Descript, manufacturer of audio editing software with artificial intelligence, is launching a new subscription tier called Descript Pro today that provides access to the overdub feature.
Overdub, first released in beta last year, is a synthesis tool that lets you effectively create an AI voice double for yourself. In this way, the software can fix a mistake in a shot while the end result still feels natural and organic. It’s a bit like creating a pure audio deepfake, just using it to fix your own mistakes instead of trying to pretend to be someone else.
According to Descript, Overdub is now an official feature when Descript Pro launches and is available when you pay for the subscription for $ 30 a month (or $ 24 a month billed annually). Before today̵
Besides Overdub, Descript’s main selling point is that audio editing is reinterpreted as a kind of collaborative Google Docs-style app, so you can edit the transcription of a podcast or other form of audio recording as if it were a Word document, and so on making the changes is reflected in the playback. Descript also acts as a video editor and transcription service. The latter includes AI and human-supported transcriptions that run for up to 30 hours for the Pro plan and 10 hours for the Creator plan with the option to purchase additional hours.
The company says its new Pro level also includes “exclusive access to advanced noise processing, custom audiograms, and flexible publishing and export options.” The treatment of noise words is a particularly useful sounding function, because according to Descript noise words like “you know”, “like” and “kind of” can be recognized, ignored and deleted to make an audio recording sounds smoother and more professional.
CEO Andrew Mason tells The edge“In our internal testing, we found that removing filler words from zoom meeting recordings reduced the length by about 10 percent.”