How this example was made
Whisper small estimated the word timings. The same unedited transcript is used here and in the kinetic example so their visual treatments can be compared.
Check word timing as well as spelling. This spoken sample illustrates the highlighting effect; long sung notes and loud backing music can need manual correction.
Rendering times are measured on the build computer for these short samples. They exclude upload, waiting in the queue and, unless stated separately, transcription. They are not a server speed or accuracy guarantee.
Original audio
Source text
Start with a small idea. Give every word a moment. Turn your sound into a story worth sharing.
Text shown in the video
Start with a small idea. Give every word a moment. Turn your sound into a story worth sharing.
Download captions ↓Source and measurement notes
These are public product demonstrations made from original text, synthesized tones and an eSpeak NG synthetic voice. They contain no customer uploads or commercial songs. English audio and on-screen text remain English in every page language.
Built with the same renderer used by the website. Measurements include one build run and can change with hardware, model loading and settings.
Published:
About this tool and its limitations ↗