Creating a talking head scene

Selecting a scene design to reveal the talking head editor

To create a scene with captions, first create a scene! You can do so by dragging and dropping the media content that you want captioned onto the canvas, or by clicking on the large green + on the canvas.

If you drag and drop your media onto the canvas to create your scene, we’ll detect that your media has narration and automatically choose a caption scene design for you.
If you choose to create a scene by clicking the large green + icon, you'll want to select one of the scene designs under the Automatic Captions header:



Once you've picked the right scene design, drag and drop your uploaded video onto that scene.

There are a few designs to choose from in most templates, and you can always select a different one by clicking on Swap design if you decide that you want the base of your talking head scene to look a little different.

Generating captions

Once you've selected the appropriate scene design and added your media, we'll generate captions for you automatically! This can take a minute or two, so please be patient:



You might notice that when we’re generating captions, another tab appears on the left side! The Transcript tab is where you’ll edit your captions, and how you’ll navigate if you want to add any overlays, call outs, or cutaways to your talking head scene.

Why are my captions very inaccurate or missing sections?

We use Google's transcription system (the same as YouTube!) to automatically detect the spoken language and create the transcript. If your captions appear incorrect, this could be for a number of reasons.

We may have detected the wrong language. To fix this, look for this pop-up near the bottom of the transcript tab and click on Does something look wrong?:



Once you do that, you'll see a drop-down menu to select the correct language:



After selecting the correct language, we will regenerate the captions for your video using the language that you selected. Note that this feature does not translate your captions, instead it'll try to create them for the language that you've specified. So if the person is speaking French, you should select French here.

If entire sections of your captions are missing, this may be because we weren't able to understand the speaker. You may be able to fix this using our audio enhancement feature which is available on upload.

You can correct mistakes in individual words using the correction feature (see below for how to do that!).

If even after all these steps entire sections of the transcript are missing, this could be because the speaker can still not be understood. We recommend in this case to record a new video, making sure the person is speaking clearly and that you are in a location with minimal background noise such as traffic and air filtration systems.

Editing captions

Once the captions are generated, the first thing you'll want to do is go through and double check the spelling, punctuation, and overall accuracy. Our robots are pretty good, but not perfect!

The captions that our AI generated automatically will appear in the Transcript tab on the left side. In this tab you’ll find the style controls for your captions at the top, so you can edit the text color, size and font:



This is also where you can turn captions off if your scene doesn’t need them! Just flip the toggle and watch your captions disappear.

To edit the caption text itself, click on the word you’d like to edit, then click on Correct. You’ll be able to make any adjustments that you need to by typing right in the transcript! It works the same way if you need to delete any extra words: just click on them, click Correct, and delete.

Note: unlike the Story tab, changes you make in the transcript tab get applied to your scene automatically!



Need another version of your captioned video in a different language? On the Professional or Enterprise plan, you can do that automatically! More here.

Updated on: 06/06/2023