Breaking Down AI Speaker Text Colors (Descript)

Tutorial-style YouTube video explaining the meaning of different transcript text colors in Descript. Covers color states for unassigned text, AI speaker assignment (loading/pulsing), generated-but-not-yet-converted AI speech (draft mode), converted/normal audio, error states (orange underline), overdub editing (strikethrough → pulsing → dotted underline), and missing-trained-speaker gaps (yellow).
When you're creating your Descript project, you'll see a lot of different colors in the transcript. So what do they all mean? Well, I'm going to be looking at this help article on help.descript.com and I'll leave the description or the link to this page in the description below, but I'm going to run through these. So if you see light blue text, it'll look like this on your timeline. And if I click this, I can expand it. So light blue text like this means that there you've written the words in, either you've pasted it in or typed it in yourself. But no speaker has been assigned yet. So this add speaker button, you'd have to click on that and then assign it to an AI speaker. So that's what the light blue is moving beyond that, if you go ahead and assign one. So in this case, they assign the name Dan. And while it's assigned, it's going to pulse like this. You see how it's going light to light to dark bluish, dark grayish. That's what's happening when it's loading, it's processing that speech, that text into speech. Next is the dark blue text. So this is when AI speech has been generated and that you're not currently in write mode, but it hasn't yet been converted to audio. So you can think of this as like a draft. So there's gonna be certain limitations to what you can do. You're not going to be able to edit this yet. You can't change the punctuation or the capitalization. You can't enter correct mode. You can't resize the boundaries of this clip until you go into your timeline and click on a button for that clip that says convert to audio. Moving beyond that, if it's black, that means it's been converted to speech, it's normal. And if you have any other video that you've brought into descript or you've done descript recording, it's going to be black. So black is telling you that this is a normal sound file. You can enter correct mode on it, you can delete it, you can change the boundaries of it. Anything you can do to a normal sound file, you can do when it's looking black on the timeline. If it's black with an orange underline, this means that the AI speech is having an issue. Something hasn't worked properly. It could be that. Oh, and it's not loading that image. But it could be that your Internet connection got interrupted. As it says There or for some, some reason, maybe their servers are down, something's not working properly and it's not able to generate that text. Moving on to overdub text colors with Overdub. Overdub is when you highlight a few words or a few sentences and then you enter this right mode where you can change the words that are spoken there. And then when you click Overdub, it's going to convert it into your AI voice. And when you have this box open, you'll see a strikethrough over the words. It's going to turn like a light gray or a lighter gray and be stricken out. And that's telling you that that is currently being edited. That's the section that basically you have selected and is what you're going to edit when you're generating it. When you've made the changes and you hit Overdub, it's going to pulse just like we saw before with the AI speaker. And you can see here on this little sample, it's going to pull us just like that. Go from light gray to darker gray until it finishes. Once it's finished, it's going to look like this black text with gray dotted underline, just like this. Oops. And it can't load it, but you can see it there. So again, this is kind of like the draft mode. It hasn't yet been converted to audio. And then moving down the next one is if you see yellow text with a gray dotted underline. This is telling you that AI speech or overdub speech has not been generated for that section because whoever you've labeled it as in this case, this person, Dan, there is no AI speaker trained for this person and therefore it's not able to turn this into AI speech. And what's going to happen is that's going to just be a gap clip on your video. There's going to be no associated sound in that section. And there's also gonna be this little icon of a person so that you can click on that and train and train or add an AI speaker to then convert that into AI speech. And then lastly we have the black text again, as with before, our black text means that this is normal audio. We can change the boundaries of the clip, we can change the audio levels, we can capitalize and change punctuation and do everything we can with normal sound files. If you found this useful, check out this video.










