How to configure my SRT latency?

The SRT latency setting on each input and output stream is key to optimize the quality of service.
The right SRT latency settings will depend on the network characteristics between the chosen Link region and the SRT device.


750 ms is a good starting point if you don’t know where to start. This will ensure reliable delivery for most common use-cases.


Depending on your network conditions, you could also typically

  • Go down to 300 ms if you have a high quality internet connectivity and delivering to a cloud region close to where you are located
  • Increase to 1500 ms or even more if you are using a lower quality internet connectivity or delivering to a cloud region located very far from where you are


The SRT latency can be configured independently for each input and each output. This means that you can optimize the distribution independently for each output, based on different network conditions


Output latency display


Link exposes latency metrics directly in the product, so you can see the actual impact of your configuration rather than estimating it.
For each output, Link displays an estimate of the latency introduced by the service, with the total value and its breakdown across three stages:

  • Input — upload from your source device to the Link input
  • In-app — internal routing and processing performed by Link itself
  • Output — delivery from the Link output to the receiving device

This breakdown makes it easy to identify which part of the chain contributes most to your end-to-end latency, and therefore where adjusting the SRT latency setting will have the most effect.


Two things to keep in mind:

  • Latency information is only available when the output is linked to an input.
  • Output latency is not available for outputs using RTMP


Troubleshoot the SRT latency


Once you create an input or an output in a region, having a look at the SRT statistics is key to understand whether the connection quality is good or not.


SRT Stream Bitrate, Packet dropped, and RTT monitoring


For inputs, the "SRT Dropped Packet" should always be null or very low, as it counts the number of lost packets after the SRT recovery took place (hence, it basically counts the number of unrecoverable packets). Having too many dropped packet is a clear indication that the link quality is low. You should increase the SRT latency in this case. Increasing the latency will give more room to the SRT protocol to recover packets, and is likely to reduce the amount of dropped packets.


The SRT "Round Trip Time" is also very important and can give you a very valuable hint on how to adjust the "latency" setting of your SRT stream. A good rule of thumb is that your SRT latency should be at least 3 to 4 times higher than the RTT.

Updated on: 07/09/2026

Was this article helpful?

Share your feedback

Cancel

Thank you!