An SRT server is the managed contribution endpoint. Configure latency and bandwidth for the real network, secure it, then use telemetry and controlled routes rather than guessing from player symptoms.
开始之前
A reachable UDP port and a firewall rule limited to the required sources when possible.
A real SRT encoder or test source and a receiver such as VLC or a downstream Callaba module.
Measured network RTT and expected peak bitrate.
设置说明
仅解释操作员需要的控件；内部字段名和实现事件不会显示。
Listener
Identity and network capacity of the SRT endpoint.
- Name and active state
Operator label and lifecycle state of this SRT server.
- Listener port
UDP port used by incoming SRT connections.
- Receiver port
Optional distinct port used by the receiver side of the configured workflow.
- Latency
SRT recovery buffer sized for the measured network path.
Start from measured RTT and test under representative packet loss.- Maximum bandwidth
Headroom SRT may use for retransmission relative to the media bitrate.
- Receive buffer
Host buffer available to the SRT receiver.
- Connection timeout
How long an unavailable peer may remain unresolved before the session is treated as failed.
Security and access
Protect media and restrict who may publish or receive.
- Passphrase
Optional SRT encryption secret shared by authorized peers.
Use a strong secret and distribute it separately from the URL.- Access mode
Require explicit Stream access objects, restrict by host, or deliberately allow broader access.
- Publisher and receiver roles
Separate clients that send media from clients that consume it.
Routing and recovery
Connect this endpoint to one or more SRT peers.
- Routing mode
Disable routing or use Callaba/public-address peer definitions.
- PUSH or PULL
PUSH sends toward a listener; PULL connects to an available source.
- Routing hosts
Ordered peer addresses used by the configured route workflow.
Observability
Use live evidence before changing transport parameters.
- Statistics interval
Frequency used for supported statistics collection or callbacks.
- Live bitrate
Shows whether media is currently arriving.
- Network RTT
Round-trip time reported by the SRT session.
- Connection and event history
Session and transport events used to diagnose disconnects.
安全的首次工作流
- 01
Create the server with an unused UDP port and a clear operator name.
- 02
Measure the real path, then set latency, maximum bandwidth and timeout for that network.
- 03
Add a passphrase and explicit publisher/receiver Stream access.
- 04
Start the server and connect one real publisher.
- 05
Connect one receiver or downstream module and confirm live bitrate plus Network RTT.
- 06
Only after the single path is stable, add routing or a backup PULL source.
工作流示例
Create, verify, edit and retire an SRT server
An operator needs to change an SRT endpoint without mistaking a saved form for a working media path.
构建方法
- 1
Create the server with its role, port, access policy and expected media settings.
- 2
Start it and wait for the endpoint state expected by the publisher or receiver.
- 3
Connect real media and verify bitrate, picture, audio and the intended downstream output.
- 4
Before editing, identify whether the field requires a stop or restart; save the change, restore the server and repeat the same media test.
- 5
Before deleting the server, remove or reassign every route, recording, Multiview tile and output that depends on it.
Secure remote contribution into production
A field encoder sends a feed over the public internet.
构建方法
- 1
Create the SRT listener and size latency from the measured path.
- 2
Set a passphrase and create a publisher Stream for the encoder.
- 3
Connect the encoder and confirm bitrate, RTT and stable session state.
- 4
Select the verified SRT source in Multiview, Recording or Restreaming.
Primary and backup SRT contribution
Production has two independently reachable feeds and needs a controlled recovery path.
构建方法
- 1
Add the independently tested primary and backup hosts to the PULL routing plan.
- 2
Verify each source alone before enabling the ordered route workflow.
- 3
Watch both source health states in Multiview or the operator interface.
- 4
For planned switching, use the preferred-route control; for reconnect recovery, observe the active session and output.
- 5
Simulate a source loss during a maintenance window and record recovery behavior.
Multilingual browser event from one multitrack SRT programme
Production provides one SRT/MPEG-TS programme with multiple finished language or commentary tracks.
构建方法
- 1
Receive the multitrack SRT source and confirm that every produced audio track is present.
- 2
Listen to the discovered tracks in SRT Server statistics and assign operator-readable language names.
- 3
Create one Web Player per audience language; set Modify Audio Tracks to Select Audio Track and map the intended source index.
- 4
Add those players to a Language list group and choose the default language.
- 5
Test switching, access rules and playback in a clean browser session.
Multilingual browser event from separate SRT feeds
Production supplies each finished language as an independent SRT feed instead of one multitrack programme.
构建方法
- 1
Receive and validate one independent SRT source for every produced language.
- 2
Create one Web Player per verified feed and name it for the audience.
- 3
Add those players to a Language list group and choose the default language.
- 4
Test switching, access rules and playback in a clean browser session.
验证结果
- A real publisher and receiver remain connected for the planned test period.
- Live bitrate confirms media arrival and Network RTT remains within the tested latency budget.
- Unauthorized stream identities or hosts are rejected by the selected access policy.
- A backup-source test produces the documented recovery behavior.
故障排除
Session connects but media breaks up
检查- Compare bitrate to available upload.
- Check RTT and packet-loss behavior.
- Review latency and maximum bandwidth.
Increase recovery headroom only from measured evidence; also correct the source bitrate or network bottleneck.
Publisher cannot connect
检查- Check UDP firewall and public address.
- Match port, mode, passphrase and Stream identity.
Test one publisher without downstream modules, then restore the verified receiver path.
The backup does not take over as expected
检查- Verify both sources independently.
- Review route order, mode and timeout.
- Observe the active session during failure.
Correct the failed source or routing definition and repeat a controlled failure test.