Latency still a significant hurdle for remote production

By Joy Tang

When the speed of data transmission is low, the broadcast of a remotely-produced event can be badly affected. Viewers are used to high-speed connections, and they can be more unforgiving.

“The biggest hurdle to widespread adoption of remote production is the challenge of latency – especially in the context of premium live sports,” observed Greg de Bressac, Grass Valley’s Vice President Sales, APAC.

“With most fans also following along on social media, even a single-second delay is unacceptable.

“Audiences want to feel as close to the action – if not closer – as they would in the venue, so there’s a strong drive towards higher resolutions and more camera angles to deliver an ever-richer fan experience.

“In turn, this will put more significant stress on the network and available bandwidth.”

Better encoding

One workaround is to use technology that enables less data to be sent over the network to achieve the same outcome. According to de Bressac, more efficient encoding technology such as JPEG2000, JPEG-XS and MPEG offers an attractive alternative.

Such technology can deliver ultra-low delays, comparable to transporting the signal over fibre, at lower-cost and still maintain the viewers’ quality of experience.

Grass Valley’s own Direct IP technology, which delivers reliable live signals across distances of up to 20,000 kilometres, is another solution to latency concerns.

“The use cases we have for Direct IP clearly demonstrate that at-home production is a viable and highly-effective alternative to a traditional outside broadcasting (OB) approach,” de Bressac said.

“We have seen a lot of momentum towards the adoption of IP infrastructures and workflows in the region, which underpin and open the way for remote production. Outside broadcast companies and production services providers are also gearing up to support a summer of sports that will involve a significant remote component in order to ensure staff safety.”

Distributed workflows

Grass Valley is already working with several customers in the region to introduce workflows that support remote and centralised work.

“We are working with a customer to set up a 100% remote production solution for live sports production in the Oceania region with only cameras and camera operators at the venues,” de Bressac shared.

“The production infrastructure will be at a different location and the control surfaces and operators will be at yet another location or locations. This distributed workflow will change the game.”

Trade-offs needed

Kevin Mockford, MediaKind’s Director, Contribution Processing & Product Management, also identified latency as one obstacle to remote production, but emphasised that there are many aspects of remote production, each of which with its own challenges and success factors.

“For example, the remote production of camera-pan and zoom shoots requires ultra-low-latency network infrastructures. This can then have a significant knock-on impact on the compression technology used (if used at all), the required connectivity bandwidth, and therefore, the cost,” he pointed out.

At the same time, rules of thumb may not apply, Mockford added. “There is often an assumption that sub-100 millisecond transmission latency is required for a remote production. However, this is not always the case,” he noted.

“In fact, remote production has even been successfully carried out using satellite links, which typically provides transmission latency of 500 milliseconds or greater. So, broadcasters should carefully consider the trade-offs in terms of what is handled remotely and what is not to ensure they’re getting the best compromise of cost and performance.”

Working in tandem

Synchronisation for cloud-based workflows is another consideration.

“Some of the main challenges we’ve seen involve complexities surrounding the move from ‘traditional’ remote production, where all the video and audio feeds are transmitted from the venue to a central production facility, to a ‘distributed production’,” Mockford said.

“In this instance, the production team is not co-located, and the production tools are likely to be software running in a public cloud instance. The challenge here is getting the live video, audio and data into the public cloud reliably, with sufficiently low delay, while maintaining synchronisation between them all.”

Mockford further elaborated that reliability covers both error-free transmission and protection against complete failure of the link. “Protocols such as SRT (Secure Reliable Transport) can provide error correction, and the use of separate network connections can protect against total failure of the network connection.

“Timing synchronisation challenges can include having the same system clock reference across the whole production workflow and maintaining frame alignment across all the video feeds.”

New solutions needed

Mockford said the SMPTE ST 2110 standards and its use of PTP (Precision Time Protocol) SMPTE 2059 is one approach to dealing with the challenges. “It is perhaps a very traditional approach, taking the established practices from the world of SDI workflows and adapting it for IP-based workflows.

“However, it is not very easy to implement in a public cloud environment. So, an alternative approach using less onerous real-time timing requirements while still delivering the needed synchronisation is key. MediaKind among others are developing techniques to do this,” he shared.