Media Transcription

Areas of application: TV, podcasts, World Wide Web


  • → Automatic transcription of spoken language used in TV programs, podcasts, etc.
  • → Speaker change detection
  • → Analysis of media data (categories, patterns, conversational features, advanced keyword spotting)

Unique selling propositions:

  • → Domain adaptation by solution partners or users
  • → On-premises solution (data protection)

Our system automatically transforms spoken language from various media sources (TV news, podcasts, lectures, etc.) into text. It recognizes different speakers and marks the points where they take turns in the transcription. This is essential for creating minutes of oral exchange. EML Media Transcription has proved its worth in several pilot projects with partners such as the German radio broadcaster Bayerischer Rundfunk. At present the language portfolio is being expanded to include the most important European languages. Further languages will soon be available as well. With advanced analysis tools, archived media can be transcribed and searched on premises.

Link: Products - EML Transcription Server, EML Speech Mining Server, EML Transcription Workplace, EML Language Model Workplace