Amazon Web Services (AWS) announced today that it plans to unveil what it calls the “Cloud Native Agile Production (CNAP)” project, an open-source initiative aimed at accelerating the creation and delivery of media during live production. Developed in collaboration with BBC’s research and development (R&D) team, Sky, and AWS Partners, including Adobe and CuttingRoom, CNAP debuts as more news, sports, and entertainment productions look to develop ancillary content for live programming. CNAP’s interoperable framework provides a single virtual store for live content, making it easier to migrate production to the cloud so that creators can work with their preferred toolsets to rapidly create video highlights clips and packages for integration into live content, social sharing, and other applications, AWS said.

AWS will spotlight an end-to-end CNAP workflow at IBC on its Stand (5.C90). The demonstration includes partner technologies from Adobe, CuttingRoom, Drastic Technologies, Techex, and Vizrt, which collectively support essential fast-turnaround media workflows demands spanning low-latency video ingest through to playout, file import/export, and web-based and craft editing. The project will be highlighted at the live panel, “Reinventing fast-turnaround news and sports workflows for the cloud,” in the Innovation Village on Sunday, September 15 at 11:15 AM.

“By providing an open interoperable approach to fast turnaround media workflows, the Cloud Native Agile Production project is breaking down workflow silos, bringing the technology to support live broadcast and streaming, social media publishing, archive/record, VOD streaming, and more into one cloud-native environment. This convergence unlocks incredible new efficiencies, flexibility, and operational benefits for content providers,” explained Chris Blandy, Director of Strategic Business Development for Media & Entertainment, Games and Sports at AWS. “We’re thrilled to collaborate with our customers and partners on this project, which will ultimately help content providers produce content in the cloud for fast turnaround workflows while also extending growth opportunities for AWS Partners.”

Cloud Native Agile Production
For the project, BBC R&D provided its open-source Time Addressable Media Store (TAMS) specification, which is publicly available on github, and stores, queries, and accesses segmented media over HTTP. AWS then began building on top of the modern open, API-driven architecture to ensure AWS customers and partners could leverage the CNAP framework to deploy custom TAMS implementations in AWS.

Presenting a cost-efficient alternative to the lift-and-shift, file-centric workflows deployed today, CNAP limits time spent processing content, according to AWS. It supports a serverless, chunked media store approach with Amazon Simple Storage Service (Amazon S3) providing the underlying storage layer. Amazon S3 enables a ‘store once, use many’ approach to repurposing media, so simple edits can be expressed as a metadata ‘publish’ rather than a new asset or exported file. This strategy reduces storage duplication, time spent processing storage, and the volume of space required for the same workload.

“Based on our experience and the data we’ve collected, the costs related to running an ingest 24/7 with the Cloud Native Agile Production project are impressively low,” explained Chris Swan, Industry Specialist Solutions Architect for Content Production at AWS. “Running a workflow on CNAP for three months costs the same as running a similar workflow on incumbent solutions for only a week.”

“Many M&E customers still run operations on premises and want to move into the cloud but require seamless interoperability between their preferred creative toolsets and AWS services. The Cloud Native Agile Production project introduces a solution with a cloud-native framework that enables industry-standard tools to work better together,” Blandy added. “Maintaining an open, interoperable approach is key to making progress on this front. The more AWS customers and partners that lean into this project, the higher the level of interoperability we can achieve so customers can work with a broader range of tools, which benefits the entire industry.”

“The Cloud Native Agile Production project is an exciting framework designed to optimize rapid media workflows in the cloud, and a significant milestone in promoting cloud adoption and enhancing workflow efficiencies,” shared Simon Williams, Vice President, Digital Media Strategic Development, Emerging Solutions and Partnerships at Adobe. “Adobe Premiere Pro is now the first desktop video editing software capable of directly accessing, playing, and editing growing video files on Amazon S3 – and we’re excited about the new capabilities and workflow efficiencies this will unlock for customers.”

The AWS CNAP demo at IBC provides an end-to-end workflow from ingest, through editing content, to gallery playback through a vision mixer. The tx darwin live media processing platform from Techex ingests feeds from the nearby Newsroom in the Cloud demo, including a live camera on the stand. Once registered in the open source implementation of the TAMS store, the chunked media is stored on Amazon S3, and a React web application hosted on AWS Amplify visualizes the content.

All content in the TAMS store is accessible to view and edit via the CuttingRoom video editing platform web user interface. A Drastic Technologies plugin then enables content editing on a virtualized Adobe Premiere Pro workstation. Using a Vizrt Vectar Vision Mixer, content from the store can be played back and mixed with live feeds. Showcasing the different ways that TAMS store content can be distributed, the demo also illustrates a TAMS store content stream to a mock news site.

CNAP’s Beginnings and Future Roadmap
AWS says its CNAP project originated out of an alignment between the BBC and AWS on the BBC’s TAMS chunked media store concept. As the two companies collaborated, they realized that the TAMS store framework could solve a larger industry challenge and set out to build CNAP with the help of multiple AWS teams. The goal was to build an open-source implementation to deploy the TAMS store in AWS, enlist partners to join the effort in April 2024, and demo an end-to-end workflow at IBC.

Prior to IBC, AWS and BBC R&D held a special event to brief partners on the project and ideate what an end-to-end IBC demo might look like. They encouraged participants to experiment with the API during a hackathon, which resulted in new capabilities such as the ability to import existing media from TAMS into the CuttingRoom web-based editing tool. “The Cloud Native Agile Production project is the answer to standardizing what we already are doing, making interoperability so much easier and cost-effective for us, partners and customers,” explained CuttingRoom CEO and Co-Founder Helge Høibraaten.

CNAP is still in early development, and the team behind it is exploring how they might apply live AI analysis to the workflow in the future, in addition to other improvements, according to AWS.

John Biltcliffe, Senior Solutions Architect for Media & Entertainment, Games, and Sports at AWS, explained just how important partner collaborations will be to CNAP’s continued evolution, “Embarking on a long-term project like CNAP, with so many stakeholders, requires time, attention to detail, and strategic communication. We’re grateful to BBC and Sky and all of our partners on the project thus far, as well as the many AWS teams who have helped us realize our vision. With IBC around the corner, we look forward to demonstrating the possibilities of CNAP to the community and gathering even more input to continue advancing the project in the greater interest of the M&E industry.”

PARIS—Motion Spell has announced that Netflix has completed an 18 month transition to integrate GPAC Open Source Software in Netflix’s worldwide content operations. The GPAC Project on Advanced Content (GPAC), is an open source software that provides a modular multimedia framework for packaging, streaming, inspecting and playing content. It provides a wide range of tools for processing media content and has a strong focus on standards.

Motion Spell SARL, based in Paris, who exclusively handles commercial licensing for GPAC, also provided consulting expertise to Netflix’s R&D division as part of the deal.

Due to the large amounts of packaged media Netflix produces daily, robust packaging tools are crucial; in addition Netflix also updates its packaging recipes. For example, they recently started publishing AOM’s AV1 packages so their packaging tools need to be flexible, and adaptable to new technologies. Netflix reaches a wide variety of devices: from TVs, to set-top boxes, to mobile devices. To that end, they rely on the use of standards, such as the ISO Base Media File Format (ISOBMFF) and thus need standard-compliant tools.

Romain Bouqueau, CEO of Motion Spell, said, “Netflix is a company with an exceptional team, clear goals and an innovative R&D programme. Combining our expertise made the transition process to GPAC very efficient. We’re delighted that Netflix has gained in flexibility, while GPAC has gained a customer who respects the value of open source software.”

Anne Aaron, Director of Encoding Technologies at Netflix, said, “As part of our strategy to build a flexible and efficient packaging platform, we needed a strong partner with deep expertise of the best encoding and packaging tools. Working with Motion Spell helped make the entire process of integration incredibly easy.“

SAN ANTONIO, TEXAS--Just prior to the IBC Show in September, NewTek, a San Antonio, Texas-based provider of live TV production systems announced the introduction of NewTek NDI, a new open standard for live production IP workflows over Ethernet networks. Broadcast Engineering Extra recently spoke with NewTek CEO Dr. Andrew Cross to discuss the current state of video over IP, the standards involved and how NewTek NDI could impact the future of live video production.

BROADCAST ENGINEERING EXTRA:What are the biggest challenges currently for using IP for live production?
CROSS: First of all, there's so many pieces to an IP production, it's a bit hard to answer this with a one line answer. Obviously, you've got the file-based part of it. You've always got B-roll, you've always got titles, you've always got graphics, things like this. Those are probably using IP today, using centralized storage systems already. That part is probably fine.

It's the live video feeds part which is a challenge because I think this is something where there are no good standards. People are using SDI, they're using HDMI and maybe a few people are still using analog video of some kind. That is just something that's remarkably hard to do in IP today.

There are some solutions. You can buy boxes that convert in and out of IP, but you're going to fight those really hard because if you've got two different parts, you aren't guaranteed they're going to have the same latency; it's a mess. If you have some sources coming in in IP and some are not, how do you make sure they're in sync with each other? That's really a difficult problem at the moment.

BEE: Newtek NDI is an open source protocol for using IP and live production. Can you explain the genesis of this standard and why did you pursue this path?
CROSS: If you look at the standards that are out there, like SMPTE 2022-6 and the new TR03 standard, in large part what they are trying to do is take a traditional SDI feed and now put it pretty much one for one over IP. And the other real place where they differ [from NDI] is that you kind of run out of bandwidth on IP pretty quickly because it's not as high bandwidth as SDI. They are looking at three to one compression and there's no standard in either of these things in regards to what kind of compression should be used, although, I think that's the direction everything's going.

Read through the standards and you’ll see what I mean. Imagine if you want to start a company and you’re thinking "OK, IP video's big. So we can do cool things with it. We can make computer applications, we can make cool pieces of technology that talk IP and put video over IP." And now read through one of these standards and try to imagine that without a million bucks behind you how you could possibly implement that in software. Because the problem is these protocols have really, really tight timing requirements.

For instance, in TR03, they need 10 microsecond timing accuracy for each scan line of video. There's just no way that a computer system without customized hardware is ever going to be able to achieve that.

The big benefit of IP is that every computer system in the world can talk with it. You can pipe it around, you can write applications and do things that were never thought of before. If you look at almost any other industry it's going to be the small companies with cool ideas; they're going to revolutionize this industry. Yet, the existing standards are really only taking SDI and putting it over IP and that just doesn't seem where the future of IP video can be. But there is a place for that. So, in many ways I don't see what we're doing is counter in any way to the existing standards.

NewTek comes from the computer end of the spectrum. We've been building real-time computer-based processing systems for 10-15 years, and we've been doing video over IP for 10 of them. Over those 10 years we've worked out how to get IP video in and out in ways that our products can work with and what we actually did with NDI is take that and open it up to everybody and with that, we can bring in a huge number of vendors that already support this.

BEE: Was NDI already available with NewTek products and now it’s just the case that you’re opening it up to everybody?
CROSS: Yes, it has gone by a variety of names. For years, we’ve had the ability to take video from a Vizrt virtual set system, for example, and send it over IP into a NewTek Tricaster. We've done the receiving, they've done the sending. We've been relatively open about giving anybody the ability to send us video. And we've had the SDK for a long time. What we've done now is opening up all parts of this—the sending part is open, but now the receiving part is open. Even those vendors making competitive products.

People making TV monitors or video monitors now can all receive the feeds from this. Right along with this is the ability to find sources on the network, so If you have a laptop running Power Point, you can run the NDI mirroring application on that PC. There are sources available for anybody doing live production work on that local network to see. So our APIs make it very easy to find sources on that network to then bring the video in or to send video out to them.

BEE:There has been a lot of talk about the fact that video switching is hampering the move to IP. Can you comment on how NewTek handles IP within the switching environment?
CROSS:  Absolutely. We were actually showing this at our IBC Show booth. We had Tricasters with 12-plus inputs all coming in over IP. And because we've been doing this for some time and we're not starting now, we've got compression and it's all standardized, built into our system. We can work with what is basically baseband compression; it's similar quality to say, ProRes or DNXHD. At IBC, this meant that we had lots of streams over a relatively low bandwidth connection with very low latency on a standard 1 Gbps connection. This is something that we're actually going to have in the market here very, very soon. Think a few weeks, not a few months.

BEE:How will the introduction of New Tech NDI affect the company's future product development?
CROSS:  A question I get asked a lot—and is directly related to this—is “why on earth would you take all this effort you've put into IP—working with lots of other companies—and just open that up for everybody else to use?” Ultimately, the reason for that is I think that NewTek believes very strongly that the future of live production is going to move towards IP and quite frankly, we make products that have done this for a long time, and are particularly well suited to this. The faster we can help get IP adoption in the market, the better—the better for us, our customers and really the better for everybody. We see the future of video transport being over IP. It just makes sense for so many reasons. Clearly, you're going to see more and more products where IP is increasingly important in them.

BEE: This is an industry that has been built on standards. Do you think that the standards-making bodies are moving too slow and is that one of the reasons why you introduced NewTek NDI?
CROSS: That's a great question. First of all, I should say our attempt with NDI is not to circumvent standards bodies. I have reached out to everybody who has worked on each one of these different standards and said "Look, we've got this huge installed base. We're making it very easy for people to build products that work over IP. We have probably about a 100,000-plus systems in the market today that can already talk these protocols. Why don't we work together on making this something that's more broadly accepted?" We're not looking to own or control this. We would be very happy to work alongside somebody else to merge what we're doing with them. We're very open to that. Our intent is not to try to lock off part of the market or anything like that. Otherwise, we never would have just opened this up. We want to enable IP video.

Clearly the standards bodies are slowly getting there, but, at least from a NewTek perspective, we tend to come from a different end of the market.

If you were a high end broadcaster and you're looking at an SDI cable and you're saying "How on earth do you get this into IP?" you're going to build an IP solution that looks very much like an SDI cable. Quite frankly, because the people driving the standards bodies have tended to be people who have been driven by the top ends of broadcast, the standards look like that. Now, I'm not being critical of them, in fact that's why I'm saying that I think there's a great opportunity for somebody like us who tend to come from the other end of the spectrum to work with somebody to make these things coexist.

I actually see that NDI, in many ways is not competitive with SMPTE 2022-6 or the new TR03 or any of the others. I think they're trying to achieve a slightly different goal. If you look at almost any area of technology— such as computing—the things that become software-based ultimately tend to win out in almost every area of technology. Custom hardware tends to not be on the winning end of that spectrum over any sustained period of time. So, what we're trying to do with NDI is make something that's very compatible with that.

BEE:When will we start seeing products or technology from other vendors that are compatible with NDI?
CROSS:  They already exist. We're going to be releasing tools within a few weeks that take NDI and attach it to any of the commodity capture cards out there. For instance, if you have a Blackmagic card that can access a network input source, if you have an AJA output card somewhere else in the building you can just send these from one place to another over IP. We have more than 100 companies that support this.