Adobe Premiere Pro - Advanced Training

How to create automatic captions in Premiere Pro

Daniel Walter Scott

Download Exercise Files

Contents

Certificates

We’re awarding certificates for this course!

Check out the How to earn your certificate video for instructions on how to earn yours and click the available certificate levels below for more information.

You need to be a member to view comments.

Join today. Cancel any time.

Sign Up
Hi everyone, yes, you read right, you can automatically make captions, goes from speech in your Timeline down here, through something called transcriptions, which we know about, and then we can click 'Create Captions', and it's magic, watch, there was no typing, and we've got captions all the way through. It's automatic, it's awesome, it's in 13 languages, let me show you how it works. 

All right, to get started I've got a bit of dialogue, it's from an earlier project, remember Tourism Island, I'm not using that sequence, I'm just keeping everything clean here. So if you do want to bring it in, under 'Exercise Files', under 'Tourism Island', under 'Audio', this is the audio file that I'm using, 'Old Irish Blessing', you can see it there, I've put on my Timeline, and I've dumped a couple of videos from under footage in there. It doesn't really matter, it's all about the audio. 

So first thing we need to do is, let's switch our kind of workspace to 'Captions', it's a bit boring but it's going to bring up this chunk. If you can't find it, under 'Window', find it under Captions, no, it's called 'Text', there it is there. The way this works, remember, is that you create a transcript first and then make captions, so it's weird that way. So let's transcribe the sequence automatically, this is fancy and new. 

So 'Create Transcript', and basically which track is it on, you can mix it, so it's going to go through, all of the tracks but that's not good if you've got music playing, and I've got this weird audio that's on this bit of footage, so I don't need that, so I'm just going to say, actually, just use Audio Track 2, see this one here, it's got all the waveform on it, and that's it, this will all be off by default. I'll go through them in a bit more detail, but I bet you, 99% of the time that's all you're going to do, it's going to go off and do things. 

The one thing that I kind of skipped over there, is that it's actually uploading it to the Cloud, so your data is going off to the internet, and I guess Adobe wants you to know that it is going to the internet, it is secure on the internet, but you know, with your permission, it's uploading it and processing it in the Cloud, yeah, so there you go, so it is done, let's have a little look at what's created. 

So remember, a transcript is created at the moment, not a sequence, so let's play it. "Let the road rise up to meet you, may the wind always be at your back." Cool, huh. It's just like match it all up, you don't have to type it out, and like, the voice recognition is amazing, like this is really new, literally came out yesterday, I've updated this video, and so, I don't know, I think for a first version it's doing pretty amazing, with this accent in English, so you have to let me know in the comments, if you're finding it troublesome in other languages. It is, I think it's released with like 14 languages or something, so that's cool. 

So yeah, we've got a transcript, so what do we do with it? Well, nothing really, we basically create that transcript, so that we can then click this button, 'Create Captions', and again, just leave everything as the default. We'll run through a little bit more details, but 99% of the time, I bet you, this is going to work, and now we've got our little captions along the top, look at them here, ready? "May the road rise up to meet you, may the wind always be…" It is so good, if you've done this manually, or maybe sent it off to get done, and paying somebody to do it, man, it's pretty good doing it here for ourselves in Premiere Pro. Again, this tool is new, so I haven't really put it through much of its paces, I've been trying to, but yeah, I'm sure I'll find some limitations, but I don't know, "and rains fall soft upon your fields, and until…" it's pretty amazing. 

Now other things in this, let's have a little look. So let's say that we're going to go back and look at just a bit more details, because this is the advanced class, I'm going to select all the captions along here and delete them, because I don't want the captions at the moment. I'm going to go back to the transcript as well, see, Captions, Transcript, and at the moment there's no, like delete transcript, so you can re-transcribe the sequence and it kind of replaces what's there. So let's do that to have a little bit more detail. 

So this is one that says Audio Clips Tagged as Dialogue, just do those ones because you might have a lot in here, and you might have different clips and sound effects. You don't want the whole track done, you just want the bits that you've tagged as dialogue. We do that, remember, when we go, you, I've got 'Essential Sound', you're the 'Dialogue', and now this one will be transcribed, and automatch while we're there. So now when we go 'Read Transcribe', we can say, actually just do the ones tapped Dialogue, so you can individually pick the ones. 

Obviously, you can pick a language, there's a bunch in there, let me know, when you're doing your work, obviously, not this particular one, because it'll be very similar, but the work that you're doing, throw something into Premiere Pro right now for me, after this video, and give me a out of 10 score of like how good it got to it. It'd be good to see over time how good these things are, and how better they go. So English, other things, let's look at the most important other feature, is, this Recognize Different Speakers Talking. This seems really good, so I'm going to cancel out of that, at the moment I only have one woman talking, so I'm going to grab another bit of audio. So I'm going to say, from our 'Exercise Files', under 'Sound', there's one in there called 'Sound 02'. Now it could be one track with different people talking, I had a look through our exercise files and we don't have anything, and I couldn't find anything on my machine, so we're going to fake it by doing this, there you go. 

We'll pretend that this, my friend, is, just me and this woman having a conversation, so what you can do is you can say 'Re-transcribe', and you can say, all 'Tagged on Dialogue', or we could do Audio Track too, wouldn't matter, but this one here, it says, 'I want to opt in', to kind of know the speakers, who's who. One thing it says is that, you know, there are new laws, there are laws around kind of this biometric privacy, about recognizing people without asking, or face recognition and audio recognition, it's all weaved into here. 

Basically Adobe wants you to know, like, "Hey, you need to have permission, to be uploading these to our servers", and be trying to recognize them. They only recognize them as speaker 1 and speaker 2, but you know, it depends on how you feel about this. You've got to at least know you've got to have permission to do it, so I'm going to enable it. You'll notice in there, that says, you can't do it in Illinois, I think it's one of the states in America, that forbid this to happen, as part of their law. I think, Texas do it as well, that might change, that might go away, it might get worse, but let's hit 'Transcribe', and what it's going to try and do, is break out the two different speakers, even when they're in the same track, so that, yeah, you can identify the different people, and separate them out, rather than it kind of blending together. 

All right, let's speed this up. So it's got it, like, let's have a look. "Be at your back, may the Sun shine warm upon your face," "All right, so this is…" Great example, Dan. In here you can see it, speakers, and I can say, actually this is, Kathy, and this is, handsome Dan, or just ansome Dan, there you go. Now I can generate captions, and we'll have our two separate speakers. Now while we're in here let's look at some of the more details on creating captions. You can create a blank track, that is when you want to hand type it in, rather than getting it automatically done from the transcript we just made. 

Everything else in here is useful, I guess, there's Caption Presets, there are, captioning presets for different specific kind of use cases. I'm not really knowledgeable about the different outputs, so you might be, you might go into a special place, and it needs, you know, you might have Teletext, and you can click in here to get the right kind of captioning done. Format and Style kind of seems similar, about Format Again, it's to do with the kinds of preset, in the way that it's displayed on here, and again, with a stream it depends on what you're doing. So if the format changes, you know what you're doing with the format, you can pick the different kinds of streams, I don't, I'm kind of blagging my way through this a bit. In terms of the style, that's something I do know about, is-- we'll do style in another video, that's when you can change the fonts and stuff, pick a default style and update it in here. 

Other things you can do when you're creating captions, how many characters go into the words that appear on the screen, do you want it to be really long lines or really short lines, you can have a play around with that. There are obviously some industry standard defaults, that are in here but you can play around with them. What is the duration in seconds, how long do they stay on screen, you might feel like I'm speaking too fast, and it needs to be slower or faster, again, the gap between captions, when one disappears and the other one appears, do you want it to be 0, so they just flash up on screen, or do you want a bit of a pause between them? Does your captions display on the text, you know, on the screen, in a single or double line? All these things you can go and adjust, but the default's pretty handy. 

Other things that you might want to do, advanced stuff is, let's go back to, our 'Captions', actually, I'm going to delete these captions again, go back to my 'Transcript', you get your transcript looking good first, and then make your captions, it's easier process. So in here I'm going to recreate them, and what you might do is set an in and out point, say you only want captions for this part, so I set my in and out points, which is my I on my keyboard, and O, wherever your Playhead is, then I can re-transcribe and I can say, actually, just do the in and out points, because you might have a very long document, I'm going to turn off the different speakers. 

The other option in here is that you might merge with an existing transcription. I couldn't think of a way that this would be useful, so you have a transcription instead of like redoing it again so it's new, you can add new audio, and then smoosh it with the existing one, I don't know when that's useful. You probably can think of it already, I can't, but you can do that. Let's cancel that just because the other options you might look at in here, we've looked at Re-transcribe. 

Now when you export your transcript it feels like, "Oh great, I'm going to export my transcript because I want some text." If you do this you end up with this really weird format. I'll put it on to my poor old desktop, my desktop's not in here, where are you, desktop? There it is. Let's stick it in and let's see what we get. Now you're going to get this weird file that you don't want, so again, Daniel Scott, Desktop, lost my shortcut, You get this one, pr, you know, prtranscript, it's a weird file, like if you try and open this in something, you get this, I can't work out something that actually, you know, that actually works with. 

So you might know, you might be like, "Oh great, it's the prtranscript that I've been looking for," that's how you do it, but what you probably want and what I want, is Export to Text File, then you export Plain Text, and you get this kind of file, not that one, you want 'Export File', and you get-- let's stick it on the desktop again, let's have a look, and that transcript.txt file, and that's probably more what you want, then you can use that, I don't know, for a text transcript for your website, or maybe going out as dialogue to people, but that's another thing you can do once you've got your transcript. 

Other things out there, pretty self-explanatory, Pauses, you can, instead of just having like, you know, not having a gap, you can actually add the dot dot, '..' to show a pause, and the Disable Auto Scrolling is this. So auto scrolling on means that when I do this, can you see, it moves along, and kind of moves up and down, depending on what's being played, You can turn that off by clicking that button there. Now there's a bit more to transcriptions, but this video is already getting long, so I'll separate them out in other videos, including how to style them, and do them manually. 

The other thing to note is that this particular feature, the automatic transcript making is a new feature, like literally came out yesterday, so yours might change a little bit, there might be a little bit of jiggling around, or another option in here, or they've changed the words, they've moved this up and this down, so if there is something a little different, check the comments below, I'll update them down there, if they're just teeny tiny ones. I'll re-record the video if there are big things, but you might just find other new features, because you're advanced now, you can look around and go, "Actually, I reckon I can work that out," if something does change, I'm totally trying to get out of making updated video for this one. I'm hoping this could be the one forever, please, Adobe, don't change it, but if they do, have a look around, check the comments, and yeah, that is automatic transcript, but really we want the automatic captions. 

All right, let's jump into the next video.