In today's video, I'm going to show you how you can leverage the new Chat GP's image model to create full shot animated themes with consistent characters and consistent scenes across board. We will use Chat Jet to generate all the visuals and then use a separate tool which I'm going to talk to you about in a second to do all our animations. This is a follow-up to my previous tutorial where I showed you the various options you can animate your chat GPT generated images.
But in this tutorial, we're going to take it a step further to show you how you can put everything together to create short animated stories leveraging chat chip. I'm going to put a link to that tutorial in the description, the previous one. I do advise you watch it right after this one.
But let's get straight to this tutorial. So when generating full animated stories, there are couple of things you need which CHP can actually help with. Obviously the first thing you need is a script.
Your script would need some characters. So, you would have to do some character design to design what your characters would look like. Then, for each scene in your script, you're going to need a shot list, which simply describes the different shots you're going to have in that scene.
For each of those shots, you would need a text to image prompt. So, you can use that to generate what that shot would look like visually. And for each of those images generated from the text to image prompt, you also need a image to video prompt which you can copy and put in any video generator to animate that shot and bring it to life.
Finally, some of your scenes might have some dialogue. So you're going to need the exact text for that. So you can put that in a text to speech generator to generate your audio.
Now I could sit here and walk you through how to generate a prompt for each of these which would probably have made this video longer. Instead of that, I've bundled all of these into a GPT. So, this GPT sort of helps you with everything that I've described right here.
You just tell it what you're trying to generate, and for each scene, it will generate all of this for you step by step, and you can just copy and paste. It's not perfect, but it gives you a good building block to start building your animations on. So, for the rest of this tutorial, I'm going to show you how to do these full animations using this GPT.
And let's get straight to it. So once you head over to chatgptt. com, you want to head over right here to the left and click on explore GPT and then simply search for animation script builder.
So you're going to see it right here, this one by Emmanuel Crown. So click on that and then simply click on start chat. So now right here in the chat, you can describe what kind of project you're trying to work on.
So you can put your entire script if you already have one there or be as detailed as possible. In my case, I'm going to simply just say, "Write a 3D cartoon animation script about a boy getting ready for school. " It's very vague because I just want to do something here for a quick test.
So, as you can see, it's gone ahead and asked some clarifying questions. How many characters do I want in the script? What does the boy look like?
Um, any other characters that we have in the scene, what do they look like? And what's my preferred animation style? It says 3D.
So, in this case, I'm going to say there should be two characters, the boy and his mom. And then I'm going to ask you to just go ahead and come up with the rest of the details cuz I'm just doing a demo here. So as you can see it started first by generating our characters.
So it's described the character for the boy and the character for the mom. So that's the first step. So next I'm going to say go ahead and create the image for the boy for the first character.
So you see it's gone ahead and generated the image for my first character. But if you notice right here, there's a little uh disclaimer there that says made with the old version of image generation. So not the new version.
So for now, OpenAI does not support the new image generation model inside of uh a GPT. It's still using Deli. So the images you're going to get from this are not going to be that great.
So I'm going to show you a workar around for that. But as soon as they support that in GPTs, I'm going to update the GPT to be able to generate the images right inside there. So you don't have to go out of this chat.
But let me show you a quick walk around for that. So, I'm going to go ahead and ask it to go ahead and create the script for scene one. I'm doing it scene by scene.
So, as you can see, it's broken down scene one right here. It's given me uh each of the shots. So, for shot one, there's a little bit of a narration.
So, it's put the narration right there. Um, a little bit of the description of the scene. Uh, the text to image prompt is right there.
And then also the image to video prompt is all right there. I can copy this, put in any of my image generator. copy this, put in any of my video generation to have the final um scene.
So, all of that looks good. So, what we're going to do instead of going right here in this same chat to continue to generate um the images, I'm actually going to go ahead and copy all of this script and everything that this GPT has generated. And then I'm going to head over and start a brand new chat.
And then once you click on the new chat right here, you want to make sure you are on GPT40. I'm going to then paste this script in here. And make sure at the top also you paste the description for the two characters uh that it generated or however many characters you have in your own script.
And the only reason once again we're doing this is just because the new Charg image model is not supported from within GPT and we want to have access to this model because it's the one that's going to help us with the consistent characters and the really nice looking images that Chad Gypty can do. Now, so that's the only reason we're doing this. Um, once that comes to GPTs, you wouldn't have to do this.
You can continue the rest of this process within the same GPT itself. So, the prompt I'm going to add below that is create a front view of the image of Ben on a white background. So, that's the first character and that's what that looks like.
So, that's the first image. I'm going to go ahead and ask it to generate the image for the mom. So, it's generated that and that's what the mom is going to look like.
So, now great, we have our two characters generated. Now let's go ahead and generate the image for each and every scene. Remember the GPT already described what each scene should look like.
So I'm going to type in the prompt create the image of shot one. Remember it described each of the shot list. And then I'm going to put here three is to two.
So that's the aspect ratio. So that's the widest the chat GPT image model can do as of today. So it at least looks almost like a 16x9.
Um hopefully they update that soon. So, I'm going to hit send to generate that. So, this is what it came up with.
So, this is the first image for shot one. I'm going to go ahead once it's done with that and say create the image for shot two. And that's what that looks like.
So, I'm going to go ahead and continue to repeat this. Create the image for shot three, shot four, and shot five. So, for all the shots that I have in the script, and that's what it's done.
It's generated all of this for me. So if you notice throughout our characters, both the mom and the little boy have remained consistent across the scene and I haven't had to do so much work with that chat just remembers all the context from the chat and is able to maintain uh the consistency of the characters even the clothing and even the environment of the room. Now, if you have a very long script, by the time you start getting to scene 8 10, it might start losing some of the coherence and what the character looks like.
So, if that happens, there's a simple fix for that. Just go back and re-upload that image of the character in one of your prompts and just ask it to generate uh the image of say shot 52 and then just put the image of your character as a reference. So, it really remembers what your character looks like.
And you can keep doing that. And anytime you have the coherence of your character being lost, just re-upload the image again. And that should uh continue to look consistent across the scenes.
All right. Now that we've generated the images for every scene, so I've done all of that. So I have all my images now and my scenes are ready to go.
Now we need to start bringing them to life by animating them. There's several tools you can use for this and I talked about that in the previous video. But for this particular video, we're going to be using Clean AI.
And I'm going to show you also a little bit of Runways Gen 4. Uh those two models do very good at animating um sort of the stylized 3D stylized uh cartoon also stylized um sort of images. So I've opened clean AI right here.
And I'm going to go right here and click on the upload image option. And now I need a prompt to give to Clean AI. So remember our GPT has generated everything for me.
So I'm going to go back to what the GPT generated. And for shot one, there's right here already a prompt for image to video prompt. So I'm just going to copy that, go back to clean AI and paste it.
Remember, like I said, it's not completely perfect. So I'm just going to make a few edits to remove some things that I don't want. But it's just again a building block.
You can tweak it however you want, but it kind of makes the work much easier for you. You don't have to keep thinking, okay, what prompt do I have to use for this particular scene? So, in my case, I've removed some of the prompts that it had there, and I'm going to add a little bit of a camera movement to just have the camera zooming in.
And then I'm going to hit on uh generate. But before I do that, um Clean AI now has clean 2. 0.
You might have heard about that. Uh from my test, when it comes to sort of uh especially like the 3D animated style images, you do not get that much of a difference from 1. 6 for most of your shots.
So there are shots where it becomes very good cuz it's very good at prompt adherence and especially when you have shots that have a lot of fast motion movement. So I'm going to show an example. So this is an example I have of one of the last scene of Ben running out of the room.
So this is with clean 1. 6 and I've done the exact same thing on cleaning 2. 0.
You can see that's a much better animation, much smoother. We don't have that much artifacts being added to the animation. I would advise start with 1.
6 six for most of the times uh you use much less credit for that and you might for the most of it get what you want. So in this case most of the shots here apart from that one last shot was generated with 1. 6.
So I'm just going to hit on generate and this is what that scene looks like. So as you can see it's a very static scene. There's not that much movement.
So there was no need for me to use uh clean 2. 0. So I've gone ahead and repeated the same thing for each scene.
So here just a couple of examples. So, the next thing now for me to do is to start to generate some of my dialogue. So, I'm going to go right here uh where we have scene two and the dialogue for Ben's mom.
It's already generated what that dialogue was. I'm going to copy that entire dialogue and put it to any text to audio generator. In my case, I've spoken about that a lot of times in this channel.
11 Labs is my go-to. I still think it's one of the best. So, I have 11 Labs open right here.
And I'm going to paste up that. Uh this is a narration. I'm going to paste that there and just generate that and it will generate my dialogue or my narration.
And that's what that sounds like. It's morning at Ben's house and someone is not quite ready to get out of bed. So, now that I have that, now we need to deal with how do we do the lip sync in some of the scenes that have lip syncing.
So, uh I can still do that in clean AI. It's not the best in most cases, but I'll show you how that works. So, I have this shot right here I've generated of Ben's mom.
uh sort of telling him it's time to wake up. So on the shot right there that clean AI generates, you can see there's the lip sync option right here. Okay.
So I'm going to click on that again. I'm going to go to upload local dubbing and I'm going to upload the file that I've generated using 11 Labs or if you have your voice actor do that for you. You can use that from your voice actor too.
And then once I've done that, I'm going to just hit on generate. Ben, time to get up, sweetheart. And if you watch my previous video, I've shown you other tools that you can use for lip syncing.
So I do advise you to watch that. You can play with those in connection with what we've shown here today. But one tool I didn't show in that video, and that's because it doesn't have much of a big generous free plan.
In that video, I was trying to cover some tools that gives you at least uh maybe monthly free credits or daily free credits. Um, Runway Gen 4 is very great at animating this uh sort of stylized images and it's also very great at the lip sync, but they give you a one time I believe it's 125 credit for the lifetime of your account. So again, if you're looking for something free to use, um, Runway might not be uh the place to go, but I'm going to show you how to do the same thing we've just done right now in Runway.
So there are two ways uh to sort of do this. Uh the first way is again once you're on runway, this is runway ML. I'm going to click on generate video.
So just like the way we did in clean, I'll click on the image and then again I can paste that same uh text to video prompt from my GPT. And we have different options from Turbo to Gen 4. Uh Gen 4 and Turbo are very good at this.
And then we're going to go ahead and just hit on generate for it to generate the animations for this. Now once it's done that there are two options to do the lip sync. So just the way clean AI has a lip sync option you can click right here on the actions and then you see there's an option right here for lip sync.
So I'm going to click on that. It's going to try to detect a human face and then right here you can upload your audio file and then I'm going to say generate lip sync. Now I'm going to play both of them so you can see the comparison between Runway ML and Clean AI.
So this is what they look like. Ben, time to get up, sweetheart. Ben, time to get up, sweetheart.
The second way to do your lips sync using Runway is using Act One, which gives you the ultimate best sort of lip sync whenever it works cuz it's very natural. You drive the performance. So, I'm going to head over back to Runway ML and click on act one right here.
So, the first thing you want to do is you can upload an image of your character or an actual video with some animations. So, in this case, I'm actually going to use the video cuz I want some of the animations to be on it. So, it could be either the video you generated using runway or the video you generated using clean AI.
So, I'm going to click one of Ben's mom right here sipping the coffee. So, it's going to take a couple of seconds to try to detect the face in the in the video or image. So, next you want to put a recording, a video recording of somebody acting out what you want the character to say.
You can either record that live or actually upload a pre-recording. So, let's attempt to do that live right here. You see, it has the camera right now trying to capture my face.
So, I'm just going to adjust that and then I'm going to hit record. Well, you said the exact same thing 10 minutes ago. Come on, Ben.
It's time to go to school. Wakey wakey. And in case you're wondering how I changed my voice, that is still done with 11 Lab.
So, if you're on 11 Labs, right here you'll see the voice changer option. So, just click on that. And then right here, upload whatever audio you have of your original recording.
And simply select any one of the voices that they have right here that you want it to sound like and click generate. And it's going to change your voice to sound like that voice. So now that I'm done with all the shots, so we've done the images, we've done all the animations, and we've done all the lip sync in cases where they're lip sync.
So the final thing is just to put everything together. And you can use any of your favorite video editor. I typically use Premier Pro, but for the sake of this video, I'm going to use Caput, which is a free video editor, and you can download Caput for the desktop from captop.
com, or you can also download it on your phone. So, I'm right here on Cap Cut. I'm just simply going to drag all my im uh videos, put them together, and cut them out and just arrange them right here.
As you can see, scene by scene. So, it's just by dragging, dropping, and using the tool right here to cut whatever part I don't want. And then the final part of all of this, which is very important, is to do the sound design.
So, I'm going to add the background music, add a little bit of sound effects and sound design to bring everything to life. And I'm going to show you the final product in a second. But before I do that, if you enjoyed this video, kindly do me a favor by hitting a thumbs up so that YouTube can recommend it for other people.
It does go a long way in helping me grow the channel. And if you did enjoy this video and want to see more content like this, please do make sure you subscribe to the channel. and I'm going to make sure more content like this show up on your timeline.
Thank you for stopping by and thank you for watching. I'm going to catch you in the next video. Keep learning.
Bye. And here's the final product. Ben, time to get up, sweetheart.
Five more minutes, please. Well, you said the same exact thing 10 minutes ago. Come on, Ben.
It's time to go to school. Wakey wakey. I made bacon.
Bacon. Okay. Uh, I found my lucky shirt.
You can't wear the same shirt three days in a row, Ben. But I love it. Walk outside.
