If you're hearing echo or delayed audio, or if your transcript looks fragmented or duplicated, you might be experiencing mic bleed. This happens when one speakerâs voice is picked up by multiple microphonesâusually when a speaker isnât wearing headphones and their voice is captured by their own mic and another participantâs.
This article walks through how to resolve mic bleed using both automatic and manual methods, along with best practices to avoid it in future recordings.
Automatically fix duplicate transcriptions
If Descript detects mic bleed when a sequence is created, a message appears in the bottom-right corner of the editor: âSections of (your_filename) have duplicate transcription, perhaps caused by mic bleed.â Click Fix to automatically remove duplicated text in the script.
Mic bleed detection works best when each sequence track contains a single audio clip. If untranscribed tracks are present, Descript can still detect mic bleedâbut alignment accuracy may be affected if those tracks are later removed.
The Fix duplicate transcripts feature does not delete or mute audioâit only removes duplicated text from the script. You may still want to clean up the underlying audio manually in the Sequence Editor.
Manually initiate Fix Duplicate Transcripts
If Descript doesnât automatically detect mic bleed, you can run the tool manually from the project files menu:
- Open the Project panel and make sure you're in the Files section.
- Click the
button to the right of the sequence.
- Select Advanced > Fix duplicate transcripts.
Revert the fix if needed
On rare occasions, Descript may incorrectly classify audio as mic bleed. If needed, you can revert the changes:
- Open the Project panel and make sure you're in the Files section.
- Click the
button to the right of the sequence.
- Select Advanced > Revert duplicate transcripts fix.
Select sequence tracks to include in your script
You can also control which tracks are included in your script using the Sequence Editor. This is especially useful when your sequence contains multiple audio tracks with overlapping content, such as mic bleed or duplicate recordings.
Excluding redundant tracks helps improve transcription accuracy and speaker labeling. We recommend isolating the cleanest track per speaker when working in multitrack sequences.
How to include or exclude sequence tracks
- Open the Sequence Editor.
- Click the ¡¡¡ icon on the track you want to modify.
- Select Include in script or Remove from script.