Dolby AC-4 Goes Cloud-Native: A Live NGA Production Workflow at IBC2026

Many events and discussions in 2026 highlight that video and audio are forever intertwined. Both broadcasters and OTT streaming service providers now use immersive and personalized audio formats within their production workflows. Video without proper audio significantly minimizes customer engagement. Premium content must provide the audience with a fully immersive experience to make consumers sign up for these services. Until recently, audio has played a secondary role for live events. This has changed with ongoing initiatives centred around Next-Generation Audio (NGA) production.

Up to now, NGA live production has primarily focused on on-premise workflows where all processing was handled in the OB (Outside Broadcasting) truck and/or studio. However, this is not very flexible and quite expensive in relation to equipment acquisition and maintenance. Moving this workflow into the cloud has a lot of benefits for broadcasters. For IBC 2026, a coalition of leading technology providers created a live demonstration taking place throughout the RAI, showing how an NGA production workflow featuring Dolby AC-4 can be easily migrated to the cloud. The following Proof of Concept (POC) shows how a Serial Audio Definition Model (S-ADM) based production workflow in the cloud can look.

Technology Partnerships That Drive the POC

Four industry-leading technology providers have driven the POC to completion. Each company focused on their core expertise, but only close collaboration enabled them to successfully assemble an NGA production workflow for Dolby AC-4 in the cloud. The showcase is presented by the following companies:

• Dolby has a long history of licensing audio technology to the broadcast and streaming industry. Their latest Dolby AC-4 and Dolby Atmos (DD+JOC) technologies are the mandatory or optional audio formats in almost all standards and specifications worldwide.

• Jünger Audio, a leader in developing professional audio processing solutions for broadcast and media applications, provides scalable real-time audio processing for on-premise, cloud and hybrid environments with its software-defined flexAI platform. This includes processing for loudness and voice management, immersive audio and Next-Generation Audio workflows.

• MainConcept is the leading solutions provider for licensing Video and Audio SDKs (Software Development Kits) companies that requires encoding and/or decoding within their product or service. Their AVC/H.264, HEVC/H.265 and VVC/H.266 SDKs are the driving force behind many broadcast solutions and, of course, this POC. As a long-term Dolby licensee, they will soon release Dolby Atmos and Dolby AC-4 SDKs to enable the full NGA processing pipeline.

• Techex shares tx darwin in the showcase. It is an flexible processing platform that demultiplexes incoming live streams, routes the audio to tx darwin, and remultiplexes the processed feeds, while maintaining frame-accurate time alignment, which is essential for any professional broadcast and streaming workflow.

Keep reading to learn how the Dolby AC-4 NGA production workflow for IBC 2026 works.

NGA Cloud Production Workflow Featuring Dolby AC-4 at IBC 2026

The traditional production pipelines for live events, like sports, musical concerts, and breaking news, as well as playout chains, are still based on a complete on-premise infrastructure where dedicated hardware takes over the processing of uncompressed audio, real-time metadata authoring, as well as the contribution/distribution audio and video media chain. As of today, cloud technologies have evolved in such a way that production workflows can be expanded beyond physical equipment without sacrificing both video and audio quality, latency or control, while leaning towards a more immersive and personalized experience.

The key enabler for such experiences is NGA, which is changing the way audio is produced and delivered. Instead of creating only a fully mixed, fixed audio program, individual audio elements such as dialog, commentary, music and effects or audio description can remain available as separate components. Together with descriptive metadata, this allows the final listening experience to be adapted on the consumer side. In this POC, the open metadata format used is the S-ADM, which is demonstrated in a cloud-based production environment for the first time in the industry.

1. At the beginning of the chain, audio and video is captured at the event’s venue or a playout is used. For this POC, we use a file player module within tx darwin that sends a high-quality video (e.g. HEVC/H.265) with the various audio channels as separate PCM tracks in an SRT-TS (Secure Reliable Transport – Transport Stream) which is sent to the cloud for the actual processing.
2. The Techex tx darwin platform then demultiplexes the feed into separate audio and video elementary streams. The audio elementary streams are sent to an SMPTE ST 302M encapsulator module within the tx darwin platform, to create an SRT-TS stream with PCM audio. This SRT-TS feed is sent to a Jünger Audio flexAIcloud EC2 instance. The corresponding video elementary stream remains untouched.
3. The Jünger Audio flexAIcloud solution does the authoring and creates the corresponding S-ADM metadata, so the renderer within the end device later knows how the actual channel layout, commentaries, languages, i.e., how all objects should be positioned, their volume adjusted, and other items should be configured. For the IBC showcase, we output eight channels, with channels 7 and 8 containing the actual metadata and the rest as audio. Once this is complete, the flexAIcloud instance creates an SRT-TS with PCM audio plus S-ADM metadata packed as SMPTE ST 337 in a single PID (SMPTE ST 302M) and sends it back to the tx darwin instance.
4. In the tx darwin instance in AWS (Amazon Web Services), the authored audio is remultiplexed with the video.. During this process, both audio and video need to be aligned and stamped with valid PTS timestamps, otherwise it could later cause async issues on the consumer side. The resulting stream is again a multiplexed SRT-TS stream with high bitrate video for best quality together with all required PCM audio tracks and the S-ADM metadata that is used to set up the Dolby AC-4 emission encoder for distribution.
5. The SRT-TS stream is sent to another AWS EC2 instance with a Live Encoder application that is powered by the MainConcept AVC/H.264 Video Encoder. Due to unstable bandwidth availability on the IBC show floor, we only encode into AVC 1080p25. However, under better network and hardware conditions, the POC can encode up to 4K60. In this case, users can also switch to the MainConcept HEVC/H.265 or VVC/H.266 video encoders that are the major codecs for the broadcast standards ATSC 3.0 in the US and DTV+ in Brazil.
   

   The S-ADM metadata is used to configure the Dolby AC-4 encoder which is included in the upcoming MainConcept AC-4 Encoder SDK. The target format for streaming is MPEG-DASH. The output used during the POC can, for example, have 3+3 presentations, i.e., three languages, English, German and French - each also available in a presentation with a louder commentator but lower M & E (Music & Effects), or even a seventh presentation with English voice-over as well as an additional audio description speaker. As this is a live demo, it is even possible to dynamically switch between offering different numbers of presentations.

   The encoded DASH segments are transferred to an AWS S3 bucket. Amazon CloudFront was picked as the CDN for OTT delivery with S3 as its origin to finalize the NGA cloud production workflow using Dolby AC-4 audio.

6. The final step of the POC, a viewer will play back and watch the OTT stream on a device. The main requirement is that the device must be able to receive, render and decode AVC, HEVC or VVC video plus Dolby AC-4 audio in either DASH, HLS or CMAF in real time. In our IBC demo showcase, we have used an Android-based tablet with an installed player application that can render and decode the Dolby AC-4 audio with its corresponding objects, which were defined in the S-ADM metadata.
   The user can actively switch between languages, turn on audio description tracks, tune dialog levels, or similar personalization features within the on-screen display
   
   

Live NGA Production Workflows Start Today

Setting up an NGA production workflow requires collaboration between market-leading technology vendors, each of them with a specific but essential piece to set up the full pipeline. 

As a POC, the IBC 2026 demonstration gives broadcasters an idea where the future is going. Consumers expect a fully immersive and personalized viewing experience to pay for premium live content where they not only look but completely dive into what they are watching. This experience is intensified by object-based audio formats like Dolby AC-4.

This POC brings together Jünger Audio's decades of experience in sound processing, MainConcept's more than 30 years of expertise solving some of the most complex audio and video technology developments and Techex’s modular, software-defined media processing platform designed for mission-critical live broadcast workflows. S-ADM provides a future-oriented metadata format for both live and file-based workflows. Dolby AC-4 provides efficient delivery of high-quality immersive, personalized and object-based audio at relatively low bitrates. The complete workflow is hosted on the AWS cloud platform. Together, these technologies show that professional NGA workflows can be deployed on-premise, in the cloud and in hybrid environments without sacrificing functionality, flexibility and quality.

Broadcasters and OTT streaming service providers can save money and engineering resources with cloud production. They don’t need to invest in purchasing and maintaining expensive hardware for occasional live events. Porting these workflows and all pipelines to the cloud lets them launch and use the instances only when needed. The rest of the time, they can be ignored and don’t create any maintenance costs. Moreover, when broadcasters deal with different formats and resolutions, they only need to pay for the hardware used for this specific event. They don’t need to pay for a high-end multicore EC2 instance which can handle VVC 8K60 when only AVC 1080p60 output is required. Users can easily adapt their requirements to their current needs. We expect this flexibility will drive the future broadcast and streaming markets.

Here is where you can find the companies who brought this demo to life for IBC 2026:

• EBU (European Broadcast Union): 10.D21

• Jünger Audio: 8.D37

• MainConcept: 1.C15

• Techex: 2.B45

• Dolby: 5.F86