Descript Transcription is accurate, fast and provides the foundation for the editing workflows that make Descript such a powerful media creation platform.
While there are a few ways to generate a transcript for your recordings, this article will focus primarily on Automatic (AI-based) transcription. Check out our other articles on White Glove transcription and Importing a transcript for even more ways to add a transcription to your media.
Adding a file to the Composition
If your file contains more than one speaker, click the Detect Speakers checkbox to specify the number of speakers in the recording. This helps Descript narrow down sections of audio that may contain different speakers for later identification.
Correcting the transcript
AI transcription is great, but it can occasionally make mistakes so the following will help you quickly resolve any spelling, punctuation or capitalization issues you may encounter.
When you've located a portion of the script that you would like to correct, highlight the text and then press the Correct button in the popover dialog, or press the letter E. Make your corrections in the text box then press Correct or the Enter key to apply the changes.
Even better, Descript will pause when you correct a word during playback and then automatically resume from the point of correction once you're done.
You can also quickly adjust punctuation by typing it into the script or holding the W button and clicking a word to alternate between a space, comma and period character.
Similarly, you can hold the Q button and click a word to toggle capitalization on or off.
You may find over time that AI transcription will consistently have trouble with a particular word or name causing mis-spellings throughout your transcription. The Transcription Glossary allows you to specify tough-to-translate words to help improve the chance that they will be translated correctly.
Words that have been added to the Transcription Glossary won't change transcriptions you have already completed, but they will be available for future transcriptions across all projects on your drive. When possible, add difficult words or names to the glossary before transcribing your file.
Speaker Label shortcut
Speaker Detective is great but may not always catch every speaker or format your paragraphs perfectly. If you're looking for a quick way to make a correction, or if you prefer to add speakers manually, click the section of text where you'd like to add a newline / label and press Command + R (Mac) / Control + R (Win).
We also have a whole article in the help center dedicated to working with Speaker Labels if you want to become a Pro so feel free to check it out.
Transcribing Multitrack Recordings
Descript even allows you to transcribe two or more audio tracks simultaneously and combine them into a single unified transcript for easy editing. To do so, drag the media files into a Composition.
If the files are close to the same length, you'll be prompted to specify if you would like to either keep the files separate, or create a multitrack Sequence. Click the Create Sequence button so Combine the files into a unified transcript.
Next you'll have the opportunity to specify labels for each of the files. If the files are already labeled the way you'd like them to be displayed in the Speaker Labels, you may proceed. Otherwise type the labels for each file in the "Add speaker" text box, select the type of desired transcription in the dropdown selector, then click Transcribe.
Frequently Asked Questions
Not a problem though, you can just hover over the script in your composition, click the dropdown arrow that appears on the right and select "Transcribe".
If the file has not yet been added to your composition, you can also click on the Media Library, select or right-click on the file, then choose "Transcribe file".
What are the solid blue underlines that appear under some words?
You may occasionally see light blue underlines under some words such as "um", "uh", "like" and "you know". These words known as "Filler Words" are contextually identified and can be automatically ignored or removed with our automatic Filler word detection feature.
What is the small dotted line that appears after making a text correction?
Every time you make a change to your transcript, Descript will map the new text to the corresponding location of the media file. We call this Alignment and it assures that your text, captions and more are properly synchronized.
Once the alignment is complete, the dotted line will disappear.
Can I manually adjust the alignment of my text?
Yes! Our article Manually adjusting word boundaries provides details on how you can adjust the timing of words in your text.
Why does Descript say "No Sequence at Playhead" when I try to edit my sequence?
A Sequence is created when you combine two or more audio tracks into a Multitrack Sequence. This message occurs when your playhead is located over a standard media file that is not a sequence file. You can tell if your clip is a sequence file by the sequence icon in the upper left-hand corner.