Many of us have heard buzz words being thrown around in the streaming industry like “low latency,” “ultra low latency,” and “realtime latency” but many are still confused about what these terms mean, and are curious if latency is important for their streaming initiatives.
This FAQ is meant to cut to the point and answer the most important questions, in a simple and straightforward way.
What is latency and why is it important for streaming video?
Latency in streaming is the time it takes for a live video stream to get from the point where a camera captures the live feed, to the device where the stream is viewed. Sometimes this is referred to as “glass-to-glass.”
Latency is important when a live video feed is time-sensitive and viewers must receive the live content in a short and predictable amount of time, like an online auction for example.
What is the difference between low latency streaming and real-time video streaming?
Low latency streaming is typically video being delivered from capture to viewer in about 20-30 seconds. Ultra low latency streaming is typically 3-10 seconds.
Real-time video streaming is delivered from the source to the viewer (often known as glass-to-glass) in less than 1 second, typically 100—500 milliseconds. For cloud gaming it needs to be as fast as 150 milliseconds, round trip.
Who is real-time streaming for?
Real-time streaming is especially important for auctions, AR/VR, financial news, live events, gaming, sports, trivia, and other cases where time-to-delivery and synchronicity of the live stream is mission critical.
What are the main causes of latency in live streaming?
1. Method of capture
A delay at the source because of the capture workflow of the publisher’s environment. For example, a video publisher may have pre and post processing requirements which cause the stream to be delayed before being sent to the receiving endpoint.
It takes time for an encoder to receive the stream, process it, and output the new stream for delivery. Encoding can take place at the source of the capture or in the cloud, both cases will produce some delay.
3. Method of delivery
A traditional CDN will need to cache live stream segments on a global network before they are able to be delivered to viewers over HTTP, this will generally add 30-60 seconds of latency to a live stream. Specialized network deployments are required to deliver sub-second latency streaming at scale.
What low latency streaming technologies are currently available?
Historically RTMP (Real Time Messaging Protocol) was the standard for low and ultra low latency streaming. This protocol is still used today for publishing, however it is now obsolete for viewing because it requires the Adobe Flash plugin which disappeared in April 2020.
Today the de facto standard for publishing any low latency and real time video is WebRTC.
WebRTC technology was originally developed for “real time communication” in a browser, which is essentially live video conferencing.
Recent innovations by Singapore based CoSMo software and Millicast have enabled WebRTC to support large-scale live broadcasting with sub-second latency audio and video streaming. WebRTC is currently the only current technology to support real-time, sub-second video streaming.
Other low-latency protocols and formats include CMAF, Apple Low Latency HLS, and the up and coming QUIC protocol. However, these protocols are not yet accepted standards.
What are my options for low latency streaming?
There are several options, the best solution for you depends on your use case and requirements. If 30-60 seconds latency is acceptable for your stream, you can use just about any existing solution on the market from CDN providers. If you need better than 30 seconds there are specialized media server software providers which can help you solve this problem.
If you want true real-time latency streaming you have two options:
Licensing a full stack infrastructure from CoSMo, or
If you want a fully managed solution, which includes provisioning features and dashboard i.e. configuring, deploying and managing your streaming business with guaranteed uptime, there is Millicast.
For ultra low and realtime latency, what are the advantages of a fully managed solution?
— No setup required, instant activation.
Simply activate your account and you can begin streaming to your audience instantly, at any scale.
The time, money and risk required to properly configure, deploy, scale and manage your own infrastructure is far more expensive than a hosted solution like Millicast.
To properly secure your streams and infrastructure from unauthorized access requires advanced knowledge and experience. With Millicast this is built into the service by industry experts who understand the importance of securing your streams.
— Future proof
Regardless of changes to device support, browsers or protocols, updates to the service are updated and deployed for you with no extra cost, time or attention required.
Want to try the lowest latency streaming in the world?
Sign up for Millicast today and get started with a 30 day free trial.