In late July this year, three companies; EVS, Cisco and Imagine Communications combined to present a “Live Production in an IP World” webinar. The companies are all members of the AIMS Alliance.

The webinar identified the benefits of moving to IP for live production and the required components to do so. Presenters explained how to design a working IP production solution and discussed the most appropriate time to make the IP move. Webinar presenters included; Kerry Wheeles, CTO of Networking, Imagine Communications,James Stellpflug, Product Marketing VP, EVS and Bryan Bedford, Bryan Bedford, Business Development Manager, Cisco Global.

IP enabling allows producers to do more with content and move content and files to the right places. A good example of IP enabling being used today is the recent Euro Cup coverage in France. Files from ENG crews and the OB truck were moved to the central broadcast center in Paris and down through the cloud for direct-to-consumer distribution. Using IP for file transfers, content access and distribution, content creators and broadcast partners keeps more people working from home, hotels and broadcast centers instead of the event venue itself.

IP will make its next inroads by changing the infrastructure within broadcast centers and studios, live events production, stadiums and new stadium TV production wiring, and OB production trucks. Why?

Trends

“Why” is dictated by today’s trends. According to Brian Bedford, Cisco sees four major pillars in the IP video market today. One pillar is the rapid movement in business models that impact direct-to-audience media engagement such as OTT and virtualization. “All are things are being impacted because the behaviors of our customers are really changing,” Bedford said.

By using a converged infrastructure that can easily handle IP video signals, new practices can be implemented. Cisco has built some broadcast facilities to handle massive scale but in many cases remain underutilized.

Leveraging data center technology and a converged infrastructure opens many interesting possibilities and opportunities, such as the ability to add new channels and services that impact business models and create more revenue streams. Changes in the economy and business models invite diversification. Tomorrow's facilities need an infrastructure ready to support future diversified monetization strategies.

Flex

IP is being introduced into the acquisition phase for several reasons. Switching technology has evolved to provide multiple bi-directional signals per cable including video, audio, metadata and intercom. The ability for live and file-based workflows has evolved significantly.

The ability to “flex” a dynamic infrastructure includes the ability to be flexible in the physical environment. This means not needing dedicated SDI facilities to support multiple environments. File-based playout, virtualization and OTT all require maximum flexibility.

Cisco’s Bedford noted the goal of making changes but ones where the operator’s workflow and environment remain relatively unchanged. Facilities typically have built consistent workflows that operators are familiar with and that have been in place for decades. The strategy is to streamline the workflow “under the hood,” without changing the workflow at the edge.

Fabric

“IP,” said EVS’ James Stellpflug “is already changing the game in contribution. Its already changing the game in distribution and deliveries, and now we look at where it will bring the next round of change in the production fabric.”

Most new TV peripheral gear, such as cameras, video servers, graphics systems, video switchers, multiviewers and audio mixers have the ability to move away from SDI with on-fabric IP connections either built-in or optional. “Cameras, microphones and audio devices, where sourcing is happening,” Stellpflug explained, “you can already buy these things today.” They have the connectors and are ready for the IP fabric, as do servers and graphics systems.

Traditional switchers usually have a physical limit with input processing based on cards, frames and input and output connectors. But, IP is changing the way new switchers are designed. Old school switchers were limited by connectors and control panels. With IP, those physical limitations are removed.

Scaling

The new generation of switchers can scale through processing modules. Limitations are defined not by connectors but by bandwidth. Off-the-shelf hardware brings the commodity-of-scale from the IT industry to facilities that can start from a video switcher and then scale it up by using an IP fabric. Presently, video switching is one of the most active areas where IP is bringing benefits to the product level.

IP inputs and outputs of off-the-shelf devices and servers can be shared among the different resources, creating pooled resources, scalability and IT flexibility. All these benefits apply directly to a production workflow. A facility can start small, grow through scale and hand all that processing power down to the production switcher functions in a live space as needed.

IP is enabling the flows of inbound and outbound content and data to and from cameras, servers, graphics, live feeds, or outbound to defined destinations. In a diverse variety of processing modules, switcher panels and multiple facilities or a remote equipment room, IP allows facilities to stretch out switcher fabrics so they can all execute and function as if it is all one combined switcher. This moves processing to the edge and eliminates the requirement that all the video signals must be individually wired to one big rack. IP technology enables totally new ways to design systems.

Architecture

System architecture for IP production systems must founded on a common set of marketplace agreed upon protocols.  For example, “The purpose of the AIMS group,” Imagine Communications’ Kerry Wheeles said “is to foster the adoption of one set of common, ubiquitous standards-based protocols for interoperability over IP in the media and entertainment business.”

Wheeles identified three requirements for adoption; technical robustness, market awareness and global adoption. In the case of AIMS, every protocol has been vetted, tested, specified and created by existing standards bodies such as SMPTE, AES, EBU, the Joint Task force for Networked Media (JT-NM) and others. The AIMS goal was to select and promote the adoption of known protocols that allow interoperability similar to that enjoyed by SDI solutions.

AIMS baseline for interoperability is SMPTE 2022-6. It allows users to lightly encapsulate SDI, transmit it over IP, and de-encapsulate back to SDI. The 2022-6 standard has been around for several years and multiple vendors produce products for it. It is in use in on-air facilities worldwide.

AIMS enables separate IP streaming audio with AES67. That audio standard supports split video and audio routing with TR04, which is a combination of SMPTE 2022-6 and AES67 all timed and synchronized by SMPTE 2059.

To save bandwidth, TR03 can move the video essence separately, send it as RFC4175 combined with AES67 for audio and SMPTE 291 metadata, forward it over the same wire and keep it all synced with SMPTE 2059 “PTP” timing support.

IS-04 is an IP network discovery and registration protocol that provides a common format for automatic discovery of devices on a network and register them for connectivity.

Each of these are folded into the SMPTE 2110 specification which is now in draft and the individual protocols are already specified and supported by many vendors.

AIMS, IABM and the Sandbox+Live IP project are working on market awareness and education, as major manufacturers and power users such as TV networks begin early adoption. The standards' implementation will be similar to the roll out of SDI. Both vendors and users are joining every week, and we can expect the trend to grow. It will be similar to what happened with the adoption of SDI when it was first introduced.

Build-out tools

Unless you’re starting from scratch in a greenfield project, the first thing a facility needs to begin an IP build out is a high-density SDI/IP Gateway. It provides simple IP connectivity back and forth with existing SDI legacy gear. Certain capabilities are necessary to accomplish this seamlessly. Kerry said one that Cisco supports is SMPTE 2022-7, “which allows you to have redundant network interfaces on every single gateway, so that you can build redundant IP networks, compare every packet coming in and seamlessly merge between them if there is any issue on either network.”

Redundancy is difficult, expensive and rare in the SDI world. But in IP it is commonplace and actually a standard when developing an IP system. IP technology easily and inexpensively enables the incorporation of full end-to-end redundancy.

On hybrid IP-SDI systems, signals must be synchronized on and off the SDI path. With multiple IP signals on the same wires, the ability for clean switching between them is important. Cisco, for instance, allows users to preset a new source on the same wire, buffer it, switch from preset to line cleanly and convert it back to SDI for on-air or critical signals.

Density is also important. A Cisco frame can handle up to 112 signals in 3RU.

Future proof

IP guarantees future flexibility. SDI and BNC connectors soon will join Y/C and RGB component analog cables in the museum of video magic. Remember EBU Rec. 601? Whether redesigning, expanding or starting out greenfield-style, IP is the future. Coax is for RF.

At the physical layer, IP over Cat 6 or fiber makes more sense than coax, particularly looking towards a UHD future. There’s little comparison between the sheer weight and size of the wires necessary for a HD-SDI UHD quad link and IP. A quad link supports UHDTV-1 resolution of 2160/60p. There’s limited bandwidth and severe distance restrictions over a coaxial quad link for HFR or HDR.

Will there ever be one standard the whole world uses? Stay tuned.

UHD

You can use SDI-over-coax for a quad link, as defined by SMPTE 424M at 3Gbs per coax. There are two types of quad link signal formats. Two Sample Interleave (2SI) has become standard on most new equipment and has only one line of inherent delay. The other format is Square Division (SQD), which takes the total resolution and breaks it into four quadrants. It was built into most earlier quad links and has an inherent ½ frame delay.

SMPTE 2082 defines a single link SDI@12Gbs signal per coax. It extends existing SDI infrastructure and supports 2160p60, but cable length restrictions generally confine it to OB and mobile production trucks. It is not capable of supporting higher frame rates.

Presently there are three methods to move UHD over IP. One is the Sony Network Media Interface (NMI). It provides a single multicast group per signal, similar to SDI or 2022-6, where everything is contained in one group. Break-away audio requires de-encoding and de-embedding, or separately on AES67.

The slight compression NMI uses allows two signals to fit in a 10 GB link for buffered destination-timed switching capabilities, similar to a frame-synched preview channel. NMI compression is virtually lossless, quite applicable for production environments and found in many of Sony’s new cameras and production switchers.

The second method is the TICO Alliance (RDD35). TICO uses virtually lossless, low-latency 4:1 compression, supporting three UHD signals per 10 GB link for destination-timed switching. Separate multi-cast groups allow for audio break-away without de-embedding.

TICO technology was developed by intoPIX, and designed for 4K/UHD transport over legacy SDI infrastructure and modern IP production and contribution networks.

The third method is ASPEN. ASPEN is an encapsulation format that packetizes uncompressed SD, HD, 3G and Ultra HD signals into an MPEG-2 Transport Stream. This encapsulation method is documented in SMPTE RDD-37. The resultant Transport Streams are formatted into RTP/UDP/IP packets according to SMPTE ST 2022-2 for IP network transport.

The NewTek NDI (Network Device Interface) enables IP workflows, supports 10Gbs with bandwidth for UHD. However as of right now, NewTek has yet to introduce any other UHD product._­­­

There is also uncompressed UHD via VSF TR-03, which does not fit into a 10Gbs Ethernet. Uncompressed UHD requires a 25-100Gbs Ethernet. It would provide separate multicast groups per essence, each containing video, audio and metadata.

The biggest problem with uncompressed UHD via VSF TR-03, Cisco’s Bryan Bedford explained “is that it’s not available yet, and it’s not immediately pending either. So while the network and the basic routing capabilities from companies like Cisco and others can already support 20-100Gbs Ethernet … there is no real support for uncompressed UHD anywhere, and very little on the road map.”

Graphics courtesy Cisco, EVS and Imagine Communications.