Latency in video streaming is the delay between the broadcaster and the viewer. A reduced latency is important for time-sensitive video content such as live sports games, news, and interactive shows.
Latency is categorized into three categories as presented below:
| Low | High | Protocols | |
| Common Latency | 20 | 45+ | HLS & MPEG-DASH |
| Reduced Latency | 6 seconds | 20 seconds | HLS & MPEG-DASH |
| Low Latency | 2 seconds | 8 seconds | LL-HLS, DASH LL CMAF, RTMP |
| Ultra Low Latency | 0.2 seconds | 2 seconds | SRT, WebRTC |
Why does latency occur during video streaming?
Latency from a video source accumulates through various stages of the delivery, however the most significant cause relates to buffering of the video content through the player. HLS and MPEG-DASH recommendations for stability are to maintain three segments between 2-10 seconds in length, thus by default there is typically a 30 second (3×10) second delay between the video source and the viewer.
Buffering is required to solve transmission interuptions due to network jitter, and congestion.
Reducing Latency
Streaming with a latency of up to 20 seconds is what we label as Reduced Latency and is made possible through reducing segment lengths and providing a high quality network close to users, typically by a CDN.
Reducing latency from the traditional 30 seconds is only recommended where network conditions are optimal, and typically not recommended if your broadcast is delivered to a different city to the viewers.
MediaCP Cloud Video offers Reduced Latency features achieving a latency of approximately 18 seconds. This is achieved by delivering smaller segments over our global CDN operating close to your viewers.
What about Low Latency, or Ultra-Low Latency
Low Latency, and Ultra Low Latency can be achieved with emerging technologies such as SRT, and WebRTC however are not as widely compatible as HLS.
