Look At What We Can Do With AI Video Now!

I recently posted this video on my Twitter account and a ton of people asked me how I did it. And this video was ultra simple to make. Now I went to my daughter's school Father daughter dance and they had this cool 360 camera where you stand on a platform and the camera sort of rotated around you and they shot this video of us and the video they gave us back. I just plugged through a tool called Kyber and it gave me this cool, like, animated effect here that makes my daughter look like an anime character and me look like a scary monster. But seeing all the comments, people wondering how I made this video made me want to do another video where I talk about all of the current video tools that are available with AI. Not too long ago I did this Twitter thread about nine AI tools that are currently available to generate videos. And it's one of my most popular tweets ever, getting over 876,000 views and 619 retweets. And it just went crazy. So I know people get excited about AI video tools and. And in this one, you can see all my tabs up here along the top of my screen. You can see that I'm gonna talk about quite a few options for getting cool effects out of videos. So let's start with D id. This is a tool where you can upload an image that looks like a person that has a mouth on it and make that person animated. So let me jump into it. They do have a free trial. You do get a certain amount of free credits on it. Once you use up those free credits, then it is a paid plan beyond that, but it's a really simple tool to go ahead and create a video real quick. So you can see down here, you can create an AI presenter or you can choose a presenter that they've already got in here, or you can add your own image and use that as the presenter. So that's what I'm going to do. I'm going to grab an image off my desktop here and you can see this is an image that I made in mid journey here. I can either upload a voice audio over here, or I can type a script and this character will speak it out. So let's go ahead and change the voice to Brandon. I don't know what this is going to sound like, but I'm going to type some text. Hello, my name is Brandon and there are a lot of papers spinning around my head right now. This is really crazy. Sure. Then I can come up here and click on generate video. Click Generate again. It's going to use one credit. I have 40 credits left now. It says my video is done. I can play it back.

Hello, my name is Brandon, and there are a lot of papers spinning around my head right now. This is really crazy.

You could see that it's actually animating his lips as he's talking. It's animating his head. He's kind of doing some subtle motions, even blinks throughout it. So it's a really cool, really simple way to animate faces. And this one's called D id. The next one I want to show you is called Leiapix. I've shown this in a few other videos in the past. Really, really simple concept. You just upload an image. Here's an image that I made of the Godfather. You can see how it sort of animates it and gives it like a 3D look here. You can change the animation length to be a lot f faster. You can change the animation style to kind of go back and forth or in wide circles or vertical or this perspective. You could change the amount of motion. You could change the focus point to far away or close up or center. And there's all sorts of different settings that you can mess with to kind of get this 3D perspective animation. So if you have a still image, this is a good way to add some motion to it. A lot of times when I'm making videos, and instead of just having a single static image for B roll over the video, it's kind of cool to add an effect like this to just give that video a little bit extra motion. This one's called Leiapix. You can find it over at L E I A P I x dot com. It's called the Leiapix converter, and this one is currently completely free to use. Next up, we've got animated drawings. This one comes from Meta. It's actually designed to upload your children's drawings and animate those drawings, but it will actually work with AI generated images, sort of. I made this in Mid Journey. It's a quick superhero image. I'm going to upload this and see if I can animate it using animated drawings here. So let's go ahead and click Next. It's going to scan the image, agree to the terms that you're basically helping them do research. And this is just for demo purposes only. Right now we're going to go ahead and tell it where the image is confined to here. Click Next. It's going to scan it. You can see it put a white outline around the character. So it found the character just fine. If it doesn't find the character, you can sort of draw the mask around it. And now it's asking us to sort of pick various points. So where's the right ankle? We'll put it right up about here. Left ankle right about here. Might come out a little funky, but let's see what we get. We'll click next. And now it's going to create an animation out of that character. And that's what we get. A little rubbery looking. But let's go ahead and make her jump up and down and wave. And that's another way to make a real simple anim out of a image that you made in something like midjourney or Leonardo or something like that. Or even a hand drawn image. Let's make her boxing here. I mean, you could see it's a little rubbery just because it's meant for kids drawings. Not really designed to use with like midjourney images. So you get this sort of like rubber woman effect here. But I don't know, it's kind of funny. Yep, yep, look at that. Just kind of something silly you can do to make static images a little more animated and fun here. Now I'm going to get into some of the like, really serious, really cool AI tools that are available for video right now, starting with Runway. So over@rwayml.com they have an entire suite of AI video editing tools. You can do motion tracking, you can add a green screen background, you can blur the background, replace the background, automatically generate subtitles, you can blur faces in videos. You could turn any video into a slow motion video. You can do inpainting where you actually remove people or objects from videos that you don't want. You can completely remove the backgrounds from videos. But what Runway's been most known for these days has been their Gen1 and their Gen2 products. So Gen1 is where you can upload a video. Let's say, for example, here's this video of this animated girl dancing here. And you can apply a image to it as well. For example, let's go ahead and apply this vibrant pattern to it. And now we can go ahead and generate a video. And it should take a combination of the existing video and apply this sort of pattern with it as well to create something completely unique. And here's what it came up with. You could see it's that same dancing girl, but now it applied this sort of pattern style over the top of it and you get this completely unique stylized effect. Now this is what they call Gen1. And you can get access to Gen1 with your Runway account. And Runway is also a freemium model, which means you can do a certain amount of things for free before having to upgrade and actually pay. But they also have something called Gen 2, which is actually still in closed beta access. So not everybody has access to it yet. But that's a complete text video. You just enter a text prompt and it will generate a video from it. So to use Gentoo, you come to Runways Discord, and if you have access to it, you'll see it over on the left sidebar you'll see this Gen 2 section. You simply type 2 and then you just give it a text prompt of what you want to see a video of a wolf walking through a forest. You'll see this little timer here to let you know that it's generating it. It creates a brand new thread over here on the left, so you can actually see some details about what it's generating. And when it's ready, it'll just pop up right here in this new thread. So I can press play and we can see our wolf running through the forest here. Let's go ahead and look at it again in full screen. I'd say it looks more like a coyote, maybe a Chupacabra. I don't know. You can kind of see it. The thing that I've noticed about Gen 2 is sometimes you got to run the same prompt five or six times before you actually get the video that you really want to see out of it. I'm going to go ahead and say I don't like my result because I didn't really feel like that looked like a wolf. That feedback is good for Gen 2 because it's reinforcement learning. You're letting it know that the video didn't match the prompt as well as it should have. If I run the prompt again, I get a completely different result. I guess it looks a little more like a wolf, at least some sort of animal walking through a forest. And that's Gen 2, which again is in private beta. But I believe it'll be rolling into the Runway dashboard Fairly soon. Now, Gen 2 is in private beta, but this one, called Model Scope, is actually available for free on a hugging face. And I'll link up the hugging face URL down below the video, but it essentially does the same thing. Now, the major downside of this one is it only generates two second videos and it puts a giant Shutterstock watermark across the videos. But I just went ahead and Entered this prompt. A monkey on roller skates. And you can see I've got a monkey sitting on the ground, I guess, playing with its roller skates. Here's another example with Spider man surfing, a panda eating bamboo on a rock and an astronaut riding a horse. And again, pretty much all these. You see this giant Shutterstock watermark on them, which I think proves to us that it was trained on a lot of Shutterstock footage. Now the next tool I want to talk about is called genmo. Now there's the original genmo and then they have a new genmo that they've been working on that is still in beta access right now. The original genmo is a text to video, or you can upload an image as a starter image and it will sort of animate that image. So, for example, if I upload this same image of the guy with the paper swirling around his head, I can describe some edits and it will tweak the video a bit. For example, I can say, turn the man into a cyborg. Let's set it to six seconds down here. Leave everything else the same and click make video. And here's the video we got out of it where you can see that it's kind of cycling through our original image and kind of putting a cyborg behind him, changing the guy's skin color a little bit, putting some glasses on him. But it gives us some unique images that could be an interesting stylistic choice. If you watch my video about how I made a short film, I used genmo to make a time lapse sequence of a city that was in ruin, time lapsing to become a city that was essentially beautified over time. They also have a text to video where you can just type a text prompt a cyborg dancing in a field of flowers. Let's see what we get. Let's generate the image. And basically what this is going to do is it's essentially going to generate that starting image and then it's going to run the same process we just saw, but based on the starting image that we generate Here you can see there's our cyborg dancing in a field of flowers. So if I check this box, it just moves to this next phase here where I can describe some edits and it will do the same thing we just saw from that last video, but from this image as the starting image. Now genmo has a new feature called genmo Chat, which is in beta. It's not available to everybody yet. I'm not totally sure when this is going to be rolling out to everybody. But I did get early access to play with it. And this also has text to video in it. So I can use natural language and say something like, generate a video of a cyborg dancing among flowers. Submit it. Sure. I can generate a video of a cyborg dancing among flowers. What kind of flowers would you like? Daisies. And then it says starting video generation. And here's the video we got out of it. A cyborg with apparently long curly hair dancing among the flowers. But you can have this generate anything you can imagine as well. You just ask the chat. You can start with an image. You can start with a text prompt. There's all sorts of stuff that you can do with this genmo chat as well, though this is just scratching the surface, but since I still have a bunch more tools I want to show you, I'm going to move on. So when I showed you this video earlier of me with my daughter at the Father Daughter Dance here, this was actually made with this tool, Kyber. If you look in my videos, you can see I've made a handful of different ones in here, including this one that I just showed you on my Twitter. The prompt that I used, I basically uploaded this video and used the prompt a happy father and daughter in the style of still from a retro futuristic film. I don't remember exactly why I picked this prompt. Lo fi, steampunk, cinematic focus. Realistic, highly detailed masterpiece by Albert Robita. So a lot of this stuff was extra prompt magic that Kyber sort of added to the mix when I uploaded my video. This is one where I started from an image and had it sort of morph my face into like an older man. This one, I started with a Nerf generation and then had Kyber run through it, which I'll talk about nerfs in a minute here. And Kyber just has all sorts of cool stuff. So if I come to create a video up here, you can begin with an image. You can transform an existing video, which is what I did with the one with my daughter, and you can add a song to your video as well. And I believe it sort of syncs up the animation to the song if you use that feature. But if I begin with an image here, let's go ahead and start with the same image that I used earlier. Continue to prompt. I want to create a video of a man with papers swirling around his head. Click, Continue. And then the style. It gives you some ideas for a style or you can type your own. So let's select oil painting here and you can see it adds all this extra prompting magic to it here. Now I can click Continue to settings. I can have the camera move in or out, rotate, do all sorts of stuff. Let's have the camera zoom out and see what happens. Show initial image to start. Let's leave that and let's go ahead and generate and see what it does. You can see here's the video that it generated from that, where it almost kind of turns it into like an old fashioned looking painting. I don't know how else to describe it. I guess. I guess it's more of an oil painting style like how he described it, but very, very artistic effect. And then if we come to create video and we transform an existing video, it works completely differently. So for example, if I upload this video of the girl dancing here and we use that as our starting video and then we continue to a prompt. This actually works kind of similar to what you would expect out of something like Gen1, which I showed you earlier, which combines the video plus whatever prompt you plug into it. So I want to create a video of a woman dancing. We'll click Continue and then for the style, let's do vector art and see what we get out of it. It should completely change the style of that video that we started with. I'm going to go ahead and leave this on the default and click Create. So now it gives some preview frames so we can kind of pick the style that we want to see from it based on these frames. Of these, I think this first one looks the most interesting to me. So we'll click Finalize Video and it's going to go ahead and run its processing. And here's what we got out of that one. And to me it looks just awesome. It's like a cool cartoon. I personally really love how this one came out. Honestly, it looks really cool. I actually think this came out better than the original that I made in Stable Diffusion. It really sort of improved, improved upon it. It got rid of some of the flickering and gave kind of more of a cartoony effect. So that's really cool. I can then upscale this video, which I think I will do. And man, am I happy with the way that one actually turned out. The next one I want to show you is called Plasma Punk. And Plasma Punk actually makes a video that's synced up to music. So it kind of puts the music first with this one. So when you start, it asks you to choose music or you can upload your own music. But let's go ahead and select Playful and let's listen to some of these and pick one. Is it Tronic? Okay, I like that one. So let's go ahead and use this one. There are other styles that you can play with here. Click around, see if there's any music you like. I'm just going to keep this moving and I'm going to pick this. Is it Tronic one here. And I'm going to move on to the next step. Now it asks you to pick a chunk of the music. It'll actually give you a recommended section. So if I reset it to recommended, it recommends this section. Foreign. I'm good with that. It'll be about 20 seconds. So I'm going to go ahead and click next. You can choose a starting image if you want. I'm going to go ahead and use the default camera settings. Some of these other camera settings are new since the last time I was in here, but I'm going to use the default because I really like the way the default came out in the past. And I'm going to click next. I'm not even going to pick a starting image. You have some various styles here. Color, photorealistic. I'm going to use colorful because I've used colorful in the past and it's always looked awesome to me. And then we can enter a prompt, a colorful Minecraft village. Trees and animals wandering around. And then for content description, let's do a Minecraft villager. I don't know. Let's see what happens. And let's click Generate video. Now I'm in queue at number 378 and here's the video that it. Definitely got those Minecraft vibes to it. I just think this generates some really cool, really vibrant videos. So I love the way these always come out. All right, so the next tool I want to tell you about is called Decoherence. And this is basically using Deforum, which I did a whole several hour long training on with some guest Experts on this YouTube channel in the past. And this really simplifies it. It gives a simple user interface to Deforum. And this one, they have a free trial. You get a certain amount of free credits and then it's paid after that. Forgot to tell you, Plasma Punk is free right now. Decoherence has a free trial, but then you have to pay for credits. This one's actually an app that you download onto your computer. So you would download it. Click on New Project. I'm just going to call it Sample Project. We'll create the project here. It asks for a style. You've got a bunch of style options here, let's go ahead and select Synthwave Punk. Under Starting frame, you're essentially going to generate a starting image. So whatever prompt you can imagine from, like a stable diffusion prompt, let's just do a beautiful cyberpunk woman with a futuristic stick, Cyberpunk cityscape in the background. And then we'll just add some extra keywords, HD 4k, ultra realistic, unreal engine detailed. And then it'll generate some images that we can use as our starting prompt image. All right, so we have some options here. We've got this image, this image, this image, and this one. I think this one's probably the most coherent. So we'll go ahead and start with this image. And then under general, I'm just going to leave the frames per second at 12. Pretty much going to leave all of this the same. But let's set the zoom so. So it's zooming in a little bit. Let's set it to 0.25 here and we'll leave all the other settings the same. And what makes decoherence really cool is you can actually sync it to audio and it will actually move the frames based on the beats of the audio. So if I import audio here, I've got this song here that I made in movert and I can open this and I will sync it to this. So if I just listen to the song back, it sounds like this. But I can tell it to separate the audio channels. And now it's separated out drums. This is what it thinks the drum sounds like. We've got bass. Let's see if it thinks it found any vocals. Not really. And then other it captured this, some of the synth sound. So let's listen to the drums again. All right, so now we've got our audio imported. Now if I go to audio reactive effects, I can add a new effect and I'll select the channel as the drums channel select effect. And let's just do X rotation so every drum beat, it'll rotate a little. We'll name this Cyberpunk Rotate. And then down here it wants us to enter a prompt. So let's go ahead and enter a prompt of cyberpunk woman dancing in a futuristic city. And you can add more prompts along the rest of the song. But I'm just going to have it cut off right here just for the demonstration purposes. So I'll bring this all the way in here like this. So it's just the exact length of this song. And we'll go ahead and generate our video and here's what we got out of this one. It's a cyberpunk woman dancing in the street with a kind of cool, artsy, deforum style to it. So the next thing I want to talk about is nerfs. And I've talked about nerfs a little bit before. But basically what a NERF is, it's when you take a whole bunch of images or a video around an object, and then it sort of turns that Object into a 3D object because it captured images at every possible angle. And these can actually be turned into some really cool videos. Now, this is a nerf that I did of my disc golf cage that's out in my yard. But what's cool about nerfs and using a tool like Luma Labs, which you can find at Lumalabs AI, is that you can actually look at your 3D object here. See, I can move it around however I want, but I can create an animation out of this. If I come up here and click on this little camera button up here, I can actually create keyframes. So right now you can see that the camera is just going to circle around it. But if I come over here to the left and edit keyframes, I can actually move this around here. I can make the camera spin and look at different angles. I can make the camera move up and down. So let's say I want to get a low angle. When it circles around to this side, you can see I'm setting a camera path for it to circle around this object here. Maybe when it gets to this side, I want it be up a little bit higher here and maybe a little bit closer in. And now you can see I've got this camera path that is going to follow. I could click render video. And now I've got this animation here where it's going to follow that exact camera path that I just created for it. You can see how it pivots up and it pivots down, and at some point it changes angles a little bit. You can see the camera kind of twists a little bit on it here. And you can just make these really cool animations with it and then download the video. Here's a NERF I tried to do of my tent when I went camping. It's not a very good one because I didn't know what I was doing yet, and I think I moved a little bit too fast. That's why everything looks a little bit blurry and fuzzy. But you can see this animation of the camera spinning around the tent from When I went camping, I could also do stuff like this, where I zoom way out from my tent and then come in through the trees at it. You know, you can make really cool cinematic effects. I could move the camera around this way, move it around this way. And by going in and just setting keyframes, I can create this animation however I want of this Nerf that I created. Now, if you're following Billevol on Twitter, this is like his specialty. You can see he makes videos like this, where he got this Nerf of the Lodi Garden in India, and then he turned it into, like, a Minecraft video. And if you go through his feed here, he does a lot of cool Nerfs where he runs them through Gen 2, or he runs them through Kyber or some sort of tool like that to get these really awesome effects. See, here's another one that he did right here. Billevol is the absolute king at doing Nerfs plus cool video effects over the top. And if I jump back to Kyber over here and click on my videos, remember that video I just showed you of my disc golf cage here? You can see I took that Nerf and I ran it through Kyber and basically got a disc golf cage in hell here, where you can see as it spins around, it's kind of changing angles on the cage. There's supposed to be fire and flame in the background. That was me experimenting with the combination of taking a Nerf video that I made with Luma Labs and combining it with Kyber. So really, really cool, creative stuff that you can do by taking one tool and smashing it together with another tool and blending these kind of concepts together to get really, really awesome videos here. Now, the last tool I want to show you is another one that's still not in open access yet, but if you get on the wait list, they are slowly opening people up to it. And that is Wonder Dynamics. This is a tool where you can take. Take any video with a person in it and replace that person with a 3D character. You may have even seen my past video where I took this video of myself walking around in this office that I'm in here and, you know, pointing at the camera and moving around. And then I replaced it with this boombox head radio robot dude, and it came out really, really cool. I also tried to make a clip from this iconic scene from Anchorman, but it didn't fully work out. You'll see Cannonball, and then he turned into a human right before he splashes. But the technology's coming A long way. Wonder Dynamics is one that's really, really fun to play with. I couldn't make a video about generative AI and video tools without mentioning Wonder Dynamics. And that's all I got for you. I know it's a lot. There's a lot of cool AI video tools that are out there, a lot to play with. It's really blowing my mind, all of the stuff that we can do and. And it especially gets cool when you start combining some of these. You take this tool, you make a video, you put it into this video to AI tool and you add some extra flavor to it. You can make a 3D image in Leiapix and then pull it into Kyber to get another effect on top of it. Or make an animation with this meta animated drawings tool and pull it into Gen 2 and see what comes out of it that way. There's just so many ways to mix and match and create really cool, creative AI videos. And I'm really excited to see what you do with some of these tools. A lot of them are completely free to use right now too. Some of them are on free trials and freemium where you gotta pay in the future. Some of them are still in early beta testing where you can just use EM for free right now. And some of them are still just purely in beta where you gotta get on a wait list. And they're not open to everybody yet. But there's just so much cool stuff you can do. So many cool tools. And I'm excited to play around with them more. And I'm even more excited to see where all these tools go and what's next for AI video. And if you want to stay in the loop with all the latest and greatest tools that are coming out, make sure you check out futuretools IO. This is where I curate all the newest cool tools that I come across. I add new tools every single day. I keep you in the loop with the AI News and the AI News section. And if all of this AI stuff is overwhelming and you just want a TLDR for the week, join the free newsletter Every Friday I will just send you an overview of what you need to know in the AI world. I'll send you the five coolest tools from the week. Handful of news articles, handful of videos, and one cool way to make money with AI. It's totally free. You can join it at FutureTools IO. And you'll only receive one email a week from me every Friday. And Yeah, that's @FutureTools IO. So go check it out. And if you like videos like this, make sure you give this one a thumbs up and subscribe to this channel. I'll make sure you see more videos like this one in your news feed. I'm doing AI News videos every Friday where I'm going to bring you up to date with all the cool news in the AI space every Friday. So if you want to stay in that loop, make sure you subscribe to this channel and I'll keep these videos flowing to you. Thanks again for tuning in. Really, really appreciate you. See you guys in the next video. Bye. Bye.