Why removing 'um' from a recording is harder than it sounds
A developer explains the technical challenges of automatically removing filler words like 'um' and 'uh' from audio recordings. The article details the creation of a tool called 'erm' that leverages OpenAI's Whisper model to identify and remove these disfluencies.
Why it matters
Automated audio editing tools can significantly improve the efficiency of content creators and podcasters by reducing manual post-production labor.
Linguists have a word for the um s, uh s, er s, and elongated versions ( ummmm , uhhhhh ) that pad spoken English: disfluencies .
The article is a technical explainer focused on software development and audio processing without political or social agenda.
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in