as it didn’t meet all the performance needs of traditional MPEG-2 TS workflows, which require specialized encoders, modulators, decoders, and costly satellite transport. The goal was to create a modern alternative to the legacy MPEG-2 TS workflow. Ateme, our partner, has a solution involving using HLS (HTTP Live Streaming) as the contribution format, which can be generated by more cost-effective encoders, supports low latency, and carries CMAF fragments. For distribution, they utilized CDNs (Content Delivery Networks), offering greater flexibility, scalability, and affordability compared to satellite systems.

Once the workflow was validated, content protection became a focus, as TCP-based feeds are inherently more vulnerable than satellite transmissions. To address this, EZDRM and Ateme partnered to develop an entitlement mechanism that secures the HLS stream by delivering encryption keys to decoders as needed. We also implemented a control system that allows feed providers to enable or revoke decoders in real time.

To strengthen security, we added an additional layer. Here, encryption keys are further protected using a public key from the decoder, ensuring only the authorized decoder, with the corresponding private key, can access the content. We also implemented key rotation to achieve OTT protection comparable to BISS-CA.

By combining advanced technologies for encoding, packaging, encryption, transport, and secure entitlement management, we have created an innovative, future-ready video delivery workflow.

This paper explores the potential of hyper-personalized content in fantasy sports, presenting an in-depth analysis of various approaches to enhance fan engagement without advocating for a specific solution. It focuses on the complexities and opportunities arising from the convergence of technology and business in delivering tailored experiences that resonate with individual fans.

The new method described in the paper uses AI tagging and personalization to deliver video highlights tailored to fantasy teams. Unlike traditional methods, this approach removes the need for costly re-encoding, instead using advanced metadata tagging and dynamic playlist generation to produce custom experiences instantly. This approach offers a significant opportunity for brands to deepen fan engagement while managing operational costs effectively.

Existing AI-based approaches for highlight reels typically rely on creating highlight-specific assets, using AI to identify relevant content, and then encoding the content for delivery as a highlights on-demand asset. However, this requires a unique asset to be built, encoded, and stored for every such highlight reel, which can introduce delays and costs for processing and delivery. Creating specific highlight reels for each fantasy team would quickly become a scalability issue. Therefore, the approach outlined here is different; it is used to assign metadata tags to each play in multiple games because then customized highlight reels can be assembled on-demand using manifest personalization to select content based on the tags, rather than having to create new assets. The use of the same content improves cache hit probabilities in CDN delivery, improving the experience through better delivery. This will all be explored, showing how fans can receive highlight reels tailored exclusively to their fantasy rosters, helping to enhance their connection to their team.

The technical intricacies involved in delivering personalized slow-motion replays will be explored; pivotal moments shown in slow-motion enhance viewer experiences, but there are challenges related to timebase adjustments and playback compatibility.

The paper will explore how these can be managed without requiring non-standard high frame rate playback by client applications, delving into how consumer video players typically support standard frame rates, like 59.94 fps, but slow-motion cameras capture content at much higher rates. To avoid player incompatibilities and user complexity, slow-motion assets are stored as if already slowed to a fraction of the normal speed (e.g., 1/4th). This allows playback at standard frame rates, delivering smooth slow motion with intermediate frames intact. Using a personalization approach, viewers can request on-demand slow-motion replays, seamlessly integrated by adjusting manifests while maintaining compatibility with common player formats and capabilities. This integration ensures accessibility across platforms and devices, alongside smooth, slow-motion playback.

The paper also emphasizes the importance of conducting a thorough cost-benefit analysis for implementing AI-driven personalization at scale. It evaluates the infrastructure requirements, including server capacity, storage needs, and processing power, alongside potential returns on investment. Enhanced fan engagement metrics, such as watch time, repeat visits, and subscription retention are explored, providing a clear financial framework for media brands to explore.

Finally, this paper aims to provide a roadmap for utilizing AI-driven hyper-personalization effectively. Future developments in interactivity and immersion might include augmented reality overlays, predictive play highlights based on fantasy team performance, and enhanced fan-driven narrative experiences. The presentation will stress the importance of staying attuned to these trends and how media brands can strengthen their connection with audiences through hyper-personalized content. By aligning technology and business strategies, the opportunity is here to open new avenues for revenue generation and deliver competitive differentiation in a crowded marketplace.

The Reliable Internet Stream Transport (RIST) Activity Group in the Video Services Forum has been working on a common industry specification for methods to augment satellite delivery using the Internet.  The satellite is used for the “heavy lifting”, and the Internet is used to “fill in the gaps”.

Such a system can be further augmented by allowing the receivers to select which programs they require.  If this can be achieved, further Internet bandwidth optimization is possible, since the headend does not need to “fill in the gaps” for the programs that are not required for that receiver.  This method can be designed in a manner that is general enough to apply to other Internet transport scenarios, and not only to hybrid satellite delivery.

The RIST Activity Group has created a backward-compatible extension to RIST Main and Advanced Profiles to allow a receiver to communicate to the sender its content requirements.  Once the sender is made aware of these requirements, it can optimize the bandwidth without losing timing compliance by using the Null Packet Deletion functionality defined in RIST Main Profile.  This extension has been published by the Video Services Forum as TR-06-4 Part 6.

This presentation is a detailed overview of the technical details of TR-06-4, with some examples from actual broadcast feeds.  It also includes practical data on bandwidth saving from actual off-air ATSC content.

This presentation explores how generative AI is revolutionizing the integrated newsroom, enabling it to operate more efficiently, accurately, and innovatively. By leveraging AI-driven tools, newsrooms can automate content creation, streamline workflows, enhance accuracy, and deliver personalized experiences to their audiences. The integration of AI in news production allows journalists to focus on high-value tasks, while routine processes such as summarization, transcription, and content tagging are automated.

Generative AI plays a pivotal role in ensuring real-time reporting, assisting in fact-checking, and reducing human error, which is crucial in maintaining the credibility of modern journalism. It also enhances newsroom collaboration and productivity by automating repetitive tasks, enabling faster turnaround times from content creation to distribution.

In addition, the paper highlights the use of AI to analyze audience behavior, personalizing news delivery and driving greater engagement. Ethical considerations are also addressed, with AI ensuring adherence to journalistic standards through automated content moderation and bias detection.

Key Benefits of Generative AI in the Integrated Newsroom:

•  Increased Efficiency:

Automated Content Creation: AI automates repetitive tasks like article drafts, headline generation, and summarization, allowing journalists to focus on investigative and editorial work.
Faster Workflows: AI accelerates newsroom processes by handling transcriptions, video editing, and metadata tagging, reducing manual effort and speeding up the content cycle.

• Enhanced Accuracy:

Real-Time Fact-Checking: AI-driven tools verify facts and check for errors in real time, ensuring news is reliable and up-to-date.
Reduced Human Error: Automated error detection improves the accuracy of reports, minimizing the risk of misinformation.

• Improved Audience Engagement:

Personalized Content: AI analyzes audience data to deliver tailored news content, increasing engagement and retention.
Automated Recommendations: AI-based recommendation engines suggest relevant stories to readers based on their preferences.

• Cost and Time Savings:

Operational Efficiency: AI reduces manual labor, lowering operational costs and allowing smaller teams to manage larger workloads.
Scalability: Newsrooms can scale their content production capabilities without significant increases in resources.

• Ethical Journalism:

Content Moderation: AI flags inappropriate or biased content before publication, ensuring adherence to ethical journalism standards.
Bias Detection: AI tools can analyze articles for bias, helping maintain balanced reporting.

Generative AI is transforming the integrated newsroom into a more agile, responsive, and efficient entity, enabling news organizations to meet the demands of a fast-paced media landscape while upholding journalistic integrity.

Archival video collections contain a wealth of historical and cultural information. Managing and analyzing this data can be challenging due to the lack of metadata and inconsistent formatting across different sources. In particular, identifying and separating individual programs within a single archived tape is critical for efficient indexing, analysis and retrieval. However, manual segmentation is time-consuming and prone to human error. 

To address this challenge, we propose a novel approach that combines vision and language models to automatically detect slates and segment archive videos into distinct programs. A vision model is used to cluster frames of the video. Using recent robust automatic speech recognition and large language models, a transcript, a summary and a title for the program are generated. By leveraging computed features from the previous slates detection, we also propose a fine-grained chapterization of the segmented programs.

We conducted experiments on a dataset consisting of 50 hours of archival video footage. The results demonstrated a high level of accuracy in detecting slates and segmenting the videos into distinct programs. Specifically, we achieved a precision of 92% for an Intersection over Union threshold set at 90%.

Furthermore, our approach has shown to have significant sustainability benefits as it is able to filter and remove approximately 20% of the content from the 50 hours of videos tested. This reduction in the amount of data that needs to be managed, analyzed and stored can lead to substantial cost savings and environmental benefits by reducing energy consumption and carbon emissions associated with data processing and storage.

Live streaming has become increasingly important in our daily lives due to the growing demand for real-time content consumption. Traditional live video streaming typically relies on single-pass encoding due to its low latency. However, it lacks video content analysis, often resulting in inefficient compression and quality fluctuations during playback. Constant Rate Factor (CRF) encoding, a type of single-pass method, offers more consistent quality but suffers from unpredictable output bitrate, complicating bandwidth management. In contrast, multi-pass encoding improves compression efficiency through multiple passes. However, its added latency makes it unsuitable for live streaming. In this paper, we propose OTPS, an online two-pass encoding scheme that overcomes these limitations by employing fast feature extraction on a downscaled video representation and a gradient-boosting regression model to predict the optimal CRF for encoding. This approach provides consistent quality and efficient encoding while avoiding the latency introduced by traditional multi-pass techniques. Experimental results show that OTPS offers 3.7% higher compression efficiency than single-pass encoding and achieves up to 28.1% faster encoding than multi-pass modes. Compared to single-pass encoding, encoded videos using OTPS exhibit 5% less deviation from the target bitrate while delivering notably more consistent quality.  

As well known today, most streaming analytics systems infer the video quality and the overall QOE from video parameters observable at the client: encoded video resolution, framerate, codec type, and bitrate. Addition parameters typically include player size, buffering events, startup time, etc. However, such predictions cannot be very accurate. The bitrate correlates poorly with visual quality, especially with the proliferation of content-aware encoding techniques.

Some more advanced systems, such as those implementing the ITU-T P.1203 standard, derive encoding quality estimates by parsing the encoded bitstreams on the client side. However, such estimates are also not ideal. More accurate assessments are possible by using standard full-reference metrics, such as PSNR, SSIM, or VMAF. However, typically, such metrics are only available on the encoder side.

This presentation will discuss the design of a streaming and QOE analytics reporting system utilizing full-reference quality scores computed and channeled to the players by the encoder. The proposed system utilizes standards-based methods of signaling. The encoder embeds the observed full-reference metrics (PSNR, SSIM, VMAF) as timed metadata (ISO/IEC 23001-10) in MP4 files carrying encoded video streams. The client extracts and passes them to the analytics system as extensions of the CMCD-parameter parameter set. The analytics system then retrieves and uses them to compute the integral QOE scores. The integration logic follows the design of the ITU-T P.1203.3, accounting for full-reference codec distortion metrics as additional factors in the model.

To test the system, we use ITU-T P.1203 datasets. The results show that our proposed method achieves significantly better precision (15-20% lower RMSE) in predicting MOS scores than the reference P.1203 metric. The paper also discusses implementation complexity and costs associated with deploying and operating a mass-scale streaming system incorporating the proposed design.

The rapid adoption of Generative AI has introduced new threats to media integrity, particularly for broadcasters striving to maintain public trust. The ease of creating deepfakes and synthetic content has lowered the cost of manipulation, presenting a significant challenge for media professionals who need to maintain credibility in the face of advancing AI. 

In this session, Dr. Eric Wengrowski, Co-founder and CEO of Steg.AI, will introduce a powerful and proactive solution: steganographic watermarking. Steg.AI’s patented, deep-learning technology embeds imperceptible, durable watermarks into media, empowering broadcasters to authenticate their content after distribution and quickly detect manipulation. As a contributor to the Coalition for Content Provenance and Authenticity (C2PA) working group—a Linux Foundation group focused on establishing open standards for digital content provenance—Steg.AI is dedicated to supporting open-source tools for secure media verification and advocates for responsible AI practices. 

In this session, Eric will:

• Provide an overview of how generative AI is reshaping the media landscape and introducing new vulnerabilities, such as deepfakes and synthetic content, and what this means for broadcasters.
• Highlight the importance of ensuring media integrity and authenticity in an environment where manipulated content can proliferate. 
• Discuss how technologies like forensic watermarking can help broadcasters maintain secure, verifiable content credentials, where all content modifications are both transparent and traceable. 

Delivering exceptional video quality is critical in today’s highly competitive video streaming landscape. However, verifying video signals is a complex process. Video streams pass through various processes within media workflows, including pre-processing, transcoding, editing, post-processing, and more. The transformations that occur can alter the properties of the intermediate video, potentially disrupting subsequent processes or degrading the final output video quality experienced by viewers. Reliable content delivery necessitates the verification of video data to ensure an acceptable quality of experience (QoE). Video data verification is not limited to the measurement of degradation in perceived quality on the viewing screen. It should also consider validation of video parameters affecting the proper functioning of media devices. For example, the Y, Cb, and Cr values are altered after lossy compression, which can result in out-of-gamut RGB data and consequently lead to erroneous functioning of display devices. Similarly, parameters like light levels, black bar widths, color bar types, telecine patterns, photosensitive epilepsy (PSE) levels and patterns, field order, and scan types (interlaced or progressive) also need validation before distribution.

This presentation will explore critical video properties and demonstrate how computer vision, image processing, and machine learning (ML) techniques can measure and validate these properties and detect defects. Attendees will learn about the various experiments, conducted by utilizing ML techniques, to quantify the quality degradation resulting from lossy compression of video data. The presentation will also discuss the challenges and future directions media professionals can take to enhance the accuracy of video quality measurements and defect detection.

The increase in remote production has seen a commensurate demand for robust connectivity and compute resources on location. Verizon’s new Private Network Enterprise AI Compute Module utilizing NVIDIA GPU acceleration is built on top of a multi-band PWN, and brings cloud infrastructure on prem. It allows for flexibility to ingest multiple camera feeds and support the latest AI-enabled toolsets without sacrificing traditional tethered or in-studio capabilities.

This panel will explore how PNEAI is supporting an “Intelligent Video Prioritization for Live Production” use case at Verizon’s booth (W2530).

Hear from the experts who built this solution and learn how Private 5G coupled with AI can enhance production capabilities, increase efficiencies in existing workflows and reduce costs all while keeping the human touch.

This study used customized hardware and signal processing techniques to measure AM radio channel noise and signal strength on various road environments in Maryland, ranging from rural to densely populated urban areas. Unoccupied (“quiet”) channels and active local AM stations across the band were continuously monitored from a vehicle traveling along designated routes.

By recording both ambient radio frequency (RF) noise and the electric (E) field strength of actual AM stations, the study revealed a significant discrepancy between traditional predictions of AM radio groundwave field strength and the actual reception quality experienced by motorists. This discrepancy is particularly evident in urban areas where substantial signal loss occurs due to building clutter.

The results, presented in both graphical and colorized map formats, highlight the substantial impact on reception when real groundwave fields are combined with building clutter loss and RF noise, a condition overlooked in traditional station coverage maps. These findings underscore the importance of considering these combined factors in transmitter site planning.  It also points to potential benefits of deploying booster transmitters to improve coverage in areas with significant signal attenuation.  The project was supported by members of the NRSC’s AM Improvement Working Group.

As electric vehicles (EVs) proliferate worldwide, various wireless charging solutions are being developed which eliminate the need for using a wired connection for charging.  These so-called wireless power transfer (WPT) systems operate with high efficiency (over 80%) and can be installed for both stationary use (for example, in a garage or at a bus stop) or for use while the vehicle is in motion (by placing charging elements within a section of roadway).  NAB has been studying the possible effects of WPT on AM radio reception, both in the NAB Technology Lab and at a prototype WPT installation in Detroit, Michigan.  This presentation will provide background information on how WPT systems operate, including the possible mechanisms for creating RF interference in the AM radio band, as well as some of the test results obtained by NAB.

Public media has always pushed the boundaries of new broadcast technologies, from its pioneering work in network distribution via satellite in the 1980s, through early use of ATSC 1.0 multicasting in the 2000s, to today’s launch of ATSC 3.0 virtual or “Broadcast Enabled Streaming TV (BEST)” channels, and flash-cut ATSC 3.0 services via LPTV operations. One of the latter facilities is owned and operated by PMVG in Cookeville, TN (near Nashville), where it is being used as laboratory for development and trialing of new public media applications. Hear what’s happening on the cutting edge of NextGen TV from public media and other experts working on these innovative projects.

The HEBA (High Efficienty Broadband Antenna) has been operating daytime for close to a decade. Recent NEC modeling efforts were successful in proving the vertical radiation pattern. The FCC granted nighttime approval in Septenber, 2024. Because the HEBA’s 73-foot height makes it a “short antenna,” the FCC rules for nighttime approval are particularly difficult to satisfy. The HEBA also uses a novel two-port design to drive two active elements in the antenna system, while the antenna elements are stacked above a 40-foot square elevated ground plane. There are no ground radials. The HEBA is immune to seasonal tuning variations experienced by many conventional AM antenna systems. This paper will explore the antenna design, provide measurements and use NEC modeling to illustrate its performance.  

Natural disasters have long been a core use case for backup communication systems in broadcasting, providing crucial information to the public during emergencies. As cyber threats become increasingly common, the need for resilient, cyber-secure backup communications has grown. This paper explores the evolving use case for backup communication systems, focusing on the addition of a cybersecurity framework that enables continuous information relay even in the event of cyber disruptions.

System Overview:

The proposed backup communication system integrates traditional emergency broadcasting with a new cyber communications module, ensuring operational continuity in the face of physical and digital threats. The system is engineered to detect, mitigate, and recover from both natural and cyber events, effectively adapting to a dual-risk environment.

Use Cases:

• Natural Disasters – The system continues to cover established protocols for weather, earthquake, tornado, and flood warnings.
• Cyber Outages and Attacks – Enhanced cybersecurity measures allow this backup system to act as a failsafe when primary digital networks face cyber-attacks or technical failures, ensuring that broadcasters maintain connectivity and can deliver vital updates.

Results & Analysis:

This section provides insights into the system’s performance during simulated natural and cyber events. Testing shows high resilience, with minimal latency and data integrity preserved in both types of emergencies. Cyber use case scenarios demonstrate improved response times and secure recovery processes, enhancing the overall reliability of broadcast communications.

Conclusion:

By addressing both natural and cyber threats, this new generation of backup communication systems presents a comprehensive solution for broadcasters. This paper advocates for the adoption of such dual-use systems to maintain public trust and ensure information availability during crises, highlighting key design principles and operational benefits.

This summary highlights the innovative expansion of a traditional emergency system to meet the demands of the current cybersecurity landscape, emphasizing practical and technical advancements.

As the media industry rapidly adopts Content Authenticity Initiative (CAI) and C2PA standards, sophisticated actors are already finding ways to circumvent these protections. This critical session explores the emerging threats to broadcast integrity that exist beyond the scope of current content authentication frameworks.

Led by Rijul Gupta, Deep Media’s founder and CEO—recently called to testify before the Senate Committee on Privacy, Technology, and the Law—this session reveals how bad actors are leveraging cutting-edge AI techniques to create increasingly convincing deepfakes that bypass traditional detection methods.

Drawing from Deep Media’s extensive experience protecting major broadcasters and government agencies, Gupta will demonstrate:

• Real-world examples of AI-generated content that successfully circumvented C2PA protections
• Technical analysis of emerging attack vectors in the broadcast space
• How newsrooms can implement robust, multi-layered detection strategies
• The critical role of AI-powered real-time detection in maintaining broadcast integrity

As the only deepfake detection solution with validated 95%+ accuracy rates and deployments across major media organizations, Deep Media offers unique insights into protecting broadcast content in an era of rapidly evolving AI capabilities.

This session is essential for:

• Broadcast security professionals
• News directors and producers
• Media technology leaders
• Content authentication specialists
• Digital security teams

Join us for an eye-opening look at the future of broadcast security and learn how leading organizations are staying ahead of emerging threats. Discover why the world’s most trusted media organizations rely on Deep Media’s advanced AI detection suite to protect their content integrity and maintain viewer trust.

This presentation will include live demonstrations of Deepfake Attacks.

The live media production industry is on a slow but inevitable journey to a software-based workflow. The industry has a proud track record of open standards for signal hand-off in live production workflows, evolving from analogue composite video to SDI to 2022-6 then 2110, as well as the significant compression options of MPEG-TS in ASI then 2022-2 and SRT etc. While these standards served linear workflows well, they present inefficiencies in compute-driven environments.

Compute workflows operate in an inherently bursty manner and the traditional linear ‘hosepipe’ broadcast signal is not an efficient one for compute to conform to for the reception and handoff of data flows. The strict specification for timing conformity in ST 2110-21 was created to ensure the integrity of transport of inherently fragile RTP/UDP packets to ensure the streams were linear enough to not result in corruption and packet loss in the aggregation of flows in switch fabrics. There is an unnecessary and significant compute and latency cost in conforming to linear UDP flows, so a compute-native exchange approach is needed and there are initiatives underway in the industry to address this.

There are also associated challenges with compute-native non-linear data exchange in the timing information pertaining to each of the media ‘chunks’ being transported. In a non-linear world, it is imperative that the timing information (and any other related metadata) of each piece of media data is captured at or as close as possible to source and then retained with this data through the processing chain.

This session delves into the challenges and opportunities of transitioning to a non-linear, compute-friendly data exchange framework. Key focus areas include:

Addressing timing and metadata challenges in non-linear data exchange.
Preserving source timing information throughout the processing chain.

Evaluating existing open proposals for compute-native media exchange.

Sharing results from performance testing, with insights into bandwidth usage, processor demand, and latency impact.

The presentation aims to outline critical requirements for live production media exchange, compare current proposals, and, if possible, feature a live demonstration of the interface in action, depending on in-room connectivity and technical setup.

Ensuring tower site safety is crucial for preventing accidents and protecting workers. Here are some key areas we will discuss in this presentation:

1. Introduction

Purpose: To ensure the safety of personnel and the security of the tower site.
Scope: Applicable to all employees, contractors, and visitors at the tower site.

       2. General Safety Requirements

Personal Protective Equipment (PPE): Hard hats, safety vests, and other necessary PPE must be worn at all times .
Training: Only trained and authorized personnel are allowed to work on the site .
Signage: Display all necessary safety signs and permits in a conspicuous location .

       3. Site Access and Security

Access Control: Secure site access with perimeter fencing and gates .
Signage: Warning signs against trespassing and indicating health and safety hazards .
Surveillance: Use CCTV systems to monitor the site .

       4. RF Hazard Management

RF Monitoring: Evaluate the site for RF hazards upon arrival using a personal RF monitor 
Exposure Limits: Do not remain in areas with RF levels at or above 100% of the maximum permissible exposure limit .
Coordination: Coordinate power-down procedures with the broadcast antenna owner if RF levels are unsafe .

       5. Fall Protection

Safety Climb Systems: Ensure all towers over 10 feet have a safety climb system .
Tie-Off Requirements: Maintain 100% tie-off while climbing or navigating obstructions .
Tagout Procedures: Follow proper procedures for tagging out unsafe safety climb systems 

       6. Emergency Procedures

Emergency Contacts: Provide names, addresses, and telephone numbers for emergency contacts .
Incident Reporting: Notify the ATC Construction Manager immediately in case of a reportable safety incident or accident .
First Aid: Ensure availability of first aid kits and trained personnel 13 .

       7. Tower Fall/Collapse Procedures

Initial Response: Dispatch to the site immediately upon notification of a tower fall .
Site Assessment: Tape off the area, assess for EH&S hazards, and confirm the tower fall .
Safety Measures: If unsafe to access, provide a statement to NOC and alert local authorities.

       8. Hazardous Materials

Hazard Identification: Cease work and notify the ATC Construction Manager if hazardous materials are discovered.
Asbestos Management: Follow specific procedures for handling and reporting asbestos .

       9. Site Maintenance and Inspections

Regular Inspections: Conduct regular inspections of the site for safety and compliance .
Maintenance: Ensure all safety systems and equipment are properly maintained and functional .

        10. Conclusion

Safety Commitment: Emphasize the commitment to safety and the importance of following all procedures .
Continuous Improvement: Encourage feedback and continuous improvement of safety practices .

As Olympic Broadcasting Services (OBS), a company established by the International Olympic Committee (IOC) to eliminate the need to continually rebuild the broadcast operation for each edition, we produce the live television, radio and digital coverage of the Olympic and Paralympic Games. Our coverage is neutral, favouring no particular country or athlete, and includes sports competitions as well as the Opening and Closing Ceremonies.

Not only do we deliver our coverage to broadcasters around the globe, we also design, build and operate the International Broadcast Centre (IBC). We help broadcasters personalise their coverage for their home audiences and represent the needs of the Media Rights-Holders (MRHs)

We ensure the high standards of Olympic broadcasting are consistently maintained from one edition of the Games to the next. As part of this mission, we make use of the latest state-of-the-art broadcasting technologies to continually improve the level of services that define television quality standards and inspire future innovations.

One of the most captivating evolutions of the past years, is the move towards Ultra-High-Definition (UHD) and High-Dynamic-Range (HDR) image-capture and -display. Our goal to provide the best possible quality to the broadcasters, adopting and embracing this transition was a given fact, although highly challenging at the scale of the Olympics. With over one thousand broadcast system cameras deployed during a single Olympic Games, we were put up against a mammoth task to transform our existing Vision Engineering Workflow.

During the past decade, we have collaborated with multiple key MRHs, manufacturers and trusted partners worldwide, through dozens of research-sessions, lab-days & test-events to develop what we know today as the “OBS Single-Layer HDR Workflow”. This unique approach has provided more than fifteen thousand hours of native UHD HDR live multilateral footage so far, distributed and viewed by millions of people around the globe through the distribution channels of the broadcasters.

An important part of the HDR Workflow, was the creation and adoption of the OBS Look-Up-Table (LUT) Package, which provides conversions between the supported transfer-functions and colour-spaces. Through several generations of the LUT-Package, we provide determanistic and neutral tone-mapping with presteigne results both in the HDR- as well as the Standard-Dynamic-Range (SDR)-domain, without compromising on image quality. The early adoption of parametric replication created by OBS, paved the way to what is called “Closed-Loop-Shading” in the broadcasting industry today.

At the heart of the Vision Engineering Workflow, sits the Vision Quality Control (VQC) area at the IBC. A team of expert Vision Engineers monitor up to 80 incoming feeds in both HDR and SDR simultaneously and are in direct contact with the Vision in Charge (VIC) positions in each Outside-Broadcast Vehicle. A controlled viewing environment, calibrated grade-one displays and agile monitoring tools create the perfect conditions to guarantee all feeds are aligned and consistent.

With multiple hundred Vision Engineers (also known as Vision Operators, Rack Engineers or Shaders) joining us from all over the world for a limited amount of time in the host cities, implementing an educational program was key to the success of the UHD HDR transition. Through different formats, all Games-time participants were able to get access to both in-presence as well as online material to bring them up to speed on the details of the Workflow.

With this presentation, OBS will provide the global audience a unique and insightful behind-the-scenes of the Vision Engineering Workflow of the Host Broadcast of the Olympic Games. The Paper will present you with more details, diagrams, numbers and schematics that have never been shared publically before. OBS remains a neutral, unbiased organisation that serves the broadcasters, as well as the athletes.

This presentation builds on previous work related to Dynamic Ad Insertion (DAI) within the ATSC 3.0 framework using MPEG Media Transport (MMT). The current study extends this research by developing a robust broadcast-only transmission solution within encoder/packager systems to enable targeted ad delivery. Leveraging broadcaster applications, NRT data distribution, DASH EventStream technologies, the implementation supports precise control over ad breaks and allows for real-time event tracking. Additionally, this architecture integrates with mobile devices featuring local application storage, allowing for preloading and personalized selection of advertisements based on individual user preferences and service viewing history. Designed with the India D2M (Direct-to-Mobile) Project in mind, this paper examines the unique requirements and solutions developed to address an environment in which all ad distribution and decision logic must be completed in a broadcast-only system, with no guarantee of broadband access. 

5G New Radio (NR) can be used to provide flexible, high-capacity and low-latency networks suitable for broadcast content acquisition or delivery, but access to suitable spectrum can be challenging. One of the enablers for private network deployments is shared spectrum licensing, such as the upper n77 band (3.8 – 4.2 GHz) available in the UK and elsewhere in Europe.

The Third Generation Partnership Project (3GPP) was created to develop mobile standards for WCDMA and TD-SCDMA and their respective core networks, and has continued to publish standards as radio access technologies have progressed to 4G and 5G. These standards define frequency bands, numerologies, duplex models and messaging (among many other things). While software-defined radio (SDR) is emerging as a viable and highly flexible solution for core and radio access network (RAN) functions, user equipment (UE) typically remain hardware based with modems that implement the 3GPP standards to ensure device compatibility. The flexibility of SDR RAN allows for wireless radio networks based on 5GNR to be built in non-3GPP defined spectrum bands, but there are no compatible devices to connect.

In the USA, broadcasters have access to spectrum in the Broadcast Auxiliary Service (BAS) band (2025 – 2110 MHz), which coincides with the programme-making and special events (PMSE) band used in the UK and Europe. This allows for rapid licensing of 10/12 MHz channels for traditional wireless camera systems, such as COFDM, that could instead be used to licence low-to-medium power private 5GNR-based networks capable of supporting multiple cameras and other IP-based workflows.

This presentation discusses the development of a flexible software-defined UE capable of connecting to non-3GPP 5GNR networks in BAS/PMSE spectrum

In the recent years, 5G Broadcast has emerged as an attractive broadcasting technology that can enable broadcasters to deliver content directly to mobile devices. This is facilitated by the cellular modem-friendly design of 5G Broadcast, which leverages key building blocks of 3GPP-compliant cellular modems—such as those present in our smartphones—to facilitate broadcast reception without requiring additional dedicated hardware for this purpose.

While 5G Broadcast presents a unique opportunity for broadcasters to expand their reach to smartphones, certain key bridges between the world of traditional broadcasting and that of cellular systems need to be built along the way, to take full advantage of this opportunity. One such bridge, that we address in this paper, is to enable 5G broadcast—a technology built on the LTE air-interface backbone—to be seamlessly deployed in the frequency bands (with associated channelization) designated for broadcast systems in varied geographies across the world.

Since 5G Broadcast targets broadcasters—as opposed to cellular operators—the most important target spectrum for this technology is the portion of UHF spectrum allocated to broadcast systems (~470 – ~694/698MHz, depending on the region). The channelization used within this spectrum can vary between 6, 7 and 8 MHz in bandwidth, depending on the geography (such as those defined by different ITU regions).

While 3GPP specifications in Release 14 and 16, introduced features for 5G Broadcast in the Radio Access Network (RAN), which—together with the corresponding enhancements in the core network—enabled the deployment of a downlink-only dedicated broadcast network using 3GPP technologies, the channelization was inherited from cellular LTE—i.e., only bandwidths of 1.4, 3, 5, 10, 15 or 20 MHz were supported. As a result, Release 16 specifications for 5G Broadcast did not natively support a seamless deployment on UHF spectrum for broadcast systems that use 6, 7 or 8 MHz channel bandwidths.

To address the above issue, in Release 17, 3GPP set out to adapt 5G Broadcast to the UHF spectrum, while minimizing the changes required to the specifications in the process. In this paper, we describe how this was accomplished.

The design leverages the fact that there are LTE bandwidths smaller than the broadcast UHF system bandwidths—namely, 3 and 5 MHz—that can be embedded centrally, within a larger 6, 7 or 8 MHz UHF channel, and used for the transmission of “always-on” signals (i.e., the Cell Acquisition Subframes (CAS), comprising synchronization signals, the Master Information Block (MIB) and the System Information Block (SIB)), associated with the MBSFN Area(s) served by the UHF channel. Such an embedding ensures that no changes need to be made to the synchronization signals and MIB (which occupy a central bandwidth of 1.08 MHz) designs, both from a transmission as well as a reception point-of-view.

Once the User Equipment (UE) decodes the MIB, it determines from the MIB payload, whether a 3 MHz or a 5 MHz bandwidth is employed for the CAS, which enables the UE to decode the SIB within the CAS. The SIB, in turn, indicates to the UE whether there exist MBSFN Area(s) with a larger bandwidth than that of the CAS—i.e., a 6, 7 or 8 MHz bandwidth—over which the UE can expect to receive the Physical Multicast Channel (PMCH) that carries the eventual broadcast content.

Subsequently, in Release 18, broadcast UHF bands (with band numbers characterized by the carrier frequency and associated channel bandwidths) were added to the 3GPP RAN4 specifications. This, in conjunction with the Release 17 work described above, allowed broadcasters to deploy 5G Broadcast on UHF bands with 6, 7 or 8 MHz channel bandwidths.

Jonny McClintock and Dr Dean Armstrong of Virscient Ltd present a paper on audio transmission over Bluetooth, and discuss how ultra-low latency can be achieved  using standards-based transceivers and low-power links.

While the definition of “low-enough” latency varies from performer to performer, it is universally accepted that lower is better, and crystal-clear that solutions like the classic Bluetooth A2DP simply aren’t good enough.

The discussion begins with an explanation of the limitations Bluetooth Classic, the vast improvements delivered with Bluetooth LE Audio, and the current situation for live performers – we’ll take a look at some research from Audio Engineering Society around the different levels of latency experienced by different types of musicians.

We will then investigate the audio chant to understand the causes of latency in wireless audio transmission before exploring how these issues can potentially be resolved.

Dean and Jonn will then demonstrate a proprietary solution called LiveOnAir which combines well known codecs such as Skylark with the LiveOnAir protocol stack to deliver an analogue-to-analogue latency below 3.5 ms, or the mor. As an alternative, use of the LC3+ codec makes some latency sacrifice for significantly reduced MIPS and increased bandwidth flexibility. With this option, links of down to 10 ms analogue-to-analogue latency are achievable, and this is the configuration that will be demonstrated LIVE at this session.  

To wrap up, we will unpack a deeper layer of technical information around the LiveOnAir solution: its wireless microphone reference design is based entirely on Nordic Semiconductor’s nRF5340 BLE SoC, which runs both the protocol stack and codecs.

The nRF5340 is an all-in-one SoC that includes a superset of the most prominent nRF52 Series features. Features like Bluetooth 5.2, high-speed SPI, QSPI, USB, up to 105 °C operating temperature, and more, are combined with more performance, memory and integration, while minimizing current consumption.

This system is a turn-key reference solution for vendors looking to design audio transport products with analogue-to-analogue latency below that offered by Bluetooth LE Audio currently. It will be of particular interest to users of MIDI and WIDI systems for music content creation, along with designers and manufacturers of wireless microphone equipment for recording and live performance.

With the rise of digital video services, viewers expect high-quality visuals, making Quality of Experience (QoE) a priority for providers. However, poor video processing can degrade visual quality, leading to detail loss and visible artifacts. Thus, accurately measuring perceptual quality is essential for monitoring QoE in digital video services. 

While viewer opinions are the most reliable measure of video quality, subjective testing is impractical due to its time, cost, and logistical demands. As a result, objective video quality metrics are commonly used to assess perceived quality. These models evaluate a distorted video and predict how viewers might perceive its quality. Metrics that compare the distorted video to the original source, known as full-reference (FR) metrics, are regarded as the most accurate approach. 

Traditional quality metrics like Sum of Absolute Differences (SAD), Sum of Squared Differences (SSD), and Peak Signal-to-Noise Ratio (PSNR) are computationally lightweight and commonly used within encoders for Video Quality Measurement (VQM) and other encoder optimization tasks. However, methods that simply measure pixel-wise differences often lack alignment with human perception as they do not account for the complex intricacies of the Human Visual System (HVS). 

In recent years, more advanced metrics have been developed to better reflect human perception by incorporating HVS characteristics. Among these, Video Multi-method Assessment Fusion (VMAF) has become a widely accepted industry standard for evaluating video quality due to its high correlation with subjective opinions. 

However, the high computational demand of VMAF and similar perception-based metrics limits their suitability for real-time VQM. Consequently, encoders primarily offer only PSNR and Structural Similarity Index Measure (SSIM) for full-frame quality monitoring during encoding. While not the most accurate, these metrics are the only options that can be efficiently deployed during live encoding, as more advanced VQM approaches would consume too much processing resources needed for real-time encoding. 

To address these limitations, we introduced predictive VMAF (pVMAF), a novel video quality metric that achieves similar predictive accuracy to VMAF at a fraction of the computational cost, making it suitable for real-time applications. 

pVMAF relies on three categories of low-complexity features: (i) bitstream features, (ii) pixel features, and (iii) elementary metrics. Bitstream features include encoding parameters like the quantization parameter (QP), which provide insights into compression. Pixel features are computed on either the original or reconstructed frames to capture video attributes relevant to human perception, such as blurriness and motion. Finally, elementary metrics, such as PSNR, contribute additional distortion information.  

These features are extracted during encoding and fed into a regression model that predicts frame-by-frame VMAF scores. Our regression model, a shallow feed-forward neural network, is trained to replicate VMAF scores based on these input features.  

Initially designed for H.264/AVC, we extended pVMAF’s applicability to more recent compression standards such as HEVC, AV1, and VVC. In this presentation, we will explain how we developed and retrained pVMAF for different video coding standards. Also, we will provide details on the open-source pVMAF code which is now available for x264 and SVT-AV1.  

Experimental results indicate that pVMAF effectively replicates VMAF predictions with high accuracy while maintaining high computational efficiency, making it well-suited for real-time quality measurement. 

There is lots of talk about Dynamic Media facilities lately, but what does it all mean? Or what should it mean? Lets discuss the how the Industry should design these requirements with lessons learned from Cloud and AI, and how we can push for an Open, Secure, and Interoperable future. 

As the media and entertainment industry embraces cloud-based and elastic infrastructures, both on premise and inside public hyperscalers, new cybersecurity and operational challenges emerge. This shift promises unparalleled scalability and flexibility, yet also introduces significant vulnerabilities that must be carefully managed. The elasticity of cloud infrastructure allows for dynamic scaling, which, while advantageous for handling fluctuating demands, brings heightened risks, including increased attack surfaces and the complexities of managing tenant isolation and data segregation in multi-tenant environments. Traditional security models often prove insufficient in addressing these new concerns, as elastic systems respond dynamically to both genuine demand and potential threats.

This presentation explores the primary risks associated with elastic infrastructures which can exploit scaling features to trigger lateral expansion, resulting in unanticipated operational costs and potential service disruptions. A key focus will be on securing connectivity and controlling “blast radius,” minimizing the scope of impact from potential breaches by implementing strict isolation and containment strategies within and between tenants. Additionally, we discuss the unique requirements for effective incident detection and response within an elastic environment, emphasizing the importance of fast response times, thorough auditing, and continuous monitoring.

Elastic architectures demand new approaches to cybersecurity that adapt to their inherent dynamism. Traditional installations relied on physical segregation and infrequent updates, yet cloud-based systems require a mindset shift toward anticipating inevitable failures and attacks. We advocate for a comprehensive security framework including supplier assessments, and adopting best practices from industries like fintech and e-commerce, which have pioneered robust responses to cloud vulnerabilities. By focusing on modular incident response training, playbook development, and automated alert systems, companies can protect themselves in a landscape where threats evolve as rapidly as the infrastructure.

Ultimately, “The Liability of Elasticity” calls for a new cybersecurity paradigm, one that balances elasticity with robust security protocols, tenant isolation, and rapid response mechanisms. We aim to provide a blueprint for broadcasters and media companies navigating the challenges and opportunities of elastic cloud environments. This includes actionable insights into reducing exposure to cross-tenant risks, developing strong governance over dynamically deployed assets, and ensuring that security measures scale in parallel with elasticity, safeguarding both operations and reputations.

Scaling video delivery is a major challenge for content providers. Content Delivery Networks (CDNs) help overcome this challenge by positioning replicated content near end-users through distributed edge cache servers. However, as traffic surges across multiple networks and regions, a single CDN cannot sustain the demand. To combat this occurrence, content providers are adopting multi-CDN strategies. 

The multi-CDN approach varies in complexity depending on delivery and business requirements. Today, the content steering standard enables the orchestration of multiple CDNs, allowing seamless mid-stream CDN switching and helping to ensure uninterrupted playback. Mid-stream switching is particularly useful in case of local outage at a CDN. Once the performance degradation of a CDN is detected, for example by analyzing the latest Common-Media-Client-Data (CMCD) logs or by external Quality of Experience (QoE) services (such as Mux, Conviva, Bitmovin Analytics), the content steering balancing algorithm seamlessly switches end-users to a “healthy” CDN. Though, this is not a fail-proof solution. We still must ask ourselves: How can we detect that the failed CDN has recovered and when we should reintegrate it into the multi-CDN portfolio? 

Various content steering studies have addressed the question of reducing the time between the CDN failing time and the time at which the client is switched to another CDN. This paper will focus on how to reduce the time between the end of the CDN failure (trecovery) and the time at which clients are switched back to the CDN. Both HTTP Live Streaming (HLS) and Dynamic Adaptive Streaming over HTTP (DASH) content steering specifications instruct the normative behavior of video clients. If the steering server indicates two CDNs in its manifest, then the client must use the first one and switch to the second, only in case of an error (HTTP 404 code). Yet, when a CDN fails, while service may continue, it does so at a very low rate, which impacts the QoE. 

In this paper we will introduce and compare four approaches to detect CDN resurrection: 

1. Sacrificing some clients: the most often used approach in the industry is to drive a fraction of the impacted population (e.g., 5 %) to the failed CDN. 
2. Probing: these mock clients continuously fetch video segments from the failed CDN and generate valuable reports on the performance they experienced during this download. 
3. Mock player requests: initiating mock requests directly from the players at set time intervals can address the limitations of the probes.  
4. Round-robin sacrifice: the content steering sacrifices end-users in a round-robin fashion by allocating the failed CDN to some end-users (preferably among those who report having their buffer filled) for a limited time only. 

Popular FAST services use manifest-driven streaming delivery where revenues are derived from ads using ad insertion/replacement methods which can vary from simple techniques to more complex and often proprietary approaches. It is a high-volume service, but its ARPU is very low compared with other streaming models. To increase revenues, FAST services need to expand to additional streaming platforms in an interoperable manner and utilize more sophisticated ad strategies. SCTE has a set of Technical and Engineering Emmy award-winning technologies that can provide scalable deployment of these services through interoperable optimizations between content providers and distributors. SCTE technologies can also be used to provide scheduling to create the entire FAST channel, such that multiple manifests can be simultaneously generated for each distribution platform originating from a single instruction set stream using SCTE 224 Event Scheduling and Notification Interface (ESNI) Viewing Policy and Audience filtering constructs. This approach can integrate with existing manifest manipulators and server-side ad insertion systems while reducing its amount of decisioning to support syndication and more complicated ad techniques.

The media and entertainment industry has undergone a significant shift towards hybrid and cloud-based workflows. As organizations leverage the growing ecosystem of Software as a Service (SaaS) solutions and multi-vendor integrations, new security challenges have emerged. These hybrid environments, which often involve on-premises storage and processing coupled with cloud-based services and remote users, present potential compromise vectors for sensitive data and assets.

This presentation explores practical strategies to address security threats and safeguard intellectual property, pre-release assets, and other critical resources within hybrid media and entertainment workflows. It will examine the diverse threat vectors malicious actors can exploit, with a focus on vulnerabilities introduced by SaaS solutions and the integration of bespoke vendor offerings.

Drawing on real-world use cases, we present a comprehensive overview of the security landscape, addressing vulnerabilities at the network, application, and user levels. The presentation discusses common IT solutions for reactive monitoring and detection, as well as emerging methods for more proactive anomaly detection. This includes the use of tools for secure communication among SaaS providers and the application of behavioral analysis to identify deviations from normal usage patterns.

The presentation discusses the challenges in applying zero-trust principles to protect media and entertainment environments. It provides practical guidance on mitigating risks such as credential reuse, lateral movement, and suspicious user behavior, providing readers with the necessary tools and resources to safeguard their organizations.

By providing guidance and insights, the paper aims to assist media and entertainment professionals in navigating the evolving security landscape and exploring effective measures to protect their intellectual property, pre-release assets, and other mission-critical resources. The in-depth examination of real-world use cases and the discussion of tangible solutions will help the media and entertainment companies address their security concerns and stay informed about emerging threats.

In today’s fast-paced broadcast and media landscape, effective team collaboration and efficient content production are essential for staying competitive. This session presents an innovative approach to teamwork using avatar technology paired with advanced content production and management tools.

Picture a project where not all team members can attend a meeting or stand-up. Avatars step in as stand-ins, providing insights and sharing essential updates to ensure your team stays aligned, productive, and on track to meet deadlines. These avatars capture key insights and perspectives from past meetings, as well as content from documents and other materials, allowing absent colleagues to stay connected and contribute meaningfully. This way, team members can seamlessly pick up from where they left off without needing lengthy catch-up sessions that drain time. Avatars do more than simply represent team members; they actively contribute by presenting client updates, sharing recorded notes, and participating in discussions using pre-recorded or synthesized responses. For example, sales teams can rely on avatars to deliver client presentations or updates during periods when key team members are unavailable, ensuring the flow of communication and client satisfaction. Marketing professionals can use avatars to share strategic insights or content plans during meetings without needing everyone present.

Our approach bridges communication gaps and supports smoother decision-making by providing context and insight from previous discussions. The result is teams that remain aligned and productive even when schedules cannot perfectly match up.

To enable avatars, we have put together and showcase several intuitive tools for creating and sharing content more efficiently. Teams can seamlessly clip, merge, and subtitle videos without the need for specialized skills. These tools support multiple languages enabling broadcasters to adapt content for diverse audiences worldwide. Our product offerings a game changer for teams looking to expand their reach and connect with a global audience. Streamlined editing processes save valuable time and help teams deliver polished, localized content very fast.

Avatar-driven collaboration and user-friendly content tools reduce production time, boost engagement, and ensure a smooth workflow. Teams no longer need to wait for the right version of content or for approval. Since there are no post-production delays; production teams can keep pace with growing industry demands and deliver consistently impactful content.

A large variety of complex, novel algorithms have been implemented to interpret the content in any form that needs to be understood to make the avatars function. Given any context, the relevant parts of the knowledge created by processing data is retrieved and then shown in the most appropriate form including audio, video, images as well as plain text. Since our algorithms are flexible and customizable, production teams can adapt the avatars to suit their specific needs. Our avatars and content production tools are suitable for everything from small-scale projects to highly intricate broadcast schedules.

Our innovations align with current trends in the broadcast industry, focusing on next-generation tools that support content creation and delivery from start to finish. Leveraging avatars for collaboration and versatile content tools ensures better communication, fewer delays, and more engaging content for audiences. It’s about working smarter and making better and smarter use of technology an integral part of the creative process.

The goal is clear: empower teams to be more productive, create exceptional content, and make collaboration seamless, even when everyone isn’t always available. Our transformative approach enables broadcasters and media professionals to embrace more efficient workflows and produce content that resonates with audiences everywhere. It represents a big step toward a more connected, responsive, and dynamic future in media production.

Volumetric Capture and Reconstruction are emerging technologies that are rapidly gaining interest for their implementation in production workflows. The ability to digitise, reconstruct and manipulate 3D representations of real environments promises to reshape the production landscape as crews are able to use these digital assets in a variety of new and exciting ways. These range from the implementation of revolutionary new previsualisation methodologies, placing the movements of performers from rehearsals into a simulation of the real space to get an understanding of how they’ll interact with complex sets and props, to reducing the need to reshoot scenes that need new camera angles, recolouring or relighting by editing the digital assets instead— cutting costs and saving resources on a production.

One of the potential roads to this technology is depth estimation. By utilising groundbreaking machine learning algorithms, researchers can attempt to allow machines to determine how far the subject of a shot is from the camera, returning a “depth map” that tracks the relative distances that objects appear to be from the perspective of the observer. Feeding that depth map into a rendering engine allows for the creation of “2.5D Assets”, or a stack of two dimensional elements that give the illusion of depth when viewed from certain angles. Adding additional cameras viewing the scene from different positions and merging the resulting depth maps will eventually allow for the transition from pseudo-3D to the real thing. In turn, this allows for the capture and digitisation of in-camera elements for future use in whatever manner a production requires.

Since late 2023, Disguise has been investigating ways to bring this workflow to our flagship media platform Designer. As active participants in pan-European consortium EMERALD and the joint funded UK/Switzerland project Fast Volumetric Reconstruction (FaVoRe), and building off work begun in the MAX-R research initiative, our engineering teams have been exploring techniques to bring depth estimation to a production, working with partners at multiple European academic and business institutions to determine the best ways to use this new technology in a production environment. This presentation will cover the development work done so far to support this integration, initial findings about what the potential limitations of its use might be in different production formats and where the team plans to go next in order to deliver a full Volumetric Capture and Reconstruction solution. 

Broadcasters and video providers are continuously looking for ways to improve the viewing experience, and when it comes to live TV production, this means improving quality and reducing latency of video streams. 

Latency is a particular issue in live playout because the time between the content being captured on camera and delivery to the end user can make or break the viewing experience. Every second counts in live TV, whether news reporting on breaking events or broadcasting a sports match or race.

JPEG XS is a relatively new video compression standard designed to provide visually lossless quality at very low compression ratios. Unlike other formats, which focus on achieving maximum compression, JPEG-XS prioritizes maintaining image fidelity while offering incredibly low latency. This makes it particularly suited for applications where speed and visual quality are critical. 

It’s hardly surprising then that JPEG-XS is fast becoming the go-to choice for live production environments. And with more broadcasters adopting JPEG-XS-enabled production equipment to handle the demands of modern workflows, especially for live premium sports or news broadcasting, there’s a clear need for cloud playout systems to keep pace with this evolution.

From cameras to switchers, the ability to compress and decompress video signals in real-time is transforming production environments. In particular, there are a number of key advantages for adding JPEG XS compression technology to live cloud playout platforms. These include:

• Near-Zero Latency – latency is a critical factor in live TV delivery because viewers expect what they see on their screen to be in real-time. JPEG XS ensures near-instantaneous transmission with minimal delay.
• High-Quality Visuals – viewers expect crystal-clear picture quality. JPEG XS offers visually lossless compression that’s essential for high-definition and ultra-high-definition content, all without sacrificing speed.
• Bandwidth Efficiency – as broadcasters move toward more bandwidth-intensive content like 4K and 8K, JPEG XS strikes the right balance ensuring excellent image quality while optimizing the amount of data transmitted.
• Flexibility and Scalability – JPEG XS can be used across a range of resolutions and frame rates, making it versatile for different production and distribution needs.

Despite the growing momentum behind JPEG-XS, few cloud playout providers offer support for the format. This session will detail the need for playout systems that can manage JPEG XS encoded content without creating bottlenecks. It will demonstrate through real-world examples how broadcasters can overcome live production challenges and take advantage of both JPEG XS technology and the flexibility of cloud playout systems and enable real-time, high-quality video stream processing with minimal latency and reduced operational costs. 

The advent of Video Super Resolution (VSR) has catalyzed a shift within the media industry, enabling the broadcast of UHD / 4K content without the costs associated with upgrading production equipment, and facilitating the restoration [KC1] of archival footage providing cost-effective methodologies to support Free Ad-Supported Streaming TV (FAST) channels.

However, the integration of VSR into the broadcasting industry requires a high-quality solution that maintains consistent quality across diverse broadcast standards, handles live content with minimal latency, and ensures compatibility with existing broadcast infrastructure. Addressing these issues necessitates sophisticated AI models to deliver high visual quality, which often demands specialized computational resources, such as high-performance GPUs. This in turn increases the cost per stream, and challenges the economic feasibility.

Broadcasters are looking into the advancements of developing efficient VSR models that maintain high fidelity upscaling on a broad spectrum of hardware platforms while maintaining real time performance for low latency media applications.

Intel is working on a number of models available under the open source Intel(r) Library for Video Super Resolution, including Enhanced RAISR[1] deployed by AWS[2] for FAST applications and Enhanced BasicVSR[3] for high quality broadcasting. The latest addition is the TSENet model (Temporally Stabilized ETDS Network), designed to achieve superior visual quality upscaling without the hallucinations typically introduced by neural network models in video processing. TSENet excels at reliably upscaling video frames that have undergone compression at various stages of the media workflow. This presentation will delve into the network’s performance on CPU, evaluating its effectiveness through both subjective and objective quality assessments.

The broadcast industry is experiencing significant challenges of production capacity and cost efficiency. Growth in live sports/events and the rise of new OTT platforms is increasing demand for programming. Rights holders and venues are looking to produce their own content to retain and grow their fan base while investing in AI to drive operational efficiency, automate content creation, and offer more tailored experiences. Meanwhile, the migration to IT-based platforms continues without clear standards, making this a top priority for the industry. 

Broadcasters are looking to move away from proprietary vendor appliances and closed ecosystems to open-source software, COTS hardware, and cloud-enabled workflows. As such, Intel is working with other IT companies, broadcast vendors, standards organizations, and end users to create consensus for an open standard for software-defined broadcast (SDB), including standardized containers running audio/video processing engines with a focus on – memory sharing, low-latency media exchange between hosts, connectivity with hardware AI/GPU engines, and interoperability with existing video standards and infrastructure. 

Software defined broadcast promises a lower total cost of ownership (TCO), flexibility in selecting applications from multiple best-in-class vendors, scalability using common IT infrastructure, and sustainability through reuse of common platforms for multiple video pipelines and workflows. 

Moving to an open SDB future requires a tremendous amount of cooperation and consensus from vendors, standards organizations, and broadcasters. At the core is a common API for media exchange, so software from any vendor can access and expose audio/video streams in a standardized, efficient way, e.g. using memory sharing within a machine and RDMA connections across multiple machines. It is also critical to provide services placement and connection management across multiple hosts to maintain high bandwidth stream processing and transfers with high QoS and low latency. Other considerations include seamless AI integration and security to prevent unauthorized modification of data between vendor applications. 

The panel will review the current state of development for SDB and propose ways forward from technical, logistical, and commercial perspectives.  Panelists will represent the viewpoints of end-user broadcasters, IT infrastructure companies, and broadcast vendors. As part of this, they will consider how emerging AI technologies can play a role in innovation and create an inclusive ecosystem. 

In recent years, with the shift to Internet Protocol (IP) production equipment and the promotion of remote program production, the demand for products compliant with the SMPTE ST2110 standard has increased.

AMWA NMOS has been standardized as a control protocol, but it only supports control, such as device discovery and connection. Vendor-specific protocols are often used for detailed control, such as adjusting audio levels and matrix control. Therefore, when using equipment from different manufacturers in IP remote production between the subcontrol room and the remote location, communication often fails, making system design difficult.

To explore the possibility of remote production that maintains compatibility between different vendors, we, the Japan Broadcasting Corporation (NHK), attempted to develop a tool that uses international standards for simple control, such as device discovery and connection, and an open standard control protocol for detailed control.

First, we used “Ember+,” which is believed to be highly versatile as an open standard, to develop a tool that can achieve detailed control such as the microphone head amp gain of the input/output unit and the ON/OFF of phantom power. The results demonstrate that this is a useful control protocol for remote environments. Using this knowledge, we attempted to develop a tool that could further promote remote production.

Therefore, we focused on the Dante standard, which has a large share of audio IP standards in the audio market. Although equipment exist that can convert Dante products used in short-distance networks into the international standard ST2110 for use in remote production, two issues make it difficult to use Dante products in remote production.

The first issue was related to control, and although products existed that could convert audio signals between the Dante and ST2110 standards, they did not support control signals.

The second issue was related to transmission delay. Because Dante was a standard developed for local networks, it is unsuitable for long-distance transmissions that involve large delays, and it will require significant effort to design a system to synchronize between bases.

Several broadcasting companies use Dante equipment, and we attempted to develop a tool that could convert between Dante and ST2110, which addressed the two aforementioned issues, to effectively use their assets.

To achieve this, we converted Dante control signals into Ember+, which was believed to be useful in remote environments, making it possible to control audio levels and matrices.

In addition, a re-clocking mechanism was installed in the converter to facilitate synchronization and bridge the ST2110 signals, thus reducing the amount of work required in the system design.

Using this converter, we can now use Dante devices that were previously difficult to incorporate into systems for remote production and broaden the range of remote production.

Currently, we are working on other initiatives for promoting remote production. One involves developing a tool to remotely control devices that are not intended for use in remote environments by converting Ember+ into general-purpose input output. For instance, we designed a cough box that supports remote control such that comments can be recorded by simply bringing a microphone and the cough box to a remote location; we believe that this equipment will reduce the burden in remote locations.

NHK plans to continue developing equipment to promote remote production while considering the needs of program production.

Eurovision 2024 was a pivotal moment for the advancement of large format LED displays in live events. Most live events like Eurovision still run baseband infrastructure such as SDI, DisplayPort and HDMI. However, in 2024 the world’s largest televised music show decided to take a bold leap and run the entire video system on ST 2110. 

This session will explore the how Megapixel and partner NEP & Creative Technology planned and delivered project flawlessly. In addition, attendees will learn more about the drawbacks of baseband technology in large format LED systems and how transitioning to ST 2110 can improve reliability, performance and sustainability of projects. 

As streaming media continues to evolve, the demand for low-latency content delivery and fast tune-in times has become paramount, especially for live events and interactive applications. In response to these needs, a new addition to the MPEG DASH standard, known as Low Latency Low Delay (L3D), has been introduced in the latest 6th edition of the spec. L3D aims to facilitate two primary use cases: reducing latency and minimizing tune-in times, thereby enhancing the viewer’s experience.

L3D packaging aligns the way DASH content is prepared for low-latency playback more closely with the packaging methods used in Apple’s HTTP Live Streaming (HLS). This alignment makes creation of a unified packaging and delivery pipeline much easier, thus simplifying workflows for content providers who deliver across multiple platforms. By bridging the gap between DASH and HLS, L3D promotes interoperability and reduces the complexity associated with simultaneous support for multiple streaming protocols.

In our implementation, we have combined a low-latency encoding pipeline with a low-latency ingest protocol and a packager that supports L3D. We developed an L3D-compliant client capable of leveraging L3D renditions for fast tune-in and channel zapping scenarios. This client enables adaptive bitrate (ABR) switching with minimal overhead, allowing for quick transitions to lower bandwidth representations when network conditions fluctuate. Such efficiency is crucial for maintaining stream quality and reducing buffering, thereby improving overall user satisfaction.

An additional benefit of utilizing L3D is the enhanced synchronization of playback across independent client implementations. Due to the nature of L3D, there is more natural interoperability that synchronizes the playhead positions of different clients. This synchronization brings over-the-top (OTT) streaming experiences closer to traditional broadcast, providing a more cohesive and unified viewing experience regardless of the client’s platform or device. It opens up possibilities for synchronized second-screen experiences and interactive applications that require tight timing control.

This presentation will delve into the technical details of our implementation, highlighting the challenges faced and solutions developed in integrating L3D into existing streaming infrastructures. We will discuss the modifications required in the encoding and packaging processes, the optimization made in the ingest protocol, and the architectural changes necessary for supporting L3D in client applications. Furthermore, we will compare the performance of our L3D-enabled pipeline against current state-of-the-art practices, demonstrating significant improvements in latency and tune-in times.

By sharing our findings and experiences, we aim to contribute to the industry’s understanding of L3D and its potential to further optimize low-latency streaming. The results will be particularly valuable for streaming engineers, content providers, and technology strategists interested in the latest advancements in streaming protocols and standards.

Distribution of content has a multitude of challenges. And there are conflicts too.

The desire to deliver the highest quality video can be at odds with the technical and commercial realities of creating an optimally efficient distribution network.

When dealing with multiple partners, affiliates or takers, not every network end-point has the same needs. Some takers may be limited by bandwidth on their incoming network, other takers may output a video resolution that differs to others or potentially when global distribution is required, at a different frame rate. The Distribution network operator will, for operational and cost efficiency reasons, want to limit duplications to the formats created.

Live transcoding, in conjunction with video processing to satisfy the needs of different takers may provide a solution for some distribution networks. However, to-date, most transcode solutions have required significant GPU processing power – frequently sited at centralized datacentres and provided as a service. This can result in multiple streams egressing from the data centre – driving up cost and placing high reliance on good connectivity from the data centre to the network edge.

For the distributor, there can be alternative processing topologies that can be more efficient in terms of network utilization and cloud OPEX costs.

A better network topology may be constructed where the content distributor originates a single format video service and maintains this format as close to the network edge as possible for all takers. This distribution format can be tailored for the majority of takers, with encoding and networking parameters optimized to provide best user experience for the majority.

For those takers who require a variation in service configuration from the majority, new technology can be deployed close to the network edge that enables optimized network topology and optimized transcode conversion to their individual requirements.

This presentation discusses new technology that democratizes live transcoding technology away from solely GPU data-centre implementations. We will discuss implementations that allow processing efficient and cost-efficient deployments across a wide range of devices that include cloud and ground-based COTS and edge-video-node implementations which allow deployment at the most optimum locations.  

With distributed transcode technology deployed closer to the network edge, this presentation also discusses the importance of network management – providing a single workflow that does not discriminate between the technology topologies of the distributed transcoding implementations.

With a distributed and democratized transcoding implementation, efficiency and simplicity of network operations rely heavily on centralized and intelligent network control. This presentation describes how network control ensures that distribution of content to end takers can be managed without presenting complications of variations of individual transcode implementation.

Through innovative operational management, this presentation discloses how a wide raging portfolio of transcoding technologies, from GPU to FPGA to pure software can be rapidly integrated to deliver the most appropriate solution. Through these advances, it becomes possible to build and configuring a distributed network at scale and at pace.

By applying appropriate, democratized and distributed technology from the head of the network through the snake of internet delivery to the tail-end of the network, this presentation will show how content distribution can be made optimally efficient by utilizing the newest and best edge-transcode technologies allowing delivery networks at scale with a rapid and cost-efficient topology.

As streaming services evolve, viewer demand for new and innovative ways to consume content is making multiview an essential feature. Today’s audiences expect experiences that are engaging, interactive, and mobile-friendly. OTT service providers must adapt by empowering fans to follow their favorite athletes, games, and programs. Users should be able to personalize their viewing experience alongside the curated director’s feed. But delivering a high-quality multiview experience is a complex challenge.

Many OTT streamers have started to introduce “Multiview”: YouTube/NFL, Peacock/Olympics, NBA, Prime Video/UEFA Champions League, and more.

But the current approaches to Multiview all have significant downsides in flexibility, scaling, and user-friendliness. 

The paper will explore four primary approaches to implementing multiview, analyzing their pros and cons: “Many Encodes”, “Many Players”, “Cloud Encoding” and “Single-Player Multiview”. It will argue that single-player multiview is superior to the other approaches.

By NAB 2025, a leading motorsports OTT streaming provider will have just launched a groundbreaking multiview service based on a single-player approach. This OTT service, offered world-wide, will lift the downsides of traditional multiview while providing an unparalleled user experience. The talk will be able to explain the technology behind this service, powered by AWS Elemental encoding.

Existing Approaches

We will first discuss the most common approaches used to date

1.  Many Encodes

Server-side encoding of multiple layouts is perhaps the simplest way to deliver a multiview experience. This approach is fairly easy to deploy, and it is currently used by YouTube, and NBC Universal used it for the Paris Olympics.

Unfortunately, every combination of streams requires a unique transcode. This solution is only practical for basic layouts with a limited amount of feeds. More then 4 feeds quickly lead to a combinatorial explosion, especially when smaller and larger feeds need to be combined as required to keep streaming bandwidth in check.

      2. Many Players

Another straightforward multiview solution is using a separate player for every feed on the screen. Apple and FuboTV do this. But different platforms allow different amounts of parallel decoders, so a per-platform approach is required. Further, it is impossible to keep players in sync, and they start fighting for ABR bandwidth leading to ABR oscillation and frequent buffering. 

      3. Cloud Encoding

In this method, each viewer is assigned their own cloud encoder for the entire duration of their viewing session. It’s like what’s done for cloud gaming. This is flexible and bandwidth-efficient, but prohibitively expensive to scale to millions of users – even if the required resources were available.

Single-Player Multivview

A novel method, Single-Player Multiview uses a single streaming engine to retrieve all feeds and manage all buffers.

On most platforms, it even uses only one single (hardware) decoder for all cameras or games in a multiview “scene”.

Retrieving only what is on the screen at the resolution that is visible, this method has an unfair advantage in:

• Streaming and CDN efficiency,
• Stream switching speed,
• ABR management,
• cross-stream synchronization,
• UX flexibility
• The number of feeds that can be watched simultaneously

In short, single-player multiview combines the best user experience with the highest distribution efficiency.

With full knowledge of everything that is happening on the screen combined in a single player, that player can adapt its ABR decisions and segment retrieval strategies accordingly. Single-Player Multiview is also the only approach that can ensure reliable sync between multiple feeds.

Wireless camera feeds are an integral source of content for programme making, and are typically done using licensed point-to-point radio links or “bonded-cellular” devices. Cellular bonding using public 4G and 5G networks has become a mainstay for electronic newsgathering and remote contribution feeds. These contributions can tolerate latencies up to several seconds. The use of private 5G to support this contribution workflow has been well documented, such as the coronation of King Charles III. However, such latencies are far too long for broadcast production, where wireless cameras are cut in with cabled systems, or where remote camera control, tally and return video are required. Since 5G is a native IP technology, networks are able to support bi-directional connectivity and facilitate these additional services alongside ultra-low latency video. This presentation explores the use of (private) 5G to support full low-latency wireless production workflows, along with discussion on device connectivity, augmenting existing wireless systems, and practical advice for configuring camera control systems over 5G.