Skip to main content

Accuracy and Language

Transcription:BatchReal-TimeDeployments:All

For more information on the full range of language packs offered by Speechmatics, please refer to the guide on Supported Language Packs.

Accuracy​

Speechmatics offers two operating points for transcription:

  • Enhanced - Optimized for the highest accuracy possible.
  • Standard - (Default) Optimized for fast Batch turnaround time.

Note that when transcribing in real-time, latency for both Standard and Enhanced operating points is controlled using the max_delay config option described here.

Note that your choice of Operating Point will impact pricing, as well as the processing cost efficiency for On-Prem customers.

{
  "type": "transcription",
  "transcription_config": {
    "language": "en",
    "operating_point": "enhanced"
  }
}

Multilingual speech-to-text​

These packs are ideal when transcribing multiple languages in the same media file or stream with high accuracy. For more information on the supported languages, please refer to Supported Language Packs.

Supported multilingual packs are:

Language PackTranscription config
Mandarin and English{"language": "cmn_en"}
Malay and English{"language": "en_ms"}
Tamil and English{"language": "en_ta"}
Spanish and English{"language": "es", "domain": "bilingual-en"}

Bilingual (excluding Spanish and English) example:

{
  "type": "transcription",
  "transcription_config": {
    "language": "cmn_en",
  }
}

Bilingual Spanish and English example:

{
  "type": "transcription",
  "transcription_config": {
    "language": "es",
    "domain": "bilingual-en"
  }
}