Speaker 1: In this video, I'll show you how you can automatically transcribe and add subtitles to your video in Subtitle Edit. Now, this is highly welcome news inside Subtitle Edit, and they've been improving every day. So having automatic transcription inside Subtitle Edit, a free application that you can use to subtitle your videos, burn in subtitles to your video is really awesome because it is going to save you time, and probably a lot of back and forth in trying to transcribe and all that. So you can just let the voice recognition do the heavy lifting for you. And you can just make minor adjustments to your subtitles in Subtitle Edit. So here's Subtitle Edit, the latest version is the one that supports that and I believe it's Subtitle Edit 3.6.5 Yes, it's that released March 2022. But I believe in future versions, this will still be available and hopefully, the VUSK models that are being used will be a little bit finer in terms of what they recognize. But it's pretty awesome. So the way I've structured this video is that we are going to install the VUSK models. And then we are going to automatically transcribe and add subtitles to a video. So this is our interface and we'll begin by adding a video. So let's just add our standard video for testing here. And our video is loaded. And if your waveform is not here, you can just click the options available here for the waveform to be generated. Next up, you'll need to click on video, video stroke audio to text. And this is the audio to text feature inside Subtitle Edit. I will begin by downloading FFmpeg, I had not uploaded it. Let's just give it a couple of minutes so that you can process these details. It's not that big, I believe. But as I as I mentioned, automatic transcription inside Subtitle Edit 3.6.5 is a game changer in that it improves your subtitling or captioning speed. And hopefully, depending on the language you want to use automatic transcription for in Subtitle Edit, the model is available and it accurately picks up whatever you actually what I picked up. So FFmpeg has been downloaded and is ready for use, click OK. And once you do that, you'll need to download LibVSK. So click Yes. And then let's just give it a couple of seconds. And as I mentioned, it's a step by step process so that you see there's nothing I've skipped and you can follow along. Pretty simple. So it's almost done. And I'd like us to see what results you get. VUSK was downloaded and is ready for use. Click OK. Now, the audio to text video interface is actually to text interface is actually going to be pop up and you can generate text from audio via VUSK forward slash CALDIS speech recognition and you can check out the VUSK website and you can see the models. You can see them from here if you already downloaded them. And then there's this option to actually download and then you can open the models folder. And there's a checkbox here where you use post processing, line merge, fix casing, punctuation and more. And if you have the model loaded, you can generate. Let's click on VUSK website. Once you click on that, it's going to take us to this website. And you can see the different models available. And I believe they support 20 plus languages. You can see the different models here, the small and this one and this one. And then the word error rate or speed and the notes and the license. And you can scroll Indian, Chinese, Russian, French, German, Spanish etc. So let's go back to subtitle edit. Now, we'll click on these three dots to download an English model because this video is in English. So let's click on this. And you can see the option to download VUSK models. So English, Chinese, French, etc, all these languages, I believe it's more or less like 12 plus languages. So let's click on download. So let's give it a couple of seconds. I believe the English model being downloaded is, let's just scroll up a little bit is the 128 MB version. I've noted people telling me that the bigger versions, the 1.8 GB version here doesn't work. I'm not sure why I've not tested it. But I'll try and see if I can test it out. Now, considering the languages available, I believe anyone or anyone that is actually working on subtitles for the languages that have the VUSK models, you can greatly improve your speed in that you can just load your video, have subtitle edit automatically transcribe it into text, append the text as subtitles to your video, and they're pretty accurate from the tests I've done. I believe it picks the language really good, but it's up to you the viewer who is watching this video to actually test it out and see if it works as advertised. So let's just go back to the VUSK website and see a couple of more models here, including Arabic, Catalan, Farsi, Filipino. So it's it's scattered across the board, right to left languages, left to right languages. And I believe most most languages are spoken worldwide. So you can always see the punctuation models and case restoration, we recommend the models trained with, I believe this folks here for English, you can actually probably just add all this. But remember, they want the size is 1.6 gigs. And then you can also check out all these other options here other places where you can check for models, which might be compatible. And my assumption is, if you go to these different websites, you might see your language there based on the creators that are actually adding these models there. And then there's the you can train your own model, ATC, if you know how to do this kind of programming, then you can do that. But let's go back to subtitle edit, it's almost done. So I think if you actually in my own opinion, the latest features that have been added to subtitle edit, really put it at par with professional, or let me just say with the paid subtitling tools out there. So it's really a good thing. Now, the model is loaded, it's the VUSK model, English US 0.22 dash L graph. So if you have other models here, they'll be available, I will ensure that this is checked for us to use post processing, the line merge fixing, affix casing, punctuation and more. And then let's just click on generate. And it's going to load the VUSK transcribing audio to text. Since our video is just a short video, a three minute video, more or less, I believe I've noted it's taking about like 30 seconds or so. But test out and see if based on the language you're transcribing, it's going to take a lesser time, less time or more time. And maybe what I probably hope is that it's possible to select, especially if you're working with a video with two or more languages, you can select if this is transcribed into text, and then select the lines that you have and retranscribe based on the model for that particular section because it might be in a different language. So that is probably something that is a little bit complex, but probably something that would be of consideration when doing this. So we're almost done. Just about a couple of seconds remaining. But I'd like us to see if probably the result is something that we can work with. Maybe it is. So here we go. And I'm sorry, this video feels a little bit long, but it's for anyone that wants to use subtitle edit to automatically transcribe and subtitle their videos. So there we go. The subtitles are available and you can now see them on the timeline here. And if you actually just play.

Speaker 2: With the recent update on the YouTube interface, it's now easier than before to download YouTube transcripts. Hi, my name is David from freelancerinsights.com.

Speaker 1: I feel that it's done a good job. Now, the only couple of things that you might need to do is just tighten the script a little bit so that you just don't have a single word appearing alone. You can just you can merge, let's say you go to merge with line before. So you get YouTube transcripts. So you just merge that instead of having it as one. You can just go tightening up your subtitle, for example, all this and see what works best for you can merge a couple of lines together. And then let's see, you could go to this section, thank you for watching and have a good time. It looks pretty good. When I look at it, number one, capitalization, punctuation, and in general, the transcription part of it. It's pretty awesome for something that is you are transcribing and automatically adding subtitles to the video offline. Pretty awesome blows everything out of the water. And let's assume you wanted to burn your subtitles. As always, you just come to the video, generate video with burned in subs, click on that. And then you'll get all the different options. You can check out the preview to see how your subtitles will look like is what they look like based on the based on the font family you have. Let's go with something like Arial. Let's go Arial. Then we can remove the opaque box. Let's preview is what it looks like. Looks good. So if you wanted to do all that, just click on generate, and then set where you want to save the video. Just click on save, it starts burning into your video, the subtitles and all that. And it's pretty awesome. So I'd recommend that you try the automatic transcription feature that has been recently introduced in subtitle edit 3.6.5. See if it works for your language. And hopefully, it gives you good output. Our video is burned, the subtitles are burned into the video, you can just play it.

Speaker 2: With the recent updates on the YouTube interface, it's now easier than before to download YouTube

Speaker 1: and also the change that we just made. So I think it's done a good job. So that's it for me. Next up, we've automatically transcribed our video using the model that we downloaded. And then finally, we've made some few edits just to show you how you can do that. And then we've burned in our subtitles to a video. So you can see you can start from start to finish that is you can automatically transcribe, make edits and burn the subtitles to the video right from subtitle edit without the need for using any other application. So that's it for me. I hope this video is of value to you. Thank you so much for watching.