Skip to content
  • There are no suggestions because the search field is empty.

Can I use a multilingual audio source for my live stream or in-person event?

Yes, multilingual source audio allows a single live audio stream to automatically detect and switch between spoken languages in real time, while continuing transcription and translating into multiple target languages.

Overview

Multilingual source language support allows a single audio source to dynamically switch between languages during a session while maintaining continuous transcription and translation output. This feature is designed to work with both live Events and Services where speakers may change languages on the fly.

With this feature enabled:

  • The source language is not fixed

  • Language detection happens automatically during the stream

  • Transcription continues seamlessly across language changes

  • Output can be translated into any number of target languages, as subtitles and/or audio translations

This is ideal for:

  • Multilingual panels or Q&A sessions

  • Events with international speakers

  • Services where interpreters or speakers switch languages mid‑stream

IMPORTANT:  You must select Soniox as the ASR engine for multilingual source language support to work.

  • Multilingual source language will not function with any other ASR engine

  • If another engine is selected, language switching will not be detected

Configuration Settings

When configuring your Service or Event, set the following options:

Engine

  • Engine: Soniox

Source Language

  • Source Language: Multilingual

This enables automatic language detection and dynamic switching during the live audio stream.

Events:

Services:

Multilingual Source Label (Optional)

On Events you may optionally set a custom label for the multilingual source audio.

Examples:

  • Floor Language

  • Live Audio

  • Main Room

This label is useful when displaying or routing multilingual source audio alongside other sources.

Expected Languages

You may either:

  • Leave this set to Auto for full automatic language detection, or

  • Specify the dominant languages you expect in the stream

Specifying expected languages can improve detection stability in controlled environments.

TIP: We strongly recommend using the expected languages feature to give the speech engine hints about the languages expected in the source for best results.

Translation Outputs

Once multilingual source audio is enabled:

  • Transcription is generated in the detected source language

  • Translations can be produced into:

    • Any supported target language

    • Subtitles

    • Audio output (where enabled)

Multiple target languages can be active at the same time.

WARNING: Multilingual source language support has specific behaviors and limitations that are important to understand before using it in production.

Please review the details here: What should I know before using multilingual source ASR in live Services and Events?