Automatically detect language

Preview

This feature is covered by the Pre-GA Offerings Terms of the Google Cloud Terms of Service. Pre-GA features might have limited support, and changes to pre-GA features might not be compatible with other pre-GA versions. For more information, see the launch stage descriptions.

This page describes how to set up a recognizer to automatically recognize the language spoken in an audio file, from a preset list of potential languages.

In some situations, you don't know for certain what language your audio recordings contain. For example, if you publish your service, app, or product in a country with multiple official languages, you can potentially receive audio input from users in a variety of languages. This can make specifying a single language code for transcription requests significantly more difficult.

Multiple language recognition

Speech-to-Text offers a way for you to specify a set of languages that your audio data might contain. When creating a Recognizer, you can provide one or more languages that the audio data might include in the languageCodes field. When you then use the Recognizer in a transcription request, Speech-to-Text attempts to transcribe the audio using the best-fit language from the list of alternates you provided. Speech-to-Text then labels the transcription results with the predicted language code.

This feature is ideal for apps that need to transcribe short statements like voice commands or search. You can list up to four languages for automatic language recognition.

Caution: Though you can specify up to four languages, constrain the language list to the bare minimum needed as a best practice. The fewer language codes you specify, the higher the likelihood that Speech-to-Text will successfully select the correct one.

Enable language recognition in audio transcription requests

Specifying multiple languages in your Recognizer resource works just like specifying a single language: simply add the new language codes to the languageCodes field. Speech-to-Text supports alternative language codes for all speech recognition methods: Recognize and StreamingRecognize.

Note: You can only use the alternative languages feature with the latest_long, latest_short, and telephony models. For more information about how to specify different models, see the Transcription Models documentation.