How to Send an Extra Language Audio Track to a DCAST Live Stream (OBS, vMix, Larix)

On DCAST, every extra language in a live stream is a separate audio feed sent to its own key. Your main encoder keeps sending the picture and the original sound; each interpreter sends only their voice. This guide shows how to create a language in DCAST, and how an interpreter sends it from OBS Studio, vMix or the Larix Broadcaster app.

Language tracks work for streams sent over RTMP or SRT. A language track always uses the same protocol as its stream.

Step 1: create the language in DCAST

1. Open your stream's settings and go to Feeds → Languages.
2. Press Add language. A new track appears with an automatic name such as "Language 2".
3. Change Name viewers see to what your audience should read in the player — for example "Español" or "Deutsch".
4. Repeat for each language (up to 16).

Create every language before you press Start. The interpreters' feeds can arrive later, but the list of languages should be complete when the broadcast begins.

Each track card shows its own connection details:

Send each interpreter only their own track's details.

Step 2: the settings every language feed should use

Only the audio of a language feed is used; DCAST ignores any picture in it. These are the recommended values (also shown in the i tooltip, Recommended stream settings, next to the Extra audio tracks heading):

Never send full-quality video to a language key. It wastes the interpreter's upload and is thrown away on arrival.

OBS Studio

OBS always sends a video track, so the trick is to make it tiny.

Interpreter on their own computer (the usual case):

1. Settings → Video: set both Base (Canvas) and Output (Scaled) resolution to 320×180, and a low frame rate such as 10 or 15 fps.
2. Settings → Output (Output Mode: Simple or Advanced): set Video Bitrate to about 200 kbps and Audio Bitrate to 160 kbps.
3. Settings → Audio: set Sample Rate to 48 kHz.
4. Add the interpreter's microphone as an Audio Input Capture source. A plain colour source or an empty scene is enough for the picture.
5. Settings → Stream → Service: Custom.
   • RTMP stream: Server = the server URL from the Languages tab; Stream Key = the track's key.
   • SRT stream: Server = the track's full SRT URL; leave Stream Key empty.
6. Press Start Streaming.

Interpreter in a booth at the venue, on the main streaming computer: OBS streams to one destination per instance. Either run a second OBS instance with the --multi launch option and its own profile and scene collection (on Windows, a portable install keeps it fully separate), or use the third-party Multiple RTMP outputs plugin, which can give an extra output its own encoder, its own small resolution and its own audio mixer track. In OBS, assign the interpreter's microphone to that mixer track in Edit → Advanced Audio Properties, and keep it off the main stream's track.

vMix

vMix routes microphones to audio buses (Master and A–G). Put the interpreter's microphone on its own bus — for example Bus B — and keep it off Master so it does not leak into the original sound.

RTMP stream:

1. Open Stream Settings and add a second destination.
2. Destination: Custom RTMP Server. URL: the server URL from the Languages tab. Stream Key: the track's key.
3. Untick Use Stream 1 Quality, and in this destination's quality settings choose a small size and a low video bitrate, AAC audio at 48 kHz, and select the interpreter's bus in the audio channels list.

SRT stream (vMix 4K and higher editions):

1. Settings → Outputs, open the settings of a free output and Enable SRT.
2. Type Caller; host, port and Stream ID from the track's SRT details; latency and passphrase as shown in DCAST.
3. Audio Channels: the interpreter's bus. Choose a downscaled resolution to keep the video small.

After starting, check the track's signal dot in DCAST and listen to the language in the player to confirm the right bus is on the right key.

Larix Broadcaster (iOS / Android)

Larix is the simplest option for a remote interpreter: a phone, a headset and a mobile or Wi-Fi connection.

1. Settings → Connections → New connection.
2. RTMP stream: URL = the server URL from the Languages tab, then /, then the track's key. SRT stream: enter the track's SRT host, port and streamid in the SRT fields.
3. Mode: Audio only.
4. Settings → Audio: AAC, 48 kHz, around 128–160 kbps.
5. Make sure this is the only active connection, and start streaming.

If the track's signal dot in DCAST stays grey in audio-only mode, switch the connection to Video + audio with the lowest resolution Larix offers and try again.

Step 3: rehearse with every interpreter

Before the event, connect everyone once:

1. Start the main encoder. Language feeds may connect before, together with or after it.
2. In Feeds → Languages, check that each track's signal dot is green.
3. Open the watch page as a viewer. The language button appears on the player bar once there are two or more soundtracks.
4. Switch through every language. Check that the right voice is on the right name, and that levels are similar to Original.
5. Ask each interpreter to speak while the original is loud, and confirm they can still be understood.

Step 4: on event day

• Run the multilingual event as one continuous broadcast. The replay keeps the languages, but if you press Stop and Start again on the same stream, it offers the extra languages only for the part after the last Start. For a break, keep the broadcast running with a holding picture instead of stopping.
• Record interpreters locally if you need separate audio files of each language afterwards. The replay in the DCAST player keeps the languages; a local recording gives you files of your own.
• Give interpreters a clean return feed of the original sound (for example a separate listen-only stream or a conferencing link), so they are not interpreting from the delayed public stream.

Troubleshooting

The signal dot stays grey. Check that the interpreter uses the key (or Stream ID) from their track, and the stream's server URL — not the main stream key.

Viewers see no language button. The player shows it when a stream has two or more soundtracks. Check that at least one language feed is arriving.

The language is too quiet. Raise the interpreter's microphone gain in their encoder; aim for peaks a little below full scale, similar to the original.

I hear the original under the interpreter. That is usually intended, as a quiet background. If you do not want it, remove the original sound from the interpreter's encoder.

A language went silent after I stopped and started the broadcast. Check that track's signal dot. If it is grey, stop and start streaming in the interpreter's encoder with the same details. Next time, keep the multilingual broadcast running continuously: the replay offers the extra languages only for the part after the last Start.