Blog Post by Charlie Kraus, Senior Product Marketing Manager
December 30, 2020
Nearly every major social media platform—with the exception of maybe Pinterest—now has live streaming built in. Live streaming is everywhere, from e-sports to actual sports to SpaceX launches, and more children want to grow up and become professional live streamers than want to become astronauts. On the technical side, platforms are now in an arms race to provide low latency live streaming to their users.
It might be easy to go live nowadays but affecting that technical marvel behind the scenes is anything but simple. Since live streaming is table stakes for your platform, the solution that can offer the broadest range of low latency streaming capabilities will outcompete the rest. Not every live streaming application requires the lowest possible latency. There are several factors that determine which solution is the best fit for a given application. Big-name streamers will flock to the platform that offers the most appropriate video latency solution because that platform will provide the best viewing experience for their audience.[KC1]
Latency in video isn’t quite the same as latency in other forms of internet media. In the latter, latency is the time it takes for a server to fetch content and send it to your browser. Video latency, on the other hand, represents the lag between the time when a frame is captured and the time when it shows up on your monitor.
As such, video latency can be a multidimensional problem. For example, imagine that you’re streaming an e-sports competition in which two well-known gamers are competing against each other. The stream doesn’t include just the action of the video game—it also includes camera shots of each gamer’s face, plus an additional video inset with a live commentator.
In order to present an optimal experience, four separate video streams—the game, the two players, and the commentator—need to be broadcast simultaneously to a vast audience. (For comparison, congresswoman Alexandra Ocasio-Cortez’s recent Twitch event reached an audience of 430,000 viewers.) If one stream has higher latency than the others, the whole presentation suffers—you could find the commentator remarking on action that took place up to 30 seconds in the past, for instance.
As the streaming audience matures, they’re beginning to demand highly interactive and technically sophisticated streaming events. Interactivity requires the lowest latency, ideally sub-second, in order to provide the best viewer experience. Today’s audiences also want a high- quality broadcast-like picture quality. [KC2] It’s almost a given that a successful streamer will have a high-bandwidth fiber internet connection, and they will choose the video format and encoding protocol that can deliver their streams with the required latency.
AWS defines ultra-low latency as being within a range of 2 to 0.2 seconds, and low latency as being within 6 seconds and two seconds. With that being said, default latency for delivering common HLS and DASH video formats can be from thirty to sixty seconds. With this high latency, it’s impossible to deliver the interactivity that audiences want.
The culprit behind this high latency is what’s known as HTTP-based chunk streaming. HTTP streaming formats such as MPEG-DASH and HLS break the video into small segments or chunks that must be buffered prior to playback. While it’s possible to reduce the size of the chunks to provide low latency, making them too small increases the chance viewers will experience video rebuffering and other playback issues.
Complicating the video delivery workflow, audiences watch streaming video on TVs, phones, laptops, tablets, and more. Each of these devices may use different video format. This means that a broadcaster must take each video stream and transmux it into popular video formats such as HLS and DASH before sending it. In addition, each stream may require configurations to optimize delivery for the best picture quality possible for the viewer’s internet connection condition. [KC3]
[KC4] Limelight Realtime Streaming uses WebRTC (Web Realtime Communication) technology, an open-source standard that can deliver streams with less than a second of latency and is well suited for interactive applications. Well implemented solutions can deliver reliable, broadcast-quality, realtime video streaming at scale. This has been a significant area of development in 2020 for Limelight’s next generation Realtime Streaming solution that is integrated with the scale and scope of global CDN capacity.
While enabling sub-second latency, Realtime Streaming also incorporates data sharing that can be used to create interactive services. We are seeing these emerge in new applications for live video streaming, including online casinos, auctions, learning and in-event sports wagering. We expect to see WebRTC emerge as the leading option for scaling the delivery of live content that is incorporated into the social, audience experience.
In short, the technical side of livestreaming matters almost as much as the on-screen talent. Make the most of your livestreaming implementation with Limelight—contact us today!