Some friends came to ask me:
I WANT TO MAKE A MOVIE LIKE THIS. WHAT'S THE AI TOOL? HOW MUCH DOES IT COST
WHEN I'M DONE TALKING ABOUT FRIENDS, I'LL WRITE ABOUT HOW TO USE AI
Complete the entire process of a short film/film that was designed entirely by you
- The core process is three steps:
- Wensheng, → Vincent, 首 First-end frame video, →, Clip-in-Clip
Today, we're going to finish a short film in the martial arts style
First step: Ventura (role, scene, etc.)
Recommended tools:
- GPT-image-2 (OpenAI model)
- Seedream 5.0 (volcanic model)
- Nano Banana Pro (Google Model)
- Midjourney v7 (Midjourney model)
Here's the GPT-Image 2.0 demonstration
Unanimously: subject + environment + style + structure + drawing words
Number 1
The subject is in a seamless background. There are no objects, tables, surfaces, floors or environmental elements - only subjects. • Production of contact film for a highly sensitive martial arts character, using photo-level real film quality, warm amber colour and cold, dark and dark contrast palettes, with a shallow view. She is a young Chinese woman in her twenties, who behaves with a fine and strict manner, with sharp eyes and high cheekbones, and has a black hair bound up and scattered over her face. She was wearing a coat of black silk, which was folded, with dark embroidery in her sleeves and necks, with dark belts around her waist, and long sleeves designed for movement. A straight sword with black paint on his left waist. Her expression was very restrained — not apathy, but restraint. The contact tablet should contain a full-faced, wide-angle lens, including feet. A panoramic lens, including feet, facing the back, looking all over. There's a close-up from above the shoulder, on the front side. Ensure that the role is identical in all positions. The context of the game is neutral and extremely simple, with ground-level and simple scenery depth. The light is even, the contrast is low and the color is natural. Shadow up, natural effect. No partition lines, text labels or graphic superimpose
Number two
The subject is in a seamless background. There are no objects, tables, surfaces, floors or environmental elements - only subjects. • Production of contact film for a highly sensitive martial arts character, using photo-level real film quality, warm amber colour and cold, dark and dark contrast palettes, with a shallow view. He was a Chinese man in his mid-50s to mid-60s, despite his age, his size and strength, and his long, long white hair and a well-cut white beard. His face was covered in deep wrinkles — a man who had experienced violence and survived. He was dressed in a floating white grey-covered silk-skinned dress with silver and gray dots, with broad belts between his waist and wide sleeves to cover his withdrawal. A sword, a white silver blade hanging between his waist. His posture is completely static — the kind of static that never wastes action. The contact tablet should contain a full-faced, wide-angle lens, including feet. A panoramic lens, including feet, facing the back, looking all over. There's a close-up from above the shoulder, on the front side. Ensure that the role is identical in all positions. The context of the game is neutral and extremely simple, with ground-level and simple scenery depth. The light is even, the contrast is low and the color is natural. Shadow up, natural effect. There are no dividing lines, text labels or graphic superimposed。
Scene 1
Create a written picture of the high-altitude frozen lake on the eve of dawn in ancient Chinese mountains. The ice is wide and flat, cracking into a tiny fractal pattern, reflecting the first dim ray of light through the ridge — the cold blues and the light and warm roses woven on the horizon. The thin snow spreads on the ice, forming a pattern of wind blowing. The peaks around them are dark contours surrounded by the first light. The atmosphere is vast, exposed and ending. Images should be made from a panorama perspective, unimpressed and free of unorganized or unnecessary items. Space should feel the atmosphere suitable for the scene, but uninhabited. The light is even, the contrast is low and the color is natural. Shadow up, natural effect。
Scene 2
A documentary film depicts an ancient Chinese mountain temple that lies at the height of the mountains. The images show multi-layered monasteries and night roofs — crooked varnish roofs with snow, snow-covered stone courtyards, red paper lanterns shining under the blue and black sky, fine sculpted pillars and aromatic stove. The candlelight leaked from the door inside. Snowflakes fall light into the air. The distant background is the peaks covered by fog. The atmosphere is cold, ceremonial and dangerous. Images should be made from a panorama perspective, unimpressed and free of unorganized or unnecessary items. Space should feel the atmosphere suitable for the scene, but uninhabited. The light is even, the contrast is low and the color is natural. Shadow up, natural effect。
Scene 3
Creates a written picture of the ancient Chinese bamboo forest that is thick on winter nights. The tall bamboo has been pulled from the snowy ground, and the light green and grey surface reflects the cold blue moonlight spilled from above. The beam penetrates the tubing, forming a strong vertical light. Light snow falls from the canopy. The ground is covered with thin snow and rotting bamboo leaves. The air is static, silent and slightly dizzy — a closed space that can become a trap. Images should be made from a panorama perspective, unimpressed and free of unorganized or unnecessary items. Space should feel the atmosphere suitable for the scene, but uninhabited. The light is even, the contrast is low and the color is natural. Shadow up, natural effect。
Mirror
The following components are presented in a mirror:
Story Outline:
A short video of the Chinese martial arts film, set up at the Temple of the Snow Mountain
A young swordman in black fought with a white-haired swordman on the temple roof, bamboowood and frozen lake
The blade blinked in the snow and in the candlelight, and she ended up letting him go, and left with a broken jade
An elegant sword dance, slow motion, dramatic lights, dynamic photography, epic ancient Chinese atmosphere。
Story:
First snow
When she fell on the roof of the snow-covered roof, her footsteps were stepped forward and the black robe was turned back。
He was there long ago, with white hair scattered in the winds of the mountains. Neither spoke. Snowflake fell between them
The amber light of the lower lantern. The hands of both men slip slowly towards the handle of the sword, like slow motion
Those were two decisions that were taken at the same moment. Then it was still. Then, draw the sword。
2 Steel blades and candlelight
The sword collided on the temple roof. His sword is as thick as a mountain, and every wall is like an unshakeable wall
Each strike is like a controlled avalanche. She's faster. She didn't break it, but she did
Sliding, silk sleeves draw arc in candlelight. One kick kicked the house to the bottom of the dark。
Where they landed, the snow blew up. The fighting proceeded along the ridge and ended up in the lower bamboo forest。
3 Bamboo Forest
The moonlight passes through the canopy and breaks into a pole of light. They travel between shadows and light
Blades only shine when the moonlights. He didn't even look at it
It's over. She jumped on another bamboo, she was down from above and he was a centimeter to the shoulder
Catch her sword. Their faces are close at this point, and they breathe in the cold into the fog. Evenly。
Both know. And the next moment they split apart, and the snow on the forest land was shaken up and turned round again。
4 Open ice, dawn light cold and pale, lights out the frozen lake. They are standing alone on the ice
There was no shelter, no walls to borrow. They slowed down, not because they were tired
It's because of precision. Every fight is calculated. The ice is broken by one step。
She pushed him back with a series of tactics that he could not respond to. His sword is out of the way。
Her sword pointed to his throat. He's still。
5 Letters
She looked at him for a long time. Blade didn't vibrate, but she didn't stab. She used the empty hand
And he went into his garments, and took out the piece which was broken, and it split in half, and the crack was clean。
She put her finger together and hold it. She lowers the sword. He didn't move. She turned around and went through the ice
The robe drags through the wind-drived snow without turning back. And the sky rose after the peaks
The lake gradually swallowed her。
Step 2: Tusheng Video
Recommended tools:
- Seedance 2.0 (volcanic model)
- Little sparrow (volcanic model)
- The dream
- Grok (xAI model)
I've set up a $280 Seedance 2.0 package for this lesson. Ten million Token
Click Model - Select Visual
Seedance2.0 models have two types: fast and standard
if you want to practice, you can use fast first. it'll be cheaper
Click to open -- open 2.0
- Note: The account balance here needs to be greater than or equal to $200, or a package has been purchased, otherwise the Seedance 2.0 series model cannot be developed
Set up sequentially: draw scale, resolution, video time, number of times generated
The scale is based on your needs
The resolution is the clarity of the picture. The higher the resolution, the clearer
There's two kinds of video time
- Number of seconds: the length of time the force has to generate, up to 15 seconds
- SMART TIME: AI AUTOMATICALLY DETERMINES THE DURATION BASED ON YOUR TEXT
Quantity generated: is running several versions of the same hint, commonly known as the "tick card"
we'll start with one, or your token may be in trouble
If you want to generate a 4K video, you need to open the Al MediaKit
IF YOU NEED TO GENERATE WATER-FREE VIDEO, REFER TO API CALL DOCUMENT (LINK BELOW)
https://www.volcengine.com/docs/82379/2291680?lang=zh
I'll use the standard version here
- A 15-SECOND 720-P VIDEO
- You can see the estimate for the lower right is $14
Images generated before uploading and spectroscopy
- If you want to tell the whole story in 15, set the parameters and click on it
- If you want to be more nuanced, you can write only first act or second act at the end of the hint
There are two possible scenarios
1. Direct start of operation through models
Red (suspected faces detected)
Solutions
Make a good figure in front and turn it into a hand-drawing style
Re-upload and run
I've produced 3 15 seconds of video
About 10/1 of the package, about $30, it's expensive
Here you can estimate the cost of Token
https://console.volcengine.com/ark/region:ark+cn-beijing/tokenCalculator
Step three: Clip Scrolling
It depends on what you want
If you want to make a complete short film, then you can import a good piece into the editing software, add a spin field and a mix。
What if someone says they won't? Use descript AI Clip Tool
Final Thoughts
Because it's the video media
- Traveling, searching for a hotel, a view, a car with equipment, inviting actors, waiting for sunrise, waiting for dark, waiting for actors, editing later..
And now all I need is..
- A computer or a mobile phone + 200 bucks budget + a little bit of creativity can make a professional video
- Like Mx-Shell