With Shorten Word Gaps feature, the user should hv the option to respect punctuation.

Given Descript's great ability to transcribe, it is astonishing how Descript missed this opportunity in the Shorten Word Gaps feature.

As it is now, the Shorten Word Gaps feature is a brute force tool, and other video editors do the same thing.

I was spending so many hours fixing the "shortened word gaps" (I typically set it to 0.01 seconds, or 0.05 secs, depending on the speaker's diction). Eventually, I created an app that adds a 0.7 sec pause after each period.

I hv no special background in computers or creating apps. I just used a free RPA automation tool (TikFlow) to make a macro that finds each period in the timeline, then drags it to the right a bit.

My macro works pretty well, even though it's a bit clunky. I pay for Descript to save me time, then I spent so much time on each video that I decided to invest 20+ hours creating an app to save me from manually adding 500 pauses into a 30 minute video.

I also created a macro in a different free RPA program (Automa) which removes 200+ filler words and phrases (Descript filler removal is limited to 20 or so). I spent a lot of time learning the Automa Chrome extension, and my app runs through a video in Descript and, one at a time (you can't do Find/Replace audio/video in Descript in bulk, it must be one at a time) finds every instance of a filler word/phrase in my list and removes it (both audio and video).

I hope Descript will return its focus to what it was great at. It seems to me, this "Shorten Word Gaps ... while respecting punctuation" feature is a no brainer. Descript can shorten all pauses. Descript knows what a period is, and knows when a pause follows a period.

Why can't Descript is great at BOTH those things. Why can't it combine those two things it's great at? It wd save me, one user, more than a hundred hours.

Even without AI, you cd programmatically set a rule: IF the gap follows a period, THEN ignore. You don't even need AI to add this fantastic ability.

If you set it to 1 sec gap, or 2 sec gap, it doesn't fix the issue. Some speakers pause 3 seconds in the middle of a sentence and race to the next sentence, then pause 3 seconds in the middle of that sentence.

All silent pauses removed, and the audio/video will sound natural, will give the viewer/listener 1 or 2 secs to process each sentence.

Yes, if AI can remove the music from a song and keep only the words being sung, Descript should be able to remove a filler word seamlessly. They just need not to cut out the word, but to more it leaving the background intact, like in the said music removal.

I would like to add my voice to this. The most recent update, in particular, is absolutely terrible for this. Horrible. That you moved the "remove repeated words" and other items into the "Remove Retakes" has made this platform useless for my company as we can edit faster, cleaner, with better results than Descript. You seem bent on releasing lots of things quickly rather than releasing quality updates slowly that provide excellent results. We can no longer trust Descript for quality output.

This has 400 more upvotes than an other feature requests on the roadmap and still just Under Consideration. I'm genuinely curious why this Descript has so much difficulty with this when im pretty sure other text-based editing apps can do this.

Also goes for the "Trim & Remove Silences" tool. Why can't Descript AI identify laughter? If you want to keep host/guests laughter, it kind of makes the "Trim & Remove Silences" tool useless.

Totally agree I would rather have a filler word kept than an abrupt removal. a smoothing video algorithm would be super cool... now that is REAL AI.

Yes! Or allow us to "preview listen" the edits before they are cut to make sure they're a clean cut