Automated speech recognition – whisply
Automated transcription extracts language content from common audio and video files in text form for further processing and use. The Research Data Centre supports you throughout the entire workflow.
With the whisply tool, based on whisper, you can process and transcribe large amounts of data in a short time via the University Library's servers. The whisply tool is language agnostic and can transcribe a wide range of languages.
Other whisply features include:
- Automatic annotation of speakers and speaker changes
- Transcription output as a .txt or .rttm file
- Automatic creation of subtitles for videos in the .srt and .webvtt file formats
Instructions for installation and usage can be found on whisply's GitHub page.
Services
The FDZ can support you in areas including:
- Advice on transcribing multimedia content
- Setting up the transcription workflow
- Customise whisply's output formats
- General advice on audio-to-text
- Support for further processing of the transcription (e.g. transformation from unstructured to structured data)
