Feature

Audio Transcription

We use state-of-the-art machine learning techniques to transform the speech into text, and make it available to you.

How does it work?

  • Manage your application on Ziggeo and create a new meta profile.
  • Apply this meta profile to your video recorder’s by adding the meta-profile parameter.
  • Videos tagged with this meta profile are now automatically transcribed after being recorded.
  • Audio Transcription can be used as closed-captioning on the video player

Supported languages

We support the following languages:

  • nl-NL: Dutch
  • en-US: English US
  • en-UK: English UK
  • en-AU: English Australia
  • fr-FR: French
  • de-DE: German
  • it-IT: Italian
  • pt-BR: Portuguese Brazil
  • es-ES: Spain, Spainish
  • es-LA: Spain, Latin A

By default, Ziggeo’s audio transcription assumes US English. Set the language explicitly with the transcript-language parameter:

<ziggeorecorder … ziggeo-meta-profile=”…” ziggeo-transcript-language=“de-DE”></ziggeorecorder>
The same value can be set on a video through the API (transcript_language). It is an explicit setting, not something we guess: a value we do not recognise falls back to the closest supported variant of the same language, and failing that to en-US. Profanity masking (censor-profanity) is available in every language on the list above.

Example

Consider this example video: Here, the automatic audio transcription would yield:

I’m Elise and I’m a bit of a coffee snob some usually found in fancy coffee shops or out of the park with my dog.

Frequently asked questions

What happens if I do not set <strong>transcript-language</strong>?
US English (en-US) is used. The language is an explicit setting rather than something we detect, and an unrecognised value falls back to the closest supported variant of the same language, or to en-US. Set it to the language your speakers actually use — transcribing one language with another language's model produces confident nonsense.
Will audio transcriptions be shown as subtitles on my players automatically?
By default we are not showing them. This is only shown if you want to do that. To do that, you would add the audio-transcription-as-subtitles (boolean) parameter to your embedding.