This guide opens the door to video generation with the Minimax neural network. You will learn how to turn both Text into Video and Images into Video, and pick up the secrets of writing effective prompts.
Part 1: Prompt basics for video generation
Generating video from images (Images -> Video)
When you build a video from an image, the first frame sets the foundation for everything that follows. That is why prompts for Images -> Video can be simpler than the ones for Text -> Video.
Basic prompt structure:
[Main subject in the first frame + Motion/Change]
Where:
- Main subject is what the neural network sees in your image
- Motion/Change is a description of what should happen to the subject
Example:
🖼 + [In the scene, a dog with glowing blue eyes stares intently at a wooden stick lying in front of him. When the dog's eyes begin to glow brighter, the stick slowly rises into the air and hovers in front of him. The atmosphere is magical, with a slight bluish glow around the objects]
Translation:
In the scene, a dog with burning blue eyes stares intently at a wooden stick lying in front of it. As the dog's eyes start to glow brighter, the stick slowly lifts into the air and hangs in front of it. The atmosphere is magical, with a faint bluish glow around the objects
Generating video from text (Text -> Video)
Unlike Images -> Video, generating Text -> Video means you have to describe everything that should appear in the frame, because the neural network has no starting image to work from.
Basic prompt structure:
[Main subject + Scene + Motion]
Where:
- Main subject is a person, an animal, an object or something imaginary
- Scene is the setting (a library, a park, a fantasy world)
- Motion is what the main subject does
Example:
[A puppy is running in the park]
_A puppy is running in the park_
[A woman with an umbrella walks down a rainy street]
_A woman with an umbrella walks down a rainy street_
[A small stream quietly flows in the valley]
_A small stream quietly flows in the valley_
Part 2: Advanced prompting techniques
To create a genuinely striking video from an image, add more detail:
Detailed prompt structure for Images -> Video
[Main subject + Motion/Change + Camera movement + Aesthetic atmosphere]
Example:
🖼+ [A cat runs rapidly through a corridor, with time distortion effects emanating from its body. As it accelerates, the walls and objects around it begin to warp and stretch, creating a tunnel-like effect. The camera follows alongside the cat, capturing the rippling waves of distorted time that bend light and space around its moving form. The scene has a sci-fi atmosphere with cool blue and white energy pulses]
Translation:
A cat runs rapidly down a corridor, and a time distortion effect radiates from its body. As it speeds up, the walls and objects around it start to deform and stretch, creating a tunnel effect. The camera follows the cat, capturing the rippling waves of distorted time that bend light and space around its moving figure. The scene carries a science-fiction atmosphere with cool blue and white energy pulses
**Tip:**
- The more precise the wording, the more precise the result.
- The richer the wording, the better the overall effect.
Part 3: Extra settings
Mastering camera movement
Camera movement is the key tool for creating emotional impact on the viewer. Neural networks can imitate a wide range of cinematic devices.
Basic camera moves
- The lens moves slowly and smoothly toward the subject, gradually enlarging it in the frame (smooth dolly-in)
- The camera drifts away from the subject unhurriedly, revealing more of the surroundings (slow dolly-out)
- The optics move purposefully up or down along a vertical line, holding a straight trajectory (vertical camera move up/down)
- The lens traces a smooth arc, creating a circular sense of movement in the frame (circular camera movement)
- The camera glides elegantly and without pause through the space of the scene (smooth camera flythrough of the scene)
- The lens rushes suddenly toward the subject, creating a dramatic accent (sharp zoom on the subject)
- The camera follows a set route step by step, as if moving along invisible rails (camera movement along a path)
- The lens locks onto a single point and does not change position for the whole scene (static camera)
- The camera shakes and wobbles slightly, creating a handheld effect (handheld shot with shake)
Complex and dynamic moves
- The lens keeps a moving character in focus at all times, repeating their movements (camera follows the character)
- The camera descends smoothly from a height, gradually approaching the ground (descending camera flight)
- The lens starts at ground level and rises gradually, opening a new angle (camera rising from the ground upward)
- The camera makes a full circular pass around a central object or character (camera rotation around the object/character)
- The lens breaks through physical obstacles, creating the effect of passing through solid surfaces (camera moves through objects or walls)
- The camera moves freely through space, imitating the flight of a drone (camera movement as if flying like a drone)
- The lens snaps abruptly and suddenly onto a particular object (fast camera jerk toward the object)
- The camera sways gently, imitating movement on the surface of water (floating camera movement, as if on water)
- The lens spirals smoothly around a point of attraction, creating a helix effect (camera circles in a spiral around a central point)
Transitions and unconventional moves
- The camera passes expressively through an architectural element, changing the scene (camera crosses a doorway/window)
- The lens glides softly from one character to another, joining them in a single scene (smooth camera move between characters)
- The camera tilts gradually away from the horizontal, creating the illusion that the whole scene is rotating (camera changes angle, creating a scene-rotation effect)
- The lens changes its depth of field right in the middle of a move (camera focus shift during movement)
- The camera moves at a noticeably reduced speed, stretching the moment in time (slowed camera movement)
- The lens makes uneven, nervous movements that heighten the sense of unease (jerky camera movement for an anxious effect)
- The camera weaves skillfully between people, keeping the main subject at the center of attention (camera moves through a crowd, focus on the main subject)
- The lens hangs in the air for a few seconds, then drops rapidly downward (camera "hovers" in the air, then falls sharply)
- The camera moves at an angle, creating a dynamic diagonal motion in the frame (sliding diagonal camera movement)
Experimental and stylized
- The lens makes a full turn around its own axis, creating a circular panoramic effect **(camera rotates 360 degrees on its axis)**
- The camera moves in ragged, unpredictable jumps **(jumping camera movement)**
- The lens dives below the surface of the water and then comes back out into the air **(camera "dives" underwater and comes back out)**
- The camera floats smoothly through space, as if gravity were absent **(camera moves as if in zero gravity)**
- The lens takes on the character's gaze, making the viewer a direct participant in events **(first person)**
- The camera travels through time periods, accelerating the events taking place **(camera moves, showing events speed up)**
- The lens jumps instantly between different locations with no smooth transition **(abrupt camera cut from one space to another)**
- The camera rotates and changes focal length at the same time, creating a complex visual effect **(camera movement with rotation and focus change)**
- The lens creates a splitting effect on the image followed by a move toward the subject (camera "disperses", imitating a drone effect)
**Tip:
**Keep camera movements to a sensible 5-6 seconds. Avoid overly complex shots, which can cause generation failures
Examples:
- Smooth dolly-in: "The camera starts on a wide shot of an autumn park, moves smoothly toward a bench where an elderly couple is sitting, and stops on their interlaced hands, showing the wrinkled skin and wedding rings in close-up."
- Slow dolly-out: "The camera first shows a close-up of a single red rose on a fresh grave, then gradually pulls back, revealing dozens of people in black standing in the rain, and finally shows the whole cemetery from above, where a lone figure remains by the grave."
- Circular camera movement: "The camera circles smoothly around a couple waltzing in the middle of a ballroom, first showing their focused faces, then the elegant movement of their bodies, and finally returning to their smiling faces as the music reaches its climax."
- Camera follows the character: "The camera follows a running child relentlessly through a busy market, rounding the stalls, slipping between shoppers, ducking under hanging fabrics, holding focus the whole time on the joyful expression on his face as he finally finds his mother."
- Camera rotation around the object: "The camera makes a full circle around a stone statue of an angel in an abandoned garden, starting with the sorrowful front of the face, passing through different angles that reveal the damage and moss on the wings, and ending with a view of the setting sun through a crack in the statue's back."
- Camera moves through objects: "The camera starts outside an old mansion, then passes smoothly through a stained-glass window, showing how sunbeams refract through the colored glass, and continues through the rooms, telling the history of the house through antique furniture and photographs on the walls."
- Camera crosses a doorway: "The camera moves slowly toward the half-open door of a shabby detective's office, lingers for a moment at the gap, showing a fragment of what is happening inside, then passes decisively through the doorway, opening a panorama of the room with the detective smoking at his desk and a mysterious woman in red standing by the window."
- Camera focus shift during movement: "The camera glides through a noisy coffee shop, focusing at first on a barista preparing an elaborate drink, then the focus blurs and moves to a girl waiting for her order, and then shifts again, showing clearly a stranger watching her intently from the far corner of the room."
- Camera moves through a crowd: "The camera starts at the entrance of a packed train station, then slips between hurrying passengers, rounds standing groups of people, passes through the noisy crowd, and finally finds and locks onto a lone man standing motionless amid the general bustle and looking at his watch."
- Camera rotates 360 degrees: "The camera starts on a close-up of the frightened face of the main character, then makes a full rotation around its own axis, showing first one part of the chaotic battle swirling around him, continuing the rotation through every element of the fight, and completing the full circle by returning to the hero's face, where fear has given way to resolve."
- Camera "dives" underwater: "The camera follows a photograph falling into the ocean, first showing it gliding through the air, then passing through the surface of the water along with it, showing the shift in colors and the distortion of light underwater, and continuing the descent after the slowly sinking picture until it disappears into the dark depths."
- Camera movement with rotation and focus change: "The camera starts on a wide shot of a dancing ballerina, then spirals around her, gradually closing in, shifting focus from her graceful movements to the strain on her face, keeps spinning faster and faster, blurring the background and creating a dizzying effect, until all attention is concentrated on a drop of sweat running down her temple."
Precise aesthetic control
Beyond precise camera movement, neural networks can accurately recognize descriptions of aesthetics and atmosphere in a prompt and generate beautiful visuals and dynamic designs to match.
Prompt examples with different aesthetic atmospheres:
- Somber atmosphere: "The camera films a woman sitting in a cafe, the woman looks up thoughtfully, the camera moves slowly to capture the street outside the window, the image has a gray tone, the colors are undersaturated, and the atmosphere is somber."
- Warm atmosphere: "The camera films a woman sitting in a cafe, the woman looks at the window, the camera moves slowly to capture the street outside the window, the image shows warm tones, rich colors and a relaxing atmosphere."
- Science-fiction atmosphere: "A man in a coat walks down the street of a modern city, the camera follows the man's movement from behind, with a glossy science-fiction atmosphere and colorful tones."
Conclusion
AI video generation opens up incredible creative possibilities without special equipment or extensive skills. The key to success is well-structured prompts and an understanding of visual storytelling. Experiment with different combinations, write down the prompts that work and develop your own style. Top up your balance and go test the MiniMax video generator in GPTunneL! Video generation technology keeps advancing fast — learn the basics today and you will be ready for the new possibilities tomorrow! And if prompting gives you trouble, use our assistant for writing the right requests for MiniMax. Just describe what you want and it will write the prompt for you!