How To Create Cinematic Ai Videos - AI Filmmaking Advanced Course

AI film making is absolutely fascinating. Some say it's destroying

fascinating. Some say it's destroying creativity, but for me it's the exact opposite. It allows me to visualize any

opposite. It allows me to visualize any idea that I can imagine, but yet for most people getting started with it feels overwhelming. They don't know how

feels overwhelming. They don't know how to prompt. There are too many tools and

to prompt. There are too many tools and they don't know how to use them. That's

why in this video I will show you everything that you need to know to make your own AI movie like this.

I heard the shot. Did you get it?

Dan, you didn't miss, did you? You never

miss. And you owe me this favor.

Dan.

>> Dan.

To make your very own AI movie, there are three steps that you need to follow after you've established your exact idea. The first step is generating your

idea. The first step is generating your scenes. After that, the second step is

scenes. After that, the second step is animating those scenes to bring them to life. And the last step, maybe the most

life. And the last step, maybe the most important step, is doing post-prouction.

That is editing it, adding sound effects, adding all kind of different effects to make your video come to life.

Now, make sure to not skip any of these steps. And now, let's get started with

steps. And now, let's get started with generating our own scenes. For this

movie that I made, there are two main characters that consistently reappear in these scenes. You have me as a sniper or

these scenes. You have me as a sniper or a hunter, and then you have the unicorn.

So, the first thing that you need to do when getting started is defining what your character is and making a hero shot. It does not matter what tool

shot. It does not matter what tool you're using. For me, I'm using an

you're using. For me, I'm using an all-in-one tool like Open Art because I can use tools like Nano Banana and Seamream, but I can also use all the different image and video generators

like Google V3.1, Clink 2.5, Sora 2. All

of them are in this tool. So I don't have to switch and it allows for some consistent workflow of generating my different scenes. Now to make your own

different scenes. Now to make your own hero shot, I'm going to go over and I'm going to use Nano Banana for this. In

the prompt box, you can simply prompt what you want your character to look like. So for the white unicorn, it is

like. So for the white unicorn, it is drinking from a shallow jungle stream at dusk, soft rays, and the morning sunlight filtering through the canopy and glinting on the rippling water. So I

have like all of these different ideas.

I used a bit of chachi with this. And if

you want to find out all the prompts that I use throughout this video, then make sure to sign up for my free community. And you can find all of these

community. And you can find all of these different prompts, including some more in-depth tutorials in that free community. There's no strings attached.

community. There's no strings attached.

It's literally free. It's the best place to hang out with other people that are enthusiastic about AI. So, make sure if your video is in 16x9 to select that widescreen ratio and then make a number

of different images. I always do four because the more the better. It gives me more different results I can pick from.

So now they gave me these four shots and out of these ones, pick your favorite and use that as your hero shot. So for

the other character, which is me, I used this image that is an AI generated image of myself that I used for a thumbnail.

Now I put it in there as an omni reference and then I gave it this prompt. So I'm literally saying keep the

prompt. So I'm literally saying keep the same person, pose, lighting, composition, change only his clothing.

So I wanted to have my face and all of that consistent. I just wanted to have a

that consistent. I just wanted to have a different outfit. Then also for this one

different outfit. Then also for this one I kept it at 16 by9 and then I hit create and then it got me this image. So

now we got both of our hero shots and these are the scenes that will be used consistently as an image reference throughout all of the generations that we do. So the next step is using Nano

we do. So the next step is using Nano Banana to generate different type of shots from your main characters. So for

example here I have the horse like that.

Now I want to have a topdown aerial view. So, I literally gave it a prompt

view. So, I literally gave it a prompt with this as the image reference. Change

the camera to a top- down aerial view looking straight down at the unicorn drinking from the stream. Now,

everything remains consistent. We just

have a top- down aerial view for this scene. And you want to do this a couple

scene. And you want to do this a couple time to get all of your scenes from your character. For example, here I have a

character. For example, here I have a bit more of a close-up. Then I have another one with the sniper scope with a POV on it. Again, the prompts for this are super simple. It's literally asking

Nano Banana to change the camera view to whatever shot you have in mind. So,

changing the unicorn is easy, but now getting into this shot is a bit harder, but still not impossible. What I did is I used this shot as a reference. Then I

gave it the prompt, which is a cinematic close side view of a mill hunter lying prone on the jungle floor, aiming carefully through the scope of his rifle. As you see, I'm describing

rifle. As you see, I'm describing exactly what I'm doing. You cannot just say like a dude laying on the ground looking through a rifle. While it still might work, it won't give you the exact result that you want. By adding all of

these different details, you can get a consistent result and you can match that with all of the other scenes that you already have generated. So, as you can see, for me, using that same prompt with

that image reference gave me this result. And this is going to be the next

result. And this is going to be the next shot that I'm also using as like an image reference in the rest of my shots.

Again, as you can see right here, we can use that original image as a reference image using nanom banana again. And then

we say the hunter looks up in the forest is a little bit darker. And now we have our next scene. We can repeat this process. For example, you can also put

process. For example, you can also put in multiple different images. So here I put in that image plus another image of me to keep the character consistent. And

then I change the prompts to change the camera to a zoomed in straight aerial top. And now we have this different

top. And now we have this different angle of me. Now, doing this a couple of times gave me all of these different angles and images of me, which we can now use to make or complete storyboards.

By the way, the only image that I didn't make through Nano Banana was this one right here. I just couldn't get it to

right here. I just couldn't get it to work, so I tried a different model. And

that's also the perk of using a tool like Open Art where you can switch to, for example, Honeyan or any other model to get the exact shot that you want.

Okay, so now I've put everything into my storyboard. I'm using Figma for this.

storyboard. I'm using Figma for this.

You don't have to. you can just simply use a Google doc. But for me, it's easy to visualize everything that I've done, what I want to have happen, and how I want to prompt it to get the exact result that I want. So now we can move

over to the next step, which is animating all of our scenes. For the

animation, we're switching over to video. Here we're going over to image to

video. Here we're going over to image to video. Now I'm primarily using Cling 2.1

video. Now I'm primarily using Cling 2.1 because it has start and end frame and Google VO 3.1 also because it has start and end frame and I want to make it look seamless. The other tool that I've been

seamless. The other tool that I've been using is Sora 2 and sometimes I also use Cling 2.5. So for the first image that I

Cling 2.5. So for the first image that I want to show you, we're going to use Cling 2.1. For the start frame, we'll be

Cling 2.1. For the start frame, we'll be using this image right here of the horse. Then for the end frame, we switch

horse. Then for the end frame, we switch it over. And now we're going to use this

it over. And now we're going to use this image of me as a hunter. And we want to have the transition happening of looking at the horse and then going straight to the hunter. So here it is very important

the hunter. So here it is very important that you write what you exactly want to see. So, we're saying the white unicorn

see. So, we're saying the white unicorn instantly raises his head at the tension, then continues drinking from the steam with falling leaves. Then,

after a few seconds, the camera zooms forward rapidly past the unicorn to a far distance, revealing a person in one continuous shot. So, that's the

continuous shot. So, that's the transition that we want to have happen.

For this one, I set the duration to 10 seconds. And for this one, you can only

seconds. And for this one, you can only use the pro quality mode, and that's fine. Like, you can just use this. Then,

fine. Like, you can just use this. Then,

you click on create, and that got me this result.

The thing I love about cling is how fast that motion is. Keep in mind though, you can always like change this or speed this up in the editing. So, for example, this part, I've sped it up a little bit to make it a bit faster. Now, the thing

that I love about Open Art is how you can extend this scene. So, let's say you want to extend this and you want to have the same smooth, seamless transition.

What you can do here is you can click on grab frame to video. This is not actually what I did, but I just want to show you right now. So, this is a technique that you can use to get those seamless transitions in between shots.

So, what you can do is you can grab any frame or you can grab the last frame. By

grabbing the last frame, it is using that as a start frame on my next project. So, now we have that last frame

project. So, now we have that last frame that we can start and use in the next scene if you want to have that seamless transition. Now, let me show you another

transition. Now, let me show you another scene that I find super interesting and where you can use a bit of a different technique. So, here I got inspired by

technique. So, here I got inspired by this series, the day of a jackal. We

have this guy looking through his scope and I found it looking super cool. So, I

wanted to recreate that scene myself.

Now, I got this image over here of me looking through the scope. I got to tell you like the eye here is not doing it correctly. Like it should have been the

correctly. Like it should have been the left eye here that is going to look through the scope. Either way, I still found it a cool image and I still wanted to use it. I was also a bit of on a time constraint. So, what I did is I used

constraint. So, what I did is I used that image. I gave it a prompt where

that image. I gave it a prompt where he's slightly moving his face around while he's still looking through the scope. And then that gave me this

scope. And then that gave me this result. Now, it's not perfect yet. We

result. Now, it's not perfect yet. We

don't have that like blurring around the scope. But we can fix this in

scope. But we can fix this in post-prouction. Now, for scene 7.5 and

post-prouction. Now, for scene 7.5 and 7.6, I got these two close-ups, and I want to have different angles. So for

this I mixed around with VO3.1 and cling 2.5. Now for the first image I actually

2.5. Now for the first image I actually used VO3.1 and then for the input I used this image of the horse and then I just set it to have like a fixed camera angle

and the horse is just drinking some water from the stream. Now the great thing about this is it also generates sound. But don't worry though if you

sound. But don't worry though if you don't have sound generated because you're using like Clink 2.5. Clink 2.5

also doesn't generate any sound. We can

still add sound effects at the end in post-prouction. So, this is the shot

post-prouction. So, this is the shot that we got from VO3.1. And keep in mind, this is done with the fast model because I wanted to save some credits.

Now, for the second shot, I actually used Cling 2.5 because in my opinion, it has better prompt adherence. And for

this one, I wanted to slowly close it eyes, and it actually did that perfectly. Like, it's like, damn, I'm

perfectly. Like, it's like, damn, I'm going to get killed today. Like, that's

that's how the unicorn was looking.

Yeah, he was screwed. Then we have the aftermath when I took the shot. So I

don't actually have me taking a shot on film. I just have this scene where I

film. I just have this scene where I look like up surprised and then I made a cut because like this part kind of got like screwed over. So I didn't want to use that. So now we have like all the

use that. So now we have like all the scenes from us taking the shot and then we get the phone call. And the phone call is actually interesting because here we got more things going on that I think are worth mentioning. The first

thing I used is Cling 2.5 again for better prompt adherence because we got slight of few things going on here. We

have the camera rotating and then we also have me getting out my like iPhone out of my pocket. This only worked using 2.5. I also picked 10 seconds so we have

2.5. I also picked 10 seconds so we have like plenty of time to have this action happening. For the next scene where we

happening. For the next scene where we have the close-up of me holding the phone next to my head, I actually used a bit of a trick for that. So what I did is I took a screenshot of like this

segment right here. Then I put that into Nano Modenna to like upscale it a little bit and to refine the details a little bit better. And then I used it as a

bit better. And then I used it as a reference right here. So we have like the original image right here. And then

I gave it my other prompt with cling 2.5. So tracking camera. The man slowly

2.5. So tracking camera. The man slowly straightens himself standing up holding his phone close to his ear and he does not say a word. He just maintains a blank deadband stare forward. Same thing

here. I used Nanomana to get a different angle from me being on the phone. Moving

on, we got this scene where the unicorn is appearing. So, the unicorn is not

is appearing. So, the unicorn is not dead. Like, that's the whole plot twist

dead. Like, that's the whole plot twist from this movie. So, we have me standing there, bit flabbergasted, and then I put the phone down and then we have this end frame of the unicorn being there. This

is simply done by just using Nano Banana and adding in a unicorn there in the background. Then using start and end

background. Then using start and end frame and giving it a very slow cinematic pen. I did the same at the

cinematic pen. I did the same at the last shot where I have like the tree and that is the complete video. Now that you got all your scenes ready, before you put them into your editor, you want to

upscale them. Now, if you go over to the

upscale them. Now, if you go over to the video, so for example, you can click on any type of video. You can already say video upscale right here. This will

bring you over to using Topaz Labs where you can upscale it to any type of resolution, for example, 4K. And you can select your frame rate. Now, you can enhance the video and this will give you

better results. You don't have to do

better results. You don't have to do this, but it will make your video look a bit nicer. After that, it is time for

bit nicer. After that, it is time for the last step, which is postp production. And for this, you can use

production. And for this, you can use any type of editor. You can use Cap Cut, Dainci, or in my case, I'm using Premiere Pro. Now, once you're in

Premiere Pro. Now, once you're in Premiere Pro, you want to import all of your videos. Then, what you want to do

your videos. Then, what you want to do is you want to put all the clips together, and you want to make a few cuts, a few transitions. With most

movies that you watch, there are not many long sequences, unless you're watching the movie 1917. That one has like a 8 minute long sequence or something. So, in my case, I wanted to

something. So, in my case, I wanted to keep it short and like fast-paced throughout some of the parts. You have

slower parts and faster parts. Now, for

the beginning especially, I wanted to have a few long shots because I wanted to slowly ease into it. So, for that, I used a bit of sound effects to make it build up the movie. Then I added in some

blurs and some speed rims to also fix a bit of the minor AI mistakes that I had.

For example, when this scene of the hunter is revealed, I didn't quite like how it was pretty pixelated, and I wanted to fix that by using a bit of a motion blur so you don't see it until

the very last moment. I also added a blur around the sniper scope because I wanted the viewer to look as if you're looking through the sniper scope. And

that's a trick that filmmakers use while editing a video to put some direction on where they want you to actually look.

Then we have sound design. And this is a whole world in his own. This could be a separate tutorial. And in fact, if you

separate tutorial. And in fact, if you want to have the more indepth behind the scenes of how you can edit your videos, then please make sure to join my free school community in there. Me and my

editor go through the video and he will help explain how we edit in the sound design for this video. Because like for this we used 11 apps and we wrote all the different sound designs. As you can see on the timeline it is quite a lot to

go over but you want to add in like multiple different sounds layer them on top of each other to really bring everything to life. Lastly we have adding in the voice over. For that I used 11 lapse again. And then we have

final touches which is a bit of color correction. So now you have completed

correction. So now you have completed the editing. You can export your video

the editing. You can export your video and then you have completed your very first AI movie. I know there's a lot to this and there's still a lot of different small steps that you have to

tweak in order to make it perfect. But

hey, this still allows you to make everything yourself. There's no editors

everything yourself. There's no editors needed. You don't need any actors. You

needed. You don't need any actors. You

don't need to go to any location. You

need any camera gear. You can make this if you have any type of ID in the back of your head. You can visualize this quickly using AI. Keep in mind the tools that I'm using now might be outdated

quickly as AI is only getting better and better. The good thing is by using an

better. The good thing is by using an all-in-one AI tool like Open Art is they constantly update with the latest new AI tools. So, for example, when VO4 is

tools. So, for example, when VO4 is coming out or the next cling or whatever, you will have it on that tool.

I will leave the link to open art in the description down below alongside with a link to my community. So, you can join it. You can go through all of my

it. You can go through all of my prompts. You can see the behind the

prompts. You can see the behind the scenes of how we made this and you can continue to chat with me over there.

Click the video that's on the screen right now if you want to learn what the best all-in-one AI tool is for you. I

compare all of them and I found a few that are pretty good. So it's worth it checking