Removing repetitive sounds like umm

I am editing an old recording and would like to remove the ums and errs my parents were prone to! I am doing so manually at the moment. Is there any universal search and delete feature similar to word processing where I could this more efficiently?

No there isn’t. Computers don’t understand speech, so it is very difficult for a program to distinguish between an unwanted “umm” and an intentional “I’m” or “him” or “mmm delicious”. I’m not aware of any software that can do universal search and delete on vocal sounds.

One thing that you can do is to look at the waveform. When at an appropriate zoom level (Zooming Overview - Audacity Manual) you should be able to visualise the words - look for words that have a longer than usual gap with a low level noise between the words and you will probably find a “umm” or an “err”.

Thanks Steve, I rather thought that might be the case but worth asking! I have been looking as you say for a particular wave pattern and that does work but is rather laborious; worth pursuing though. Thanks again


I’ve tried several things to get rid of such audio ‘noises’ using an automated approach. This video was helpful in some respects for removing breaths and gives you insight into using AutoDuck and some other features of Audacity.


Steve’s Dynamic Mirror plugin, (on mirror, rather than duck),
would remove the need for a control track in that workflow.

Thanks Trebor, I’ll take a look!