Only remove silence from "Shorten Word Gaps" (Strip Silence)
Bryan Schiele
If there's laughter or unknown words after transcription, the Shorten Word Gaps feature wants to remove those sections, leaving a choppy edit if you wanted to select "Apply to all". For example, in this screenshot, the name "I-17" (a road here in Arizona) isn't recognized and would be cut if I used the "Apply to All" option.
T
Thomas Craig
I've been asking for this for 2 years now. The fact that Descript treats "unknown" as silence is bonkers. It should be based off the audio input signal. Silence has nothing to do with transcription.
Weldon Johnson
Want to be the first person to post about this in 2025. Still can't believe they don't have this. It would save me so much time and everyone else. Hours a month. I shouldn't have to pay extra for this (audacity does it for free) but I would.
S
Support Team
Merged in a post:
Strip Silence in Sequence Editor
J
Jonathan Bumhoffer
In sequences with with multiple speakers and tracks, I would love a strip silence feature - protools has this feature - analyze the tracks and completely mute the track when no adequate signal is happening (like the person isn't talking), like in a podcast when the other speaker is talking it would completely mute the audio of the person who is not talking.
Even when no audible audio signal is coming through there is still a noise floor - audio level created by the components and also ac/heat and room noise, breathing, mouse clicks, etc..
This would reduce the step of dumping all of the tracks into a DAW just to strip the silence and then have to export them out and dump them into descript.
Belief Hole Podcast
How is this request not on the roadmap yet? I never use this this tool - instead I do it Reaper using gates, which works much better, but would be so nice to have this feature working as part of the Descript process.
Canny AI
Merged in a post:
Removing Silent Parts
Utku Sen
I record podcasts in Turkish, and for the most part, everything goes smoothly. However, there are times when I need to use English sentences during the podcast. Unfortunately, the transcriber gets confused and is unable to transcribe what I say. While this is understandable, the problem arises when I use the "Shorten word gaps" feature, as it removes those English parts since the words are not recognized. This poses a significant issue for me, as I am unable to use that feature. It would be helpful to have an option to remove silent parts (not depending on the transcription).
Mark Frost
Not having this adds hours to my workflow. Its so frustrating that such a simple piece of software like audacity can do it but descript insists on its over engineered 'word gap' feature which sucks.
D
David F
Any update on this? I will cancel my subscription because of this.....
C
Chris Ridgeway
This outstanding issue (for 3 years?) still means that we never use the word gaps feature. It's a selling point but not useful. It cuts into laughter or wanted non-word dialog, and does it so often we wind up undoing most of the edits if we do it in batch. And transcription errors with it missing the beginning or ends of words (still very common for us--say one or two transcription mistakes a minute), means that the word gaps tool destroys all those as well, meaning I've introduced hundreds of errors into the audio i have to go back and check.
If this actually did silence only, it'd be miles more useful.
If somehow it used AI to be super smart, I'd love that too, but please give us this great feature to use but can't!
A
Aaron Bay
Is this still planned? Right now this is probably my most frustrating issue. I can't trust descript to "Apply to all" with word gaps because then I spend almost as much time going through and undoing a bunch of them. Also I see this is in Descript Classic, but this is still an issue in the new Descript as well.
Belief Hole Podcast
Aaron Bay Yeah I thought this had been moved to the roadmap of updates as well? Maybe they gave up on it?
Kimberly Bridges
Descript also is already really good at recognizing mic bleed, and if it stripped those sections from tracks Non Destructively it would be helpful for the editing process. If you strip silence in another DAW to then import into Descript, you can't pull the audio back if the strip cut off the tail of a word.
Load More
→