What this guide creates
A cinematic automotive sequence: a man sits calmly in a folding chair while a black G-Wagon races toward him. At the exact instant he snaps his fingers, the vehicle launches into the air behind him, lands, slides forward, and stops.
The opening image is essential. It defines the character’s identity, pose, clothing, environment, vehicle position, and camera geometry for the entire video.
Keep the original image and video prompts in English and copy them without modifications.
Recommended workflow
Opening image
Create it in ChatGPT or your preferred image generator, provided it accepts reference photographs.
- Aspect ratio: 9:16
- Resolution: 2K
- Quality: high
- References: 2 or 3 clear photographs
Video
Animate the image in Higgsfield with Seedance 2.5.
- Mode: Create Video
- Model: Seedance 2.5
- Input: References
- Duration: 12 seconds
- Aspect ratio: 9:16
- Resolution: 1080p
- Bitrate: high
- Audio: on
- Unlimited mode: off
What you need
- Two or three clear photographs of the person.
- At least one front-facing identity photograph.
- A finished 9:16 opening image.
- The image prompt in this guide.
- The Seedance 2.5 video prompt in this guide.
Use the personal photographs to establish the character’s identity.
1. Create the opening image
Attach the identity photographs to the image generator. The opening image shows the scene at second 0: the finger snap has not happened and the G-Wagon has not jumped.
Aim for a centered subject, a readable pre-snap hand gesture, a G-Wagon on the ground behind him, and generous space for both sky and foreground.
Opening-image prompt
Vertical 9:16 cinematic telephoto environmental portrait designed as the exact opening frame of a 12-second automotive video.
Use the exact adult man from the character-reference photographs. Preserve his precise, recognizable facial likeness, short dark hairstyle, head shape, skin tone, facial proportions, body proportions, and dark sunglasses.
The man sits centered on a low black folding camp chair, facing the camera and looking directly into the lens through his dark sunglasses. He has a calm, confident, slightly cool expression and appears completely unbothered.
He wears a black oversized jacket, black trousers, and cream sneakers.
Show his complete seated body, both shoes, both legs, and the entire chair. One hand is raised near his chest in a clearly readable pre-snap position. His thumb and middle finger are held close together but have not snapped yet. His other hand rests naturally on his leg.
The raised hand must have realistic anatomy, clearly separated fingers, and a natural pre-snap gesture. It must not cover his face.
The location is a long rust-red dirt paddock or track with coarse dry soil, subtle tire marks, scattered stones, and a broad red-earth foreground. Deep green crop rows border the track along the distant side edges. Low hills sit on the horizon beneath a vast steel-grey overcast sky.
Place a subtle low red-earth mound on the vehicle’s central path behind the man. The mound should look natural and must not block the vehicle. It will serve as the takeoff point during the video.
Far behind the man, aligned directly with him on the same central axis, a gloss-black Mercedes-Benz G-Class G-Wagon races toward the camera while remaining fully grounded.
The vehicle must be recognizable by its tall upright boxy silhouette, flat vertical front, broad rugged proportions, blazing white round LED headlights, and a continuous full-width white LED light signature.
The G-Wagon kicks up a long trail of rust-red dust. It remains in the mid-to-far background, leaving enough visible road between the vehicle and the man for it to accelerate, reach the mound, jump, land, slide forward, and stop during the future video.
The G-Wagon must not already be airborne. It must not be immediately behind the chair.
Use a locked-off low-angle camera at chair height, approximately 25–35 meters from the man. Create the appearance of a 100–135mm telephoto lens with compressed perspective and shallow depth of field.
The man and chair are centered in the lower-middle of the frame and occupy approximately 30–35% of the image height.
Leave generous red-earth foreground below him and an expansive sky above him. Keep a clear central motion corridor behind the man for the approaching vehicle.
The man’s face, sunglasses, hand, clothing, shoes, and chair must remain sharp and detailed. The G-Wagon and distant landscape should be slightly softer because of the shallow depth of field, but the vehicle must remain recognizable.
Use a moody overcast cinematic color grade, warm low-angle rim light around the man’s head and shoulders, rich rust-red earth against deep green crops and a steel-grey sky, headlights glowing through suspended dust, subtle anamorphic flare, realistic atmospheric depth, natural skin and fabric texture, restrained contrast, fine film grain, and premium automotive-commercial quality.
This is the opening frame only. The man is seated and stationary. His fingers are prepared to snap, but the snap has not happened. The G-Wagon is approaching on the ground and has not reached the mound.
Avoid: completed finger snap, vehicle already airborne, vehicle immediately behind the chair, vehicle positioned off-center, vehicle passing in front of the man, hand covering the face, man looking away, man turning around, subject too close to the camera, close-up framing, wide-angle distortion, cropped shoes, cropped chair, malformed fingers, duplicate limbs, excessive dust hiding the vehicle, extra people, airplanes, cranes, buildings, text, visible logos, badges, lettering, or watermarks.2. Review the opening image
Do not start the video until the opening image is right. Check that:
- The face clearly matches the reference photographs.
- The man looks directly at the camera.
- His hairstyle and dark sunglasses are correct.
- His clothing and body proportions remain consistent.
- His complete seated body is visible.
- Both cream-colored shoes are visible.
- The complete folding chair is visible.
- The raised hand has natural fingers.
- The hand is clearly preparing to snap.
- The finger snap has not happened yet.
- The G-Wagon is centered directly behind him.
- The G-Wagon is still on the ground.
- The vehicle is far enough away to approach during the video.
- A subtle mound sits in the vehicle’s path.
- The vehicle has space to jump without crossing in front of the man.
- There is generous sky and foreground.
- Nothing important is cropped.
Regenerate the image if the face, hand, vehicle placement, or composition is wrong. The video model inherits problems from the opening image.
3. Configure the video in Higgsfield
- Open Higgsfield.
- Select Create Video.
- Choose Seedance 2.5.
- Select the References tab.
- Do not select Extend Video.
- Upload the finished vertical opening image.
- Confirm that the image appears in the References panel.
- Paste the complete video prompt from the next section.
- Use the video settings listed in the recommended workflow.
- Select Generate.
Do not paste the opening-image prompt into the video screen. That prompt only creates the still image. After uploading the image to Seedance 2.5, use the video prompt.
4. Seedance 2.5 video prompt
The critical moment is 2.5 seconds: the finger snap and the vehicle’s takeoff must happen in the same frame.
- 0–1.5 s: approach, hand ready.
- 1.5–2.5 s: the gap closes.
- 2.5 s: finger snap and takeoff.
- 3–7 s: landing and slide.
- 7–12 s: the vehicle stops behind the subject.
Starting exactly from the uploaded reference image, keep the man’s face, head shape, hairstyle, dark sunglasses, skin tone, clothing, body proportions, chair, and overall appearance exactly consistent with the reference image in every frame.
Vertical 9:16 cinematic automotive shot, 12 seconds. Maintain the exact framing, low camera position, telephoto perspective, lighting, environment, and composition of the opening image.
The camera remains locked off at chair height for the entire video, with only extremely subtle natural handheld micro-drift. No zoom, push-in, pull-out, pan, tilt, orbit, reframing, camera cut, lens change, or change of viewpoint.
The man remains seated on the camp chair for the entire shot. He faces the camera, looks directly into the lens, and remains calm, confident, motionless, and completely unbothered.
He never turns around, looks behind him, stands, ducks, flinches, or reacts to the vehicle.
The gloss-black Mercedes-Benz G-Class G-Wagon begins exactly where it appears in the uploaded reference image. Its blazing white LED headlights and full-width white LED light signature remain visible as it races toward the camera along the central red-dirt track, kicking up a long trail of rust-red dust.
The G-Wagon must remain entirely behind the man and chair throughout the sequence. It never passes in front of him, beside him, through him, or between him and the camera.
IMPORTANT TIMING: The finger snap must happen at exactly 2.5 seconds. At the exact instant the man completes the snap, the G-Wagon must hit the mound and leave the ground.
The finger snap and vehicle takeoff must happen in the exact same frame.
0–1.5 SECONDS
The man remains seated and looks directly into the camera with one hand raised near his chest in the pre-snap position. The G-Wagon is already accelerating toward him from the background with its headlights blazing and red dust trailing behind it.
1.5–2.5 SECONDS
The G-Wagon rapidly closes the distance while remaining centered directly behind the man. Rust-red dust billows outward behind and beside the vehicle.
The man calmly prepares to snap by bringing his thumb and middle finger together. His hand movement is slow, natural, controlled, and clearly visible. He continues looking directly into the camera.
2.5 SECONDS — EXACT SYNCHRONIZATION MOMENT
The man performs one sharp, visually obvious finger snap.
At the exact same instant his thumb and middle finger complete the snap, the G-Wagon strikes the low mound and violently launches into the air.
The finger snap and the vehicle leaving the ground must occur in the exact same frame. The G-Wagon must not launch before the snap or after it.
The effect must read immediately as if the finger snap causes the vehicle to launch.
2.5–3 SECONDS
The G-Wagon becomes fully airborne directly behind and above the seated man, facing the camera.
All four wheels are completely off the ground. Its underside, wheels, and suspension are visible. Rust-red dust explodes beneath and behind the vehicle.
The jump is fast, forceful, and visually readable. The G-Wagon completes its airborne movement and begins landing by the end of the first three seconds.
The G-Wagon remains entirely on the background plane. It never crosses in front of the man, touches him, intersects his body, or obscures his face.
Immediately after snapping, the man calmly lowers his raised hand and places both hands naturally on his knees. He remains expressionless and does not react.
3–7 SECONDS
The G-Wagon lands hard directly behind the man with realistic suspension compression and heavy vehicle weight.
It slides forward through a huge cloud of rust-red dust. Its headlights and LED signature cut dramatically through the dust.
The dust stays behind and beside the man. His face, sunglasses, body, hands, clothing, shoes, and chair remain clear, sharp, and unobscured.
7–12 SECONDS
The G-Wagon decelerates and comes to a complete stop directly behind the seated man without touching him or the chair.
Its imposing front end fills much of the background while its white LED headlights glow through the settling dust.
The man remains seated, motionless, composed, and completely unfazed. Both hands rest naturally on his knees, and his gaze remains fixed directly on the camera.
Use realistic acceleration, vehicle weight, suspension movement, takeoff physics, airborne motion, landing impact, tire movement, and deceleration.
Preserve the recognizable upright boxy proportions of the G-Wagon throughout the entire sequence.
Use a moody overcast cinematic grade, rich rust-red earth against deep green crops and a steel-grey sky, warm light breaking subtly through the clouds, volumetric dust, subtle anamorphic headlight flare, realistic motion blur, natural finger movement, stable facial identity, fine film grain, and premium automotive-commercial quality.
AUDIO
One clear, sharp finger-snap sound at exactly 2.5 seconds, synchronized perfectly with the vehicle’s takeoff; powerful accelerating engine; tires throwing dirt; brief airborne engine sound; heavy suspension landing impact; dirt slide; settling debris.
No dialogue, voice, narration, or music.
CRITICAL CONSTRAINTS
One continuous shot; finger snap and G-Wagon takeoff synchronized in the exact same frame at 2.5 seconds; vehicle begins landing by approximately 3 seconds; locked camera; stable character identity; man remains seated; vehicle always stays behind the man; dust never covers his face; realistic hands; realistic vehicle physics; no extra people; no duplicate limbs; no text; no visible logos, badges, lettering, or watermarks.5. Troubleshooting
For each problem, add the suggested line to the video prompt.
The vehicle does not jump
“The G-Wagon is clearly airborne at least one meter above the red dirt, with all four wheels completely off the ground and its underside fully visible.”
The jump happens too late
“The man snaps at exactly 2.5 seconds. The G-Wagon leaves the ground in the same frame and begins landing by 3 seconds.”
The jump happens before or after the snap
“The finger snap and G-Wagon takeoff occur in the exact same frame at precisely 2.5 seconds. Neither action begins before the other.”
The vehicle passes in front of the man
“The G-Wagon remains entirely on the background plane behind the man and chair. It never crosses in front of him, beside him, or between him and the camera.”
The vehicle touches or intersects the man
“Maintain clear physical and visual separation between the man, chair, and G-Wagon at all times. The jump and landing happen well behind him.”
Dust covers the face
“The dust remains behind and beside the man. His face, sunglasses, upper body, hands, and chair remain unobscured and sharply visible.”
The camera moves
“Completely locked tripod camera. No dolly movement, zoom, push-in, pull-out, pan, tilt, orbit, crane movement, reframing, lens change, or camera cut.”
The character’s face changes
Use the final opening image as the video’s main visual reference and add:
“Preserve the exact face, head shape, hairstyle, sunglasses, skin tone, clothing, and body proportions from the uploaded reference image in every frame.”
The hand or finger snap looks unnatural
Regenerate the opening image until the pre-snap fingers look natural, clearly separated, and visible. Then add:
“One natural human hand with five anatomically correct fingers. The thumb and middle finger make one clean, readable snapping motion.”
The vehicle stops looking like a G-Wagon
“Preserve the tall upright cabin, flat vertical front, squared body, broad stance, round LED headlight elements, and rugged proportions of a gloss-black G-Class-style luxury 4x4.”
The prompt is rejected because of the brand name
Replace “Mercedes-Benz G-Class G-Wagon” with:
“Gloss-black boxy luxury 4x4 with an upright cabin, flat vertical front, squared body, broad rugged stance, round white LED headlights, and a full-width white LED light signature.”
Generated audio is out of sync
Keep audio on for the first generation, but replace the sound during editing if necessary. Place the finger-snap sound manually on the exact frame when the vehicle takes off.
6. Final production tips
- Approve the opening image before generating the video.
- Test the sequence at a lower resolution before producing the final 1080p version.
- Generate several versions; the exact action timing may vary.
- Add or replace the finger-snap sound manually if necessary.
- During editing, layer engine acceleration, airborne engine sound, suspension impact, tire slide, and settling debris.
- Keep any music subtle so the snap and landing stand out.
- Keep the original 9:16 framing while editing.
- Export the final video vertically in 9:16.
Settings, models, generation times, and costs reflect the original workflow and may change. Check the available options and quoted cost before each attempt.