Stream with Broadcast-Level Latency

Accueil / Blog / Stream with Broadcast-Level Latency


In our first and second blog posts in this series, we discovered how Ateme’s solutions enable the ultimate viewing experience when it comes to sports events. But what if the technology streaming these unforgettable sports moments can’t keep up? What if it delivers a few seconds later than other platforms? That’s where Ateme’s end-to-end technology steps in. With it, streaming services achieve broadcast-level latency using Low-Latency HLS and DASH, elevating the way we experience live sports.

Viewers Want Broadcast-Level Latency for Sports Streaming

The latency of OTT multiscreen delivery has traditionally been around 30-40 seconds behind IPTV and broadcast. That explains why I might see the results of a race 30 seconds after my neighbor. This longer latency is due to the different stages required to process and deliver OTT video content. So, to meet the broadcast-level latency expectations of sports event viewers, you must adapt each step of the end-to-end video-processing and delivery processes. Let’s explore these steps, and how latency can be optimized at each one, to delight sports fans.

Step 1: Video Encoding

Video encoding (or video compression) is the first processing stage to consider, and obviously requires some latency optimization. However, while optimizing, we also need to ensure that we keep a pristine video quality, even with 4K HDR content. With the Ateme TITAN encoders, video headend operational teams can manage the tradeoffs required to optimize video quality and latency.

Step 2: From Transcoder to Packager, Leveraging the Power of CMAF

Once it’s been compressed, content must be sent to the packager. This is step 2. While multi-bitrate, multicast TS streaming has historically been used between the transcoder and the packager, there is great incentive to transition to the use of HTTP and CMAF. The benefits of this transition include:

  • Reduced end-to-end latency thanks to CMAF, which enables tiny chunks of video (a few hundred milliseconds each), instead of the traditional longer segments (2-10 seconds each);
  • Simpler on-prem and cloud workflows by leveraging HTTP-based transmission and the CMAF standard, which enables a single video container to be used for both HLS and DASH video delivery;
  • Reliable transmission leveraging TCP instead of traditional broadcast UDP-based protocols;
  • Stable service thanks to CMAF ingest, which optimizes redundancy and avoids service interruptions in all failover use cases.

Step 3: Packaging

Content has been compressed and sent to the packager. Next, in step 3, it must be packaged. Low-latency content can be packaged to viewers in either fMP4 or CMAF formats. Transitioning to CMAF as the core OTT format delivered to viewers also provides great value, including:

  • Drastically reduced latency: The short chunk size significantly reduces the length of the buffer required for playout, allowing the player to start displaying after only a few milliseconds.
  • Simple workflows thanks to the CMAF ingest format, which defines XML-based data. This is simple to manage by the origin and packaging chain. It also avoids potential issues when converting subtitles and captions – a common pitfall.

Learn more here about the value of CMAF to optimize the viewing experience.

Ateme’s NEA just-in-time packager for live streaming maximizes Quality of Experience at low latency for viewers. But low latency is just part of the QoE requirements – as viewers want rich UI experiences with trickplay (i.e. the ability to fast-forward, rewind, pause, or skip to certain parts of the content) and multi-audio (e.g. allowing viewers to choose from different languages, commentary tracks, or audio descriptions). The Ateme NEA packager enables all this, as well as broader audiences for your content with robust subtitle and multi-language audio functionality.

Broaden Device Reach with HLS and DASH standards

While there are several ways to deliver low-latency streams, Low Latency DASH (LL DASH) and Low Latency HLS (LL HLS) provide the best approaches to reach the largest audience of sports fans. Ateme’s NEA packagers can package linear streams as either LL DASH or LL HLS within a second.

There are, however, legacy HLS and DASH devices that cannot support any low-latency approach. The good news is, these devices can still leverage the same low-latency segments. This approach simplifies the overall end-to-end system, although the viewing experience will be based on legacy latency.

Step 4: The CDN. Reducing Costs while Improving the Viewing Experience

The content has been compressed and packaged. Now, it’s time for delivery to viewers through the CDN – step 4. To deliver the partial CMAF segments required for the low-latency experience, the CDN requires support for Chunked Transfer Encoding for LL DASH and either HTTP/2 and/or byte-range requests for LL HLS. An additional improvement is enabled when the packager uses byte-range signaling in the HLS manifest, which allows the same segment to be used for both HLS and DASH. Compared to other approaches for low latency, the combination of CMAF segments for both HLS and DASH devices and HLS byte-range manifests can reduce core-to-edge traffic and associated cache storage requirements by 50% or more, reducing the overall costs to deliver the optimal low-latency user experience.

Achieving Broadcast-Level Latency, at Scale, One Step at a Time

As we have seen, each step of the video-delivery workflow needs to be optimized to ensure low latency, at scale, with pristine video quality.

The good news is that Ateme has invested in all areas of the OTT end-to-end video path to meet the same broadcast-level latency that viewers expect when viewing sports events live, in streaming. Stay tuned for our next blog article to find out who has already deployed it!

Have you ever been frustrated by a viewing experience where the latency was not low enough? Share with us in the comments below!



Leave a comment

Your email address will not be published. Required fields are marked *

Ateme