Transcription
What's going on? Hey, if you are trying to create an AI clone of yourself using your video or your audio, I'm going to show you how to do it right now in this video using what I believe is the best AI cloning software out there, which is called Hey Genen, hygen.com, combined with 11Labs, which is 11labs.io. I'm going to show you right now based off of massive amounts of experience how to create a really good clone cloning both your voice which is going to be used through 11 Labs and then your actual video of yourself using Hey Gen.
So I'm Jared Larkin and uh follow along at milehighleguide.com if you'd like to learn more uh tools, resources, and marketing ideas where I can help you. I want to show you right now what I've been working on. It's taken me quite some time. I've cloned myself. I've attempted to do it now five different times using AI. I'm going to share with you what's worked, what has not worked. That way you don't say make the same mistakes that I did when I was initially trying to figure all this stuff out.
So, what this looks like here, let me just show you a quick video. This is kind of the final product of one of the videos that I was working on. Let me go ahead and play it. Quick heads up. If you use Chat GPT, they just rolled out a new personalization feature that lets you control the tone and style of how ChatGpt responds. So, what you're looking at right now is in the bottom right corner the cloned version of myself, which is a cloned voice and video, which is pretty darn good. I'd say the only feedback I have for it, I'd say the voice is dialed. Little bit of too much hand movement, but it's all good. when you add in the B-roll footage and the captions and the audio, I feel like it all comes together really, really nicely. And that's what I'd say overall where we're at right now with AI and video cloning is I think when you add in the additional effects to it, um, you really can't tell anymore.
So, let me show you how to create something like this or at least get to the point of cloning yourself. So, first things first is you're going to need to go to heygen.com. What's going to happen is you can start off with the free account, start there, go ahead and just try to, you know, get going here. But eventually, what's going to happen is if you really want to use this system, you're going to want to upgrade to the $29 per month option. But wait until you have to do that. Just start off free. Once you're inside, what you're going to see here is you're going to see kind of your homepage. What you'll want to do is you want to go down to avatars and you're going to click the create new avatar. All right. I've already kind of screenshotted some of this for you just to show you what this is going to look like. But what you're going to do is you're going to create your avatar. First thing it's going to ask you to give you some instructions to follow. Um after you do that, you're going to need to also um essentially read a script saying you give permission to create that. So you'll go ahead and upload that. Um and then it's going to review your footage. It's going to rank your footage based off of what it believes is good or not good. And then eventually you can work with your avatar like you saw that I had before.
So, what does this look like here? You can see here I have cloned myself one, two, three, four, five times. I'm going to share with you what works, what doesn't work. Um, first things first, here's like the original version of myself. In a perfect world, if you can use a high quality video camera, right? A DSLR camera, put it up on a tripod, uh, use a microphone, right? That's always going to help. But what actually is interesting is that video that I just shared with you is actually the one video that I uploaded that was directly just from my iPhone which is shot right here on a tripod in 4K. And the only thing I have though I do have these DJI microphones. You can pick up these lavalier microphones for a couple hundred bucks. They are fantastic and they can really really improve your audio um even just recording off of your phone. So quick tip right there.
Now what you're seeing here, let me go and play with you. play my first original clone version. This is this is actually not a clone of me. This is me talking. And you'll notice this is not how I talk. So, I'm going to tell you what happened here. >> Hey there, my name is Jar. >> So, you can see how I'm pausing and I I actually look like a robot. And that's actually the real me. Um, this is actually the cloned version, like the actual Hey Gen output using that video sample. Let's play that. Thanks for joining me today for my AI class today. Um, I'm going to be sharing some of my So, you can see here the lips are kind of off. It's not that good. Um, here's what happened was when I first started going through this and following the instructions, it will tell you over here on the instructions page when you go to avatar, create new avatar, and um, say we're just going to create a new one from scratch, right? We're going to choose clone a real person. And then within here, right, it's going to choose your option if you're still uh, depending on what subscription you're on, but it's going to allow you to at least have one avatar. I'm on the $29 a month version right now. So, I'm just going to go back over here, do create new avatar. When I click through here though, what it's going to say is it's actually going to go ahead and it's going to actually um it's going to ask me to to follow a set of instructions. It's going to play a short video. And what that looks like is it's basically going to say um pause between every sentence. In my opinion, you really don't need to pause. I think you just kind of talk how you have to talk. You know, if you don't if you don't like, you know, ramble through the next sentence, obviously don't do that. But just be yourself. Um, the other thing to be mindful of is just not moving your head too much because if you move your head too much, that will start to affect the clone version. So, if you can kind of speak, use your hands. If you speak with your hands, I use my hands a lot. If anything, I need to use my hands less because that last clone version was a little bit too much. But, just be mindful of using your hands, kind of keeping your upper body still if you can. And then you're just going to talk about literally anything you can. Goal here, they say 2 minutes. I try to do at least five minutes on my clones. um to get that good output because at the end of the day, right, the more sample size footage that the AI can work with, the better output it's going to give you. So, I'm always aiming for five minutes. How I do that, five minutes is pretty darn long actually when you're just staring at a camera and talking. I use my watch, I set a timer, so once five minutes hits, it starts um vibrating so I know when I hit that five minute mark. Um also too, you want to look the camera just like I am right now. look the camera right into the lens the entire time, which is um uncomfortable and really weird to do, but you're going to do that just again to stay focused and have the good clone.
So, you can see here this clone number one was bad. I then went into this and created clone number two. >> Hey there, man. >> And kind of, you know, went through this whole process again. You'll see here another video. And then as most clone number three, same kind of deal. I backed up this time. You can see here I kind of screwed up the lighting on this. Again, other video tips, of course, right? You want good lighting. Um, right. The more lighting the better. Natural lighting, all that stuff is important. Um, here's another one where I was trying this one out. >> But, uh, here I am. Here's here's the DSLR camera. I shot this on a Sony A7S3. >> And then what you'll see here is the one that I actually used was the one off of my iPhone because I was having some issues with the footage uploading and actually processing. Once you upload the footage, it looks something like this. You're going to upload it. You're going to, you know, confirm it's you view. You're going to say, "Hey, this is all my all me. You can use it." And then it's going to tell you what's good and what's not good. But you can see here, right? Especially the facial uh movements. Camera is stable. Maintain same position in frame, right? I'm just moving my hands. Really, that's that's about it. Um, other things, mouth is visible, yes. Uh, strong background contrast. Sure. Subtle body movements, debatable. Um, one person, obviously, just me. At least 2 minutes of footage. Yes. I said I I always go for about 5 minutes of footage. And then video quality, right? We we don't want to shoot on a potato. Um let's use any I don't know any camera or phone that's been created in the last I don't know 8 years, right? Uh quiet background obviously. And then of course just you. So that you can see here I actually ranked bad on this one. Brief pauses between sentences. But as you could tell from my first clone when I paused too long it just made the whole video look weird. So I don't know. Um that's just my feedback for you.
Now boom. Once you upload it, it says your avatar is being created. Circle back. I don't know, take about maybe an hourish to finish the avatar once you've uploaded it. Um, you're going to do that same process, right? And then hopefully you don't have this situation, but look what happened here for me uh on a few of those clones is I kept having the avatar not actually processed and I had to work with unfortunately had to work with tech support to get that dialed in. So, um, if you have that issue, you might want to contact tech support. I kept having it unfortunately happen to me. So once you've done that here like I've done it's going to create your avatar and then you're going to see it here right within um down here within your recent creations as well as anything else you've made. Now what you want to know is when you're working with your avatar you can then start to test it out. Um you can also of course record right on your computer which I don't know I've got a MacBook Pro which is kind of old. You can see it here. Hey, this is but um it's not high quality. So, just be mindful of that. But point here is when it came down to it, the iPhone actually was the video I ended up using for my clone, which I find quite ironic.
Once you've gone ahead and created that avatar, you can now start to use it. So, you can click create up here and you go create video and then you can do create video agent, right? And let's go ahead and click through here. Here we can write something or what you can also do is you can create an AI studio which is what I like to do. So what I usually do is I go create video create an AI studio and then once this loads within here um you'll see right you can pull in uh the clone if you have multiple clones you can choose from them um and then it's going to give you what it's going to do is it's going to actually give you the default clone voice that is out of hey genen which may or may not be good for you. I've I've talked to a few people who've done this. Some people there's has come out good and then some other people uh it's come out really really bad. So let's go ahead and play mine. This is my Hey Jen voice. Let's see how it sounds here. >> Hey there. Here's how I sound. Hope you like it. The right voice brings avatar videos to life. Now I don't know if you can hear that but it's really bad. So what that means here is if if you have a bad voice like I do. >> Hey there. Here's how I sound. Hope you like it. I don't like that at all. >> Right voice.
What you're going to do is you're going to actually then segue over into 11labs.io. Now, again, start off with the free account. Uh, start there until it forces you to upgrade, but eventually it's going to make you upgrade here for $22 a month once you connect everything and start making videos. So, you can see this starts to add up quickly. you know, you're spending about $50 a month combining your AI voice with your AI video avatar to really make the ultimate cloned version of yourself. But what you'll do is when you sign into 11 Labs, you're going to walk through some of the steps. Go down to voices. Um, you're going to want to create or clone a voice. And then here's where it's going to make you upgrade, right? You're going to want a professional voice clone. If you're serious about this, you want a good clone voice. You want them to actually do a good job, right? Um, now you can see here it says requires at least 30 minutes of clean audio. Um, I've been looking into this pretty extensively and really I think a good goal would be at least 50 minutes to 2 hours. Again, the more you have, the better. Now, if you don't have, you know, 50 minutes of you um, talking, of course, that's going to be a task that you're going to have to complete. And we're talking about it's just you talking. There's there's no one else talking. In a perfect world, it's you uh you know, recording your voice, whether it's, I don't know, talking to your phone, uh using some sort of uh microphone or just whatever, doing a tutorial like this and then extracting that audio um or taking it out of a video. If you have an existing video, you could use um a program like Cap Cut to export the um audio out of it and then upload it into here um into 11labs.io, which is what I actually did. But again, I would say aim for at least 50 minutes because that's what I did. Um, and I like mine.
Once you do that, you're gonna see your voice under here. Here's my voice. >> Innovation starts with a question. What if >> that sounds aggressive? Um, but there it is. And what you'll want to do next is when you go in here, you're going to start to look at the voice. You can go into the voice ID. You can edit it. Um, you can really start to tweak this thing and mess with it here um, so that you like it. But once you do that, the next step, I'm not going to go through this step by step, but what you'll need to do is you'll need to go back over here into Hey Genen, and you're going to connect your voices. Um, and you're going to bring in a new voice. You're going to do that by integrating a third party voice. And then you're going to bring in the the actual API key. Okay. Um, the API key will be given to you through 11 Labs. You're going to copy and paste that in so that they integrate. And then you can select your new voice which right here's mine right here imported uh the voice clone and we'll bring that in. Okay. So let's see if this thing >> here's how I sound. I hope you like it. The right voice brings avatar videos to life >> compared to the one they gave me which is the original one. Let's play that. >> Hey there. Here's how I sound. Hope you like it. >> It's very quiet. >> So I always choose my integrated 11 Labs voice.
So now I can see it here and now I can go back in here and then what I do is this is now I'm working off of my avatar. I'm working off of my voice and now I can just essentially, you know, copy and paste in or type in whatever I want to create my script. So, what I do is I go to chat GPT and I actually created a GPT project that writes scripts for me automatically um based off of how I talk and how I sound. And you could do this, too. All you do is um I'll give you the prompt. If you'd like it, reach out to me, uh Mile High Title Guy, and I'll give it to you. But once you do that, you can actually use the custom prompt to sound just like you and create little scripts that you can then um copy and paste right into hey Jen and have it uh sound like you, right? So once I type this in here, here's what's going to happen. It is going to start to analyze this text and it's not going to actually show me the voice with the video unless I choose this generate button up here which will cost me some credits if I click on it. Um, but if I at least want to hear the audio, that's that's step one in my opinion. Make sure it just kind of flows really good. What you can do is down here you can press this play button. >> Ever notice this new drop down inside chat GPT? This is actually a big update. You now get three modes. Okay. So, it's not showing the actual uh me moving, right, the 3D version of myself, but it will show me uh I can hear the audio. Now, if you don't like the audio or you want to tweak it a little bit, you can click here on your voice clone. You can change the speed. You can change the volume. All right. I don't mess with the accent because I want it to sound like me. If you wanted to go more advanced settings over here, right, you can click on the voice and you can start to really mess with um similarity, stability, right? Style, exaggeration, etc., and save that. Okay.
So once you dial that in and you're happy with it, your final step is to then go ahead and generate the video, right? Again, this is where it's going to cost you money. I'm at the $30 a month option. I would have to upgrade if I wanted 4K, but you know what? 1080p actually looks good to me. Uh 25 frames per second MP4 video, obviously. Turn off the watermark. Click submit. And then it's going to go ahead and load to this screen. And it's going to start processing. And I'm going to have to wait until it gets to 100%. Once it gets to 100%, I can then download this video. Uh, here's one that I just did. Let me click on this and it will show you the lips matched to the audio. Let's take a look. Okay, come over here. Ever noticed this drop down inside chat GPT? This is actually a big update. You now get three modes. Auto means chat GPT decides how much thinking is needed based on your question, which is great for everyday use in which is pretty good. Um, obviously I'd speak a little bit faster if I really wanted to tweak this in. Uh, my teeth are looking pretty white in that. That's cool. But in my opinion, that is good enough. I'm going to download this. I'm going to click download here. I will then call it whatever I want. And then truthfully, the reason why I'm not too worried about this is like I showed you at the beginning of this video is when I go and I edit this within inside my application within I use Cap Cut. I'm going to take this video and and I'm really a big fan of the short form talking headstyle videos that look kind of like this. So here I am within Cap Cut. I'm going to create a new project. I'll bring that video in. I'm going to import it and I'm going to quickly remove the background and I'm going to start to edit it. So I'll bring this clip in. Here it is. Next step is I want to go over to remove BG. Click auto removal. This will remove the background. I will then after that bring in the B-roll footage that I want to use that will go below this. Um that will be essentially in the background while I speak. Right? So then I'll take this this this background B-roll footage is just a screen recording on my iPhone. I just air dropped it over in right into uh Cap Cut. I'll take this. Obviously, I want me smaller. So, I'll bring down the size here. Bring it over here, right? And then I will, of course, start to edit this, add captions, and do my thing. >> Ever notice this drop down? >> Also, what's nice about this is, is you can see here, I paused a little bit in between sentences, which I'll just want to splice out these gaps and make it look real clean and just more direct.
So, that's a tutorial on how I'm going to make my content. Um, if you don't use Cap Cut, you can of course also use Hey Gen. They do have their own AI studio within here, right? So, if I went in and wanted to create within the AI studio, I would work off of my existing AI avatar, type or copy and paste in my script. And then over here on the right hand side, you can start to see that you can use AI to create um, create videos, create images, right? You can use B-roll content, shapes, all the things. Add your music here and then generate it out. So, whatever floats your boat. But hopefully this was helpful for you. I hope this saves you time when using heygen.com and you don't have to go through creating five different videos like I did. And I hope you get it right the first time. Good luck to you. If you found this helpful, follow along milehighitgu.com where I share real estate tools, resources, all sorts of marketing ideas. If you're in Colorado, um, sign up. Would love to see you at some of my classes and events where I'm teaching tons on AI and marketing. Hope this was helpful. I'm Jared Larkin. See you in the next.