Multitrack transcription, automatic speaker labeling, and more in Descript 1.3

Andrew Mason
Descript
Published in
2 min readMar 14, 2018

Descript 1.3 is now available, and it’s a big one.

Multitrack Transcription with Automatic Speaker Labeling

We’re thrilled to be the first automatic transcription service to offer multitrack transcription. If your audio recording has speakers on separate tracks, this is a huge deal.

First, it means you get automatic speaker labeling with perfect precision. Second, by isolating voices, you get more accurate transcripts.

Finally, you can export your session while retaining the separate tracks — which means Descript can now be used in the editing workflow for anyone making multitrack podcasts.

We’ve got many more multitrack features to come, but this is an exciting start, so give it a try, it’s a huge step forward for automatic transcription!

Project Search

Search across all of your Compositions and Transcripts

Paragraph Timestamps

We added a timestamp indicator to each paragraph.​

​Word count / duration indicator

Descript now shows you the word count and duration of your current selection.

Shortcuts to edit audio while in Correct Text mode

Attention power users: now you can edit audio in the script without leaving Correct Text mode!

  • Option-Delete to delete audio
  • Option+Command+V to include audio in paste

With this feature, you can use Descript exclusively in Correct Text mode, if you want. If you care about this sort of thing, we’d love your feedback.

Improvements and Bug Fixes

  • Support for Team accounts
  • New warning when Descript detects an invalid alignment
  • Warn Dropbox users with large files about file corruption risks
  • Allow import of mp4 audio-only files (rename to .m4a)
  • Fixes an issue where speaker labels could move around when editing text
  • Improvements to error messaging when a transcription fails during compression
  • Fixes to inline notes that sometimes caused documents to break
  • Fix for a crasher when deleting ‘room noise’ at the start of an audio region
  • Ensure that new compositions added from the sidebar are inserted correctly into the folder tree
  • Fix for an issue where the WE playhead would not ‘play’ for initially blank compositions
  • Fix for an issue where tabs would sometimes create region breaks

--

--