Easily Create Long Videos with Consistent Characters using VEO 3 (VEO 3 Tutorial)

What would happen if I jump off this

balloon? Of course not. I won't be that

stupid to do that.

Did you notice every other thing changed

in this video apart from me? As you can

see, I have been consistent throughout

this video in different scenes. Don't

worry, Charles, my creator, will show

you how he created this long video with

just VO3. Over to you, Charles.

[Music]

Lovely.

What I'm about to show you right now,

nobody else is doing. I was working on a

project when I figured out how to use V3

to create long videos with consistent

characters. Now, V3 has a flow feature

that can be used to make long videos.

But the problem is it only works with

V2. And with V2, the visuals are not

high quality and there's no sound,

making it a pretty terrible feature to

use. An even bigger problem is that V3's

frame to video feature doesn't allow

users in certain countries to upload

photos containing people which you then

use to create videos. It's annoying. It

didn't even work for me and I'm in

London. Strange.

But don't worry. I found a simple way

around it. In this video, I'll show you

an easy method to create long videos

with consistent characters using V3. And

this works in any country as long as you

have access to V3. My name is Charles.

Let's do this.

The first step is to open V3. Now, if

you don't know how to access V3, just go

to Google and search for V3. You're

going to see the first link right here.

Right click on this first link and then

click on try flow and click create with

flow. And boom, you have access to V3

where you can create your videos and all

that. Now the normal method people do is

just to go to V3, put in a prompt and

then generate the video. Now I'm going

to show you a method. I'll first show

you the normal way of doing it but with

a specific type of prompt. So listen,

the first thing we're going to do is to

go to charg. So we're going to tell GBT.

So based on the idea that I have, I'm

going to say write an image prompt to

generate an image of a man standing

inside a hot air balloon in the sky. And

now shajib will give you a prompt. And

this is the prompt that you can use. But

don't go ahead and use this prompt. Let

me explain. You're going to do this next

step. You have to describe the character

and use it with this method I'm about to

share with you. So you're going to say,

describe all characters in this scene.

En sure you add the name, physical

description which has the skin color,

hair type, nationality, age, gender and

outfit which has a top wear, bottom

wear, footwear, jewelry including outfit

color of all characters in a prompt one

paragraph. Now if you send this what

will do is it will give you the

description of the character in that

image idea that you have given it to.

Right now, if you want to create a long

video or a consistent character, you

must have like a character in mind. So,

you do this to get the description of

the character. Now, you can see that

Chip has described the character inside

the prompt, the image prompt. But we are

going to do something. We're going to

split this prompt into two. We want the

character description to be separate

from the action. So we going to tell

Chajibility split the prompt into

character and action. And boom,

Chajibility has split the prompt. This

is the character description which is

very important and the action. Here we

can just say what we want the character

to do, which is what we want. We've

gotten the first part. Now we're going

to use this to generate an image. So

we'll take this copy this prompt here

and then we're going to change this from

text to video to frames to video. So for

this demonstration for this video we'll

be using frames to video throughout.

Right? So once you have changed it to

frames to video you click on this plus

here and then upload an image or you

just generate an image. Now, if you want

to upload an image, you could actually

do it. But if the image contains a

person or a face or something that

identifies a person, Google View would

reject it, right? This part of their

policy thing and all that. So, in some

countries, they don't allow people to

upload images and use those images to

generate a video for some reason.

They'll rather have you generate the

image on their platform and then turn it

to a video. So, we going to click on

generate image and then we're going to

enter a prompt here. So the prompts that

we copied we are going to enter it here

and then click on generate here and boom

it has generated four images. Now we

have these four images that we can use.

We have our character set. Okay. So what

we can do is to click on any one. We can

just click on this and then click on use

this image. And you can see the image

has appeared here. So the next thing

we're going to do is to add a text

prompt that we can use to generate a

video. Now if you've not subscribed to

my channel, subscribe. There are a lot

of helpful videos here. Like this video,

share and also comment. All right. So

once you have this, what you're going to

do is to go back to chip. We are going

to tell to turn this prompt that we have

to a video prompt in V3. Now I'm going

to say write V3 prompt following the

same template. And then I give

jajibility the idea of what I want the

person to do in the video and then what

I want the person to say. For example,

itan should say what would happen if I

jump off this balloon and then jumps off

the balloon and then I send this to

charge and boom chip has generated v3

prompt for me. So I don't need to think

about anything. And what you want to do

in all your prompts is to split it to

have character and action. So even if

you have three characters in the image,

four characters, you want to split the

character section part where you define

all the characters and then the action

what the characters are doing. So now

once we have this we going to copy this

video prompt and then go back to V3 and

paste it here. So you can see we still

have our first frame which is the image

we generated and then we have the

description which is the character and

action and is on frames to video. Now I

used V3 fast for this but you can use

fast or quality depending if fast is not

giving you the best result switch to

quality. So I can't say oh use fast or

use quality. Sometimes quality your 3

quality gives you a bad video but then

when you use V3 fast it gives you a

better video and then sometimes vice

versa V3 quality gives better and then

fast gives like a tacky kind of video.

So depending on the situation just

change but I'm going to use V3 fast for

this. Now we have this. I'm going to

click on generate. And boom, we have our

results. Here we have the first video.

So let's play this video. What would

happen if I jump off this balloon?

So So you can see we have our first

video. Now this is what most people do.

And then once they have done this,

they'll now go back to flow and try to

extend it in flow. They'll take this

video to flow and try to extend it. If

you don't know how flow looks like, let

me show you how flow looks like. So, for

example, this is flow. They'll take you

to flow here and they'll start

extending, they'll click on this plus

sign and then click on extend,

right? To create an extension of this.

But you can see here if I try to

generate jump off, let me just write

this. If I try to generate, it says this

feature isn't supported on V3 fast.

Switch to V2 fast. So V2 fast is not the

best option to use for this because it

doesn't give you high quality video. It

doesn't give you sound. So forget about

flow. Just watch this method I'm about

to show you. Right now we have the first

video. How are we going to make this

consistent throughout every other video?

This is the trick. Now what you're going

to do is to go back to frames to video.

Make sure that here is frames to video

and ensure that this is in view 3.

Either fast or quality works well for

view 3. And then you click on this plus

sign here for the first frame. Right.

And then click on generate. Now what

we're going to do this time around since

you can't upload an image that has a

person in it because of the policy

guidelines. What you can do is to

generate your characters. All of them.

You have to generate all your characters

in green screen or white background.

Right? So for example, I'm going to

generate this person again.

All right. To start the long video, I'm

going to generate this person again on a

green background. I'm going to generate

the front shot and the midshot. Let me

show you what I mean. So I go back to

Chpt and then enter this character

description here of the person. And then

I said write a prompt to generate an

image of Eton on a plain green

background. Now you can use plain white

background, plain green background,

plain anything. Right? So it will be

easy for the AI to remove detach the

background. Now once we have this, we

send the prompt and boom, it has given

us a prompt that says full body portrait

of Eton. This this this this and then it

says with a plain solid green background

studio lighting, which is what we want.

So we have the prompt to generate full

body portrait of our character in green

background. Now we take this a bit

further. Next we're going to say write

for midshot from waist up because we

want a shot where the person is just

close just from waist up and boom chip

has given us prompts from waist up. M

short portrait waist top of a tan and

blah blah blah blah blah plain solid

green background studio lighting what we

want right so once we have this we head

over to charg we are going to copy the

first prompt copy the full body prompt

so now we have this we just going to go

back to v3 we take the first result

which is the full body portrait prompt

and then we come back to v3 to generate

the image we paste it here and Then we

click on generate and boom it has

generated four images for us. Full body

of the person right. So you can see we

have this. So I can just select this

person here and click on use this image.

So I I can use that image. Now to

generate the second image you can just

cancel this. We are not ready to use it.

Click on this plus sign and then click

on generate image and then go back to

charity. Then copy this midshot prompt.

So we're going to paste the midshot

prompt here. So it will generate another

set of images and then we click on

generate and boom, we have another set

of four images ready for us to use. You

can see all these images look good and

they look the same. So I'm going to use

this image right here. I'm going to

select use this image. As you can see,

Google View has added it to this first

frame. Now, you can go ahead and start

generating whatever you want. So, what

we're going to do, you can even cancel

this and then go back. If you click on

this plus sign, it's going to take you

to the library where you can just select

any of the ones that you have generated,

right? The ones that you have selected,

use this image. So, these are the two. I

can just decide to use them. So, I'm

going to use this one first for now. I'm

going to click on it

and you can see the image here the first

frame. The next part is the prompt which

is very important because if you just

put a prompt and say oh generate this it

will just give you a funny result but

there's words but there are keywords

that you can start with that will tell

v3 to switch up and remove the

background. So we are going to go to

charg enter this prompt write v3 prompt

following the same template instruct v3

to change the background in the uploaded

image write new background. So charg

tell your tree change the background and

then you write new background and then

here I gave it the idea of what I want

you to do and say itan walking on the

streets of London says of course not I

won't be stupid to do that then walk to

shake someone like a gangster then we

send this prompt and boom charge has

given us a prompt a v3 video prompt to

use as you can see in this prompt is

specified replace the plain green

background in the uploaded image with a

lively London street scene which is what

I want. Streets of London which is what

I want the background to be right

walking in the streets of London. So it

has changed it here in the prompt and

then it says um what the character would

say I want you to say is also here. Of

course not. I want that stupid to do

that. So I'm going to copy all this.

Paste the prompt here and click on

generate. And boom we have our first

video.

Of course not. I won't be that stupid to

do that.

So now this is view 3 fast I use. Let me

switch it to V3 quality. So I can just

switch it here and select the image and

enter the prompt. Okay. So now we have

told you to replace the playing

background and then we send the prompt

again. Boom. We have our video. Let's

play it.

Of course not. I won't be that stupid to

do that.

Now you see this quality looks better.

I'm going to play the full video later.

We have this first video. Now, how can

we make sure that this character is the

same in the next video? Simple. We just

use the same image. We'll just keep

using the image and then changing the

background. But now, let's try a

different style of changing background.

Which is the most

professional way of doing it, right?

Which is the best way of doing it? We're

going to tell right v3 prompts following

the same template. instruct v3 to change

background in the uploaded image at the

start of the video in the first frame

write the new background and then I said

oh he seated in the back of a car seat

and it says did you notice every other

thing that's changed in this video and

blah blah blah so if you noticed I said

start of the first frame so I'm telling

V3 as you generate this video the first

frame should show the new background

don't wait for like two seconds or one

second, which is what he did in the last

video. So, this is the perfect prompt,

right prompt that you need to use in

your generations. Boom. We have this. We

have the character and we have the

prompt here, the action. As you can see,

it says in the first frame of the video,

replace the plain green background in

the uploaded image with the interior of

a moving car's back seat. You can see

here I told Jajibility that I want the

character to sit to the back of a car

which is what he did here. So we copy

this prompt which looks absolutely

beautiful. Select the image. We can use

either of these but let's keep using

this because it's a midshot video and

then place the prompt here. Once we have

this prompt here, we just send it and

boom, we have it ready. Let's play the

video.

Did you notice every other thing changed

in this video apart from me? As you can

see, it's not really playing properly

because I'm just playing it straight

from V3. So, you have to download the

video first and play it for it to show

properly. But yes, this gave me what I

want. It changed instantly from the

green background to what I want, a

background I want in the first frame,

which is what we want. So, we have our

first video and we keep doing that for

the next part. So, for the next part,

I'm going to use that same prompt. So,

I'm going to use the same prompt for

this. Write v3 prompt following the same

template. instru to change the

background the uploaded image at the

start of the video and the first frame

write new background and then what I

want um the character to do the action

now you can just use this for different

parts of the video use the same template

I'm going to leave all the prompts that

I've talked about or used in this video

in my telegram channel so if you have

not joined my telegram channel click the

link in the description of this video to

join my telegram channel there'll be a

link there for you to join once you join

it you all these prompts ready for you

to just copy and use. So don't bother

trying to copy it and all that. So now

we have this. We can just copy this. All

right, we copy the result here. We copy

this for the next video and then we

select the image. So we paste the prompt

and then add the image that we want to

use. This time around we're going to

select the full portrait of the person

which is this one here showing the full

body. And then we send the prompt and

boom, we have the video set. For this, I

wanted the person to stay in a park so I

can play it

inconsistent throughout this video in

different scenes.

I'm going to play it later in this video

so you can see how it looks like. For

this next video, you're going to enter

the new prompt and then upload the image

and send it. And then you just do that

for all your videos. That's how I was

able to achieve this video right here.

Now, once you have all your videos,

you're going to do this next part, which

is another important step you have to

take. Now I know most people just want

to like generate a 30 seconds video or 2

minutes video or 5 minutes video

straight. No, that is not the best thing

to do. Number one, strategically is not

the best because if you want to create

an engaging video, you need to change

scenes. You need to change shots. You

need to, you know, add things in the

video that will make sense. So you want

to do it beat by beat. And 10 seconds is

a sweet spot. And also, you know, you

don't want to spend money generating one

minute, two minutes video to find out

that most of the videos are crap. Like,

you don't like it. So, you have to

regenerate and spend dollars, hundreds

of dollars again. So, no, you don't want

to do that. You just want to do it bit

by bit and then put it all together,

which is what I'll be showing you in the

next part. So, download all your videos

and take it to capcot.com.

Now what's left for you is to upload all

your videos on Capcot or any video

editing tool and put them together to

form the long video which has that

consistent character. Now what you must

do is to ensure that the first frame

which has this person on the green

background is cut off. And if you notice

that first frame the person didn't see

anything. So you just ensure that you

trim it. Let me zoom in. For each of the

clip you trim it out. So it's going to

start from where the person is speaking.

Okay. So for all the clips just ensure

that you cut out the areas that has this

green background, right? The character

you used in generating the video. So

just do that. Once you're done removing

those beginning parts, you just arrange

the media clips.

Arrange them accordingly how it's

supposed to be, how it's going to flow.

Once you're done arranging your videos

based on what your idea or your story

is, just click on export or make any

other changes. Add transitions, add

effects and all that to make it lively,

engaging. Once you're done, just click

on export and you have something like

this. What would happen if I jump off

this balloon? Of course not. I won't be

that stupid to do that.

Did you notice every other thing changed

in this video apart from me? As you can

see, I have been consistent throughout

this video in different scenes. Don't

worry, Charles, my creator, will show

you how he created this long video with

just V3. Over to you, Charles.

[Music]

Wow, it's beautiful, right? And it's

very simple. Anyone can do it. So all

the prompts I use in this video, I've

left it in my telegram channel. The link

to join my telegram channel is in the

description of this video. So check it

to join my telegram channel. If you have

not subscribed to this channel,

subscribe. There are a lot of helpful

videos here. A lot of videos that will

help you. So check them out. And also

wait, like and share this video to

someone that you know is going to help.

And check out this video because I

created it for you. Check it.

[Music]

[Applause]