1. What is SRTLA?
SRTLA sends SRT video over several network paths in IRL streaming. The upside: you no longer depend on a single cellular or Wi-Fi connection. How SRT works over one path is covered on the SRT page.
| Component | Job |
|---|---|
| Sender | a suitable encoder, such as Moblin or BELABOX with an SRTLA pipeline |
| Network paths | at least two connections that are as independent as possible, for example cellular from different providers plus Wi-Fi |
| Receiving end | SRTLA receiver or relay that turns the paths back into one video stream |
| Receiver | OBS takes in the signal and passes it on to the platform |
2. Setting up SRTLA: how the setup works
An SRTLA stream needs a compatible sender, an SRTLA receiving end, and a receiver such as OBS. If bonding should cover outages, use at least two internet connections that are as independent as possible, for example cellular networks from different providers.
- Prepare the sender and the network paths, and check each path for stability on its own.
- Set up your own SRTLA server or choose a managed relay service.
- Set the stream ID and target address in the sender, receive the signal in OBS, and test a real connection loss.
The technical steps are on the dedicated pages. This page stays at the overview level.
What does your own SRTLA server cost?
According to the publisher, the receiver container used here is licensed under CC BY-NC. For private, non-commercial use there is no license fee. You may not offer a paid SRTLA hosting or rental service with this image without extra permission. For any other commercial use, clear the permission with the rights holder. Either way, VPS rental or your own hardware, power, internet access, and maintenance are not free.
amd64, not these local setups.3. Reading RTT correctly with SRTLA
RTT stands for round trip time: a data packet travels to the receiving end and the reply comes back. With SRTLA, the cellular, Wi-Fi, or Ethernet paths in use can show different RTT values. A distant or overloaded path often responds more slowly than a close, stable one.
- Low, stable RTT: the path responds quickly and evenly.
- High RTT: feedback and packet retransmission take more time.
- Fluctuating RTT: often points to jitter, cell handovers, or congestion.
- RTT plus packet loss: read them together. A single measurement is not enough for a judgment.
SRTLA can use several paths, but the connection with the lowest RTT does not automatically take over the whole stream alone. The actual split depends on the sender, the receiver, and the state of the paths. Very poor paths can still drag the setup down and should be tested on their own.
4. SRT vs. SRTLA: the differences
| Area | SRT | SRTLA |
|---|---|---|
| Network paths | usually one connection | several connections at the same time |
| Main goal | reliable transport over one connection | redundancy and use of multiple connections |
| Receiver | ordinary compatible SRT receiver | dedicated SRTLA receiving end or relay |
| One path fails | the stream depends on that connection | other active paths can keep the stream going |
| Setup | simpler, with fewer components | needs extra networks, a receiver, and individual tests |
| Data usage | usage of the single stream plus protocol headroom | extra overhead, and possibly several (internet) plans in use |
| Typical use | stable single connection or controlled uplink | mobile IRL stream with changing network quality |
5. Why use multiple connections?
- LTE + 5G: different modems or providers can be available in parallel.
- Multiple providers: an outage or weak coverage on one network does not have to hit the whole stream.
- Wi-Fi as an extra path: in the right situations, another independent network can help.
- Relay receiving end: on the server side, the network paths are merged back into the actual video stream.
6. Your own server or a relay service?
If you want to host it yourself, you can set up your SRTLA server on a fresh Debian or Ubuntu server with our tested automatic installer. One install command walks you through the steps. Questions about updates, firewall, and SSH port stay interactive on purpose. After that, you manage the server and its access rules yourself.
If you would rather not run your own server, you can use a turnkey relay endpoint instead.
For a high-quality production IRL stream, we recommend looking at a professionally operated relay server first. Redefined Streaming offers ready-to-use SRTLA endpoints with low latency and tiers up to 50 Mbit/s or 2160p60. A stable transport can noticeably reduce visible blocking, gray image areas, and frozen frames on fluctuating cellular connections.
7. Example workflows
Smartphone with Moblin → two mobile network paths → SRTLA receiver / relay → OBS → platform.
BELABOX with a compatible SRTLA pipeline → cellular and Wi-Fi → SRTLA receiver / relay → OBS → platform.
The exact menus and target addresses depend on the sender and the relay you choose. Check the guide for your sender, and test the loss of a single network path before you go live.
8. Common questions about SRTLA
What do I need for SRTLA?
A compatible sender, an SRTLA receiver or relay service, and a receiver for the merged signal. Two independent internet paths make sense if you want the redundancy that bonding gives you.
Do I need my own SRTLA server?
No. You can use a managed relay endpoint. Your own server gives you more control, but you also take on updates, firewall, reachability, and ongoing maintenance.
When is SRT enough, and when is SRTLA worth it?
With one stable connection, plain SRT can be enough. SRTLA becomes interesting when you want to use several network paths at once on the move and a single dropped connection should not stop the stream.
How many connections do I need at minimum?
If bonding should cover outages, use at least two internet connections that are as independent as possible, for example cellular networks from different providers. Test each path on its own first.
Does SRTLA automatically use only the fastest connection?
No. The connection with the lowest RTT does not automatically carry the whole stream by itself. How the stream is split depends on the sender, the receiver, and the state of the paths. Test very poor paths on their own. And always read RTT together with packet loss.