Real-time demo

February 14, 2017 9:57 am

Transcribe your voice to text in real-time using the new Speechmatics speech recognition system (beta) – in many languages.

The technology is designed to be consumed on-premises, on-device or in the cloud. Please contact us for more information at:

To give our real-time technology a go for yourself – simply select your language and get talking. But remember to check out our ‘how to get the best out of real-time speech recognition’ information for optimal results using our free trial demo.

Real-time speech recognition is limited by the quality of the audio presented to the system.

To get the best out of our system – follow the simple steps below.

  • Speak clearly however, there is no need to speak slowly
  • Ideally, your voice should flow over the microphone, not directly into it for optimal results
  • Literal translation of people ‘thinking out loud’ is very hard to read, so – if possible – think about what you want to say before you say it
  • If you have time, record some test audio and play it back to check it all sounds okay before you try it live
Your environment:
  • Record in a quiet environment to minimise background noise
  • Try to avoid multiple people speaking at the same time
  • Minimise reverberation. Sound can sometimes bounce off flat walls and muddy the signal
Your technology:
  • Use a good microphone, such as a USB noise cancelling (or directional) microphone – a head mounted microphone is preferable
  • Record at 16kHz or greater if possible
  • There is no need to compress the audio, if you do, please don’t over compress – use 96 kbps AAC or 128 kbps MP3 or better
  • Use two channels if available