Low latency means different things to different users with different requirements. Interpretation of what low latency really is also varies significantly over a wide spectrum. According to some experts, low latency can be achieved at anything between 10 and 2 seconds, but even then, there are terms and standards that measure that metric in different ways.
A recent example gave a hand wave test to explain how start-up time can also impact latency and measure how close to real time ‘live’ streaming can be. The demonstrator said, ‘I wave my hand on the camera and you’re watching on your device. How many seconds later do you see the waving hand?’ The start-up time and latency do correlate and relate, and this will impact the result when measured. Is it really that simple though?
Many experts agree that standard streaming is HTTP-based, commonly HLS. Standard streaming with ‘normal latency’ is often 30-45 seconds or more at best, and reduced latency, the industry goal for most, can be measured in a wide range between 18 seconds down to five. Although this is an industry goal, five seconds, according to most experts, is the typical latency for HD cable and not via the internet, and anything less than ten seconds is widely considered to qualify as low latency.
This means that in some circles, anything less than five seconds is to be considered real time, and in others, only less than one second can qualify as real time.
The point is that there is no set-in-stone measurement for low latency, and let us remind ourselves that a few industry professionals don’t even recognise the term. Over such a wide and significantly varied spectrum, it is difficult to pin down an industry standard. This is an area that is constantly evolving and across the industry, specialists are working towards improving latency.
The Streaming Company understands that low latency means different things to different clients with different needs, and for some it is a bonus, when for others it is a necessity. Talk to us today to discuss your requirements for low latency.