 Hey guys, welcome back to my youtube channel. This is Daniel Rossell here Today's video is going to be regarding how to generate automatic subtitles in Kaden live by using speech recognition and specifically how to do this if you are using Kaden live in Ubuntu uh now I feel like subtitling has kind of become the theme of my youtube channel for the past um past few weeks at least because I've been using subtitles in all sorts of contexts specifically for language learning and this is a really really nice feature that will basically allow you to Take let's say a video blog. So what I've done firstly is um, I have put a video blog I recorded about an hour ago hence the Same t-shirt same same room different hour And we're gonna just install i'm gonna show you guys how to install the speech to text in uh, Kaden live on Ubuntu Now before I do that just one one question that people might have is well You know, I've shown in videos the last few weeks how youtube has very nice Sub subtitling features you upload something like this video blog to youtube One day later or so youtube has automatically subtitled it So what would be the point if this tech exists in using automatic subtitling here? So there is a reason the reason would be that if you want to firstly embed your subtitles Directly into the video file the best workflow these days for creating subtitles is an AI human combo in other words some kind of algorithm which we're going to be Uh downloading from the internet in this video And then correct those as a human right the algorithm is going to get some words wrong It's not going to capitalize words correctly stuff like that So you can do this workflow inside of Kaden live and that gives you the option of saying You know, I want to embed in the video track English subtitles And then I want to actually export the subtitle file And put that as my caption file on youtube It's just a kind of more robust professional way of subtitling than doing it in youtube. I would I would argue Anyway, here's how to set it up. So there's two ways you're going to encounter this in your Kaden live First one is project subtitles And You may have been subtitling in Kaden live and said who was this fancy looking speech recognition thing Now this is what it's going to look like after it's being configured I'm using now the vosk model blah blah blah and you get to choose timeline zone timeline zone or selected clips very important To pay attention to this because Timeline zone is is the little blue line here on my editing timeline So you need to stretch that out or use keyboard shortcuts to get it all to the video Or what I can do as well another another method I can do here is again This is a video blog I'm just going to ungroup the video on the audio layers. I'm going to select The audio only because clearly we're not subtitling video and then repeat the process project subtitles Speech recognition and then point to two selected clip that would work too Now another way to see how you are set up is by going settings configure and here you're going to have a One of the options is going to be speech to text Now this is again after I've set it up But I want to draw your attention to what it says at the bottom here Speech to text is configured and it says we have vosk 0.3 0.43 And we have srt 3.5.2 Now srt might clue you in To the fact that that's the python module for generating subtitles because srt Is kind of the most common subtitle file format So what that particular python module does is generate those subtitles But that's so it's a separate process. Firstly, we listen to audio We decipher what the words are that requires a speech to text model and secondly we're going to use a Python module for actually generating the subtitles. I'm just breaking down the text So it's a little bit more logical what we're actually doing here instead of just putting random commands into the computer So it says downloads. So firstly, there's custom models folder And if you're an advanced user, you're really into this stuff. You can train up your own Uh, it would be super cool, but I don't have time for that right now But I'm just using the english us And despite the fact that I have clearly got a irish accent It's doing pretty okay with me. Not wonderful, but it's as good as youtube put it like that So, um, you you can follow this link here and it's going to take you over to I'm just going to drag it into my screen here alpha c-fi models And the one I went for was v osk model en us dot zero two two. It's a 1.8 gig Download, uh, so it's on the large it's on the larger side But uh, they have a bunch of other languages like chinese russian french german spanish So and the cool thing about this is you can have as many languages as you want here You just need to tell when you're running the speech recognition. You just need to say Uh, this is english. So use the english, uh language file, please the next thing I did So I downloaded this 1.8 gigs and then I kind of told there's a plus icon. It's pretty self explanatory You just say tell the tell kate and live where it may expect to find this model on your computer Uh, I I saved mine here program speech to text and I just pointed the um I pointed the Kate and live to that folder I think I went for this one and clicked open or something I'm not going to do it again because it's like many things on linux once it's working Do not touch it. So um, it is working right now. So I'm just gonna that was what I did there So the next stage in the process is installing the Modules, uh, so this was what I had to do. I'm gonna drag over my notepad I've prepared the commands that you are going to need I'll also leave these in the video description just to be helpful Um, out of the box my ubuntu Has python running on it yours probably does too But I didn't have these two modules So all I needed to do was install these two modules. The first thing I needed to do was install this Um, this utility called a pip3. It's super useful and it'll just Uh make installing python modules kind of effortless So that was my first command sudo apt get install python 3 pip And then I use these commands pip3 install vosk and pip3 install srt to install these two packages And I hope I made I hope that was clear What each one is doing right there The first one is the the heavy stuff. It's uh doing the speech attacks The second one is just automatically shoving in those uh subtitles So once you're once you're there, you're pretty much good. That was the hard part done So basically just to do a quick recap Firstly, you want to download your text to speech model You want to point kate in life to us? You want to make sure you've got these two modules installed vosk and srt And then it even has a nice little configuration checker So you can make sure you're good to go on all this and there's an update utility there too So we are now ready. I think so remember what I said that You need to make sure you've selected so I'm going to just do like intentionally only one minutes Because I'm going to just do this live and to show you guys so that it won't take forever So I'm going to automatically do subtitles on a minute of this now here what I want to go project I want to go subtitles. I want to go speech recognition I'm just going to do selected clips because there's no reason for this algorithm to parse the video because uh Uh, there's just video there. So I'm just saying hey, just take the audio layer and we're just going to do a minute And then click process And then make yourself a cup of coffee Um, kate in life being kate in life You know sometimes this is going to crash the machine Sometimes it's just not going to work because the computer is going to bad moves. Just make ah, this is working. So we're um Sorry, I'm aligned kate in life for no reason Now because it's a minute, we're running through this fairly quickly If it was 20 minutes the full project It's going to take 20 times as long to generate your subtitles and remember this is actually a two part process I'm going to put myself down so you can see the magic happen once we hit 100 boom. You see it Our subtitles were added to the video Automatically, I did not do this the computer did this and we can see the subtitles. They've been embedded. Welcome back to my youtube channel This is Daniel, uh My usual intro. This is Daniel Rose Hill. So you guys can see the changes I need to make. I want to add a capital D I want to make rose. Oh r o s e h i l l But it's a very very good start and another cool thing about this is that I can Choose to Embed the subtitles if I want to and just in case you don't know how to do that You simply can Hide the subtitles layer just like this by clicking on the eye What's going to happen now is that if I render out this video It's not going to include the subtitles, but I can edit the subtitles in katin live and then Export the subtitle file like this and it will export the subtitle file. So really Versus doing this in youtube For a it's more flexible b if you're not uploading videos to youtube, but to some other platform like vimeo or Anything for that matter This can be really really useful because if you're not uploading to youtube needless to say you're not going to have The youtube subtitling utilities there I think that's pretty much it. I think this will be a good guide for anyone looking to use this feature In katin live on a bunch of I hope this was indeed useful Um, and thank you guys for watching You're going to get more videos from me to feel free to click on the subscribe button and remember If you want your subtitles to be accurate, you should speak slowly and clearly Thank you guys for watching. Have a great day