Speech To Text in LOPs (replaces LiveTranscribe)
This video covers three speech-to-text operators available in LOPs - AssemblyAI (cloud-based), Kyutai (windows only local streaming), and Whisper (local - streaming/push-to-talk/file processing). I show how to setup and why to use each operator, for a variety of ways to use real-time transcription in TouchDesigner. Please let me know if there is anything confusing or unclear. Excited to see what you build with this.
Recent changelog + updated docs - https://dotdocs.netlify.app/changelog/
grab LOPs >> TOX
Comments (3)
amazing work! my lops cannot see my 4090- how do i get the cuda working so i can run some live subtitles? :-) Error from SSTWhisper / Faster Whisper "[WORKER] Invalid handle. Cannot load symbol cudnnCreateTensorDescriptor"
video unavailable
ah! I had scheduled my youtube video for tomorrow at noon. My b should be up in 2 minutes