Lock the first frame with Magnific One and a prompt like this:
A crowded overlook at the Monument Valley visitor center, Arizona-Utah border, in the late afternoon, with the iconic red sandstone buttes rising out of the flat desert valley so the viewer instantly recognizes the place. In the foreground, a handful of adult tourists lean on the low stone wall and wooden railing of the overlook: a man in a faded red baseball cap seen from behind holding up his phone, a woman in a wide straw sun hat and a sleeveless denim shirt in profile, a backpack strap and a shoulder cut by the edges. Across the valley stands the West Mitten Butte, a massive vertical red rock tower with a sheer flat summit hundreds of meters above the desert floor, with the East Mitten and Merrick Butte further back. On the flat top of the West Mitten, near the edge, a tiny group of adults is barely visible around a small barbecue grill, with a thin plume of grey smoke rising from that point into the pale blue sky. The people and the smoke are very small and distant, readable only as a few colored dots and a soft grey thread.
COMPOSITION:
Shot type: wide, distant
Angle: eye level from behind the railing crowd, looking across the valley, horizon tilted a few degrees
Placement: tourists filling the lower third and the left edge, the West Mitten in the upper center-right, the smoking summit in the upper right third, the other buttes fading into warm haze on both sides
Look-room: right
Foreground: mild occlusion, a blurred tourist shoulder and a raised phone at the bottom left edge
Crop: handheld, slightly imperfect.
CAMERA PACK:
Body: ordinary mid-range Android phone, held vertically at eye level over the heads of the crowd
Lens / zoom: 3x digital zoom from the main lens, the glass slightly smudged
Aperture or phone mode: normal photo mode, auto exposure set for the bright sky
Focus: the far butte reasonably sharp, the nearest tourists a little soft
Film or sensor: small-sensor phone texture, heavy processing from the digital zoom, JPEG compression
LIGHT:
Source: low late-afternoon desert sun from the left
Quality: warm, raking light on the red rock faces, long shadows across the valley floor, dusty haze
Color temperature: warm golden daylight, about 4500K
Shadows follow that source.
PHOTOGRAPHIC CHARACTER:
Fabric weave, seams and creases on the cap, hat and denim.
Grain or sensor texture that matches the camera pack.
Candid stance, catalog-unaware.
Highlights a little hot, the sky slightly clipped, the shaded rock faces blocked up and noisy.
Smeary watercolor detail from the digital zoom on the butte top, the tiny group and the smoke.
Crunchy phone sharpening halos along the butte edges and the railing, blocky compression in the haze.
Slight overall softness from the raised, shaky grip.
Plain, slightly over-saturated color straight out of the phone, an ordinary snapshot by a friend.
Use character references if needed (like I did) and animate it using seedance 2.5 and a prompt like this:
[STYLE + CAMERA + ATMOSPHERE]
Real amateur phone video, one continuous 30-second take with no cuts, filmed by a tourist at the crowded Monument Valley visitor center overlook. Handheld vertical phone, shaky grip, restless digital zoom that keeps pushing in and pulling back, smeary detail, compression and autofocus hunting. Warm late-afternoon desert light, dusty haze.
[IMAGE REFERENCES]
[ref_3] is the first frame: the video starts exactly on it. [ref_2] provides the redneck man's face, build, hair, beard and outfit. [ref_1] provides the woman's face, build and long wavy hair.
[CAST — LOCKED]
On the summit there are ONLY three living beings the whole video: the man, the woman and one black cat. Nobody else is on the butte: no friends, no family, no other people. The small figures seen from far away in [ref_3] are just these two. Everyone else is a tourist at the overlook, behind the phone.
[TIMELINE]
0-1s: Exactly [ref_3]. Tourists murmur in English, German and Japanese and point at the smoke on top of the West Mitten, surprised.
1-4s: [Crash digital zoom to medium] The phone zooms hard across the valley and settles on the flat summit of the butte, grainy and soft. Clearly visible: a big charcoal grill with one huge thick tomahawk steak sizzling and smoking, a big black studio monitor speaker on the rock, a dusty RV parked behind, an inflatable kiddie pool at the cliff edge with the woman sitting in it holding an empty glass bottle, and a red backpack lying on the rock. Only the man, the woman and the cat are up there.
4-8s: [Second push-in, tight] The man flips the giant steak at the grill. The woman lifts the empty glass bottle to her lips and finds it empty. She frowns, turns the empty bottle upside down and shakes it at him, with an annoyed shrug.
8-10s: [Zoom out to medium] The phone pulls back to reframe the whole camp. The man turns around toward her.
10-15s: [Pan to the RV, slow push-in] A black cat trots out from under the RV with a folded dollar bill in its mouth, walks up to the man, sits at his feet and drops the bill. The man picks up the bill, looks at it, nods at the cat.
15-19s: [Ease out to medium] He tucks the bill in his pocket, grabs the red backpack from the rock, slips it on and tightens the straps.
19-22s: [Quick half zoom out, chasing him] He runs straight to the edge of the butte and leaps off into the void. The tourists gasp and scream in shock.
22-28s: [Long free fall, jerky zoom out in steps] No parachute. He drops fast down the sheer red rock face, a tiny figure shrinking against the cliff, the phone fumbling back out step by step while the tourists scream louder and louder.
28-30s: [Wide] At the very last moment, just above the desert floor, the parachute bursts open, the word "TECHHALLA" readable on the canopy from afar. The tourists explode in relieved cheers, laughter and applause, phones raised.
[AUDIO]
Tourist chatter in English, German and Japanese, gasps, rising screams during the fall and relieved cheers close to the phone, desert wind. Heavy southern rock plays from the studio monitor the whole time, distant and tinny, louder while zoomed in. The steak sizzling. A small meow. No subtitles.
[STYLE & QUALITY BOOSTERS]
Photorealistic phone footage, real digital zoom artifacts, coherent physics, only the same two people and one cat on the summit, no other people up there, one empty glass bottle and one red backpack throughout, stable identities, natural motion blur, parachute fabric snapping open realistically.
Lock the initial frame using One and then animate it with SD 2.5. here's the prompt
[STYLE + CAMERA]
Real raw phone video filmed by a neighbor from a porch across the road, one continuous 30-second take, no cuts. Fern top left and porch railing bottom left stay in frame. Handheld wobble, clumsy pinch-zooms that go soft and blocky. Golden-hour sun, compressed phone quality.
[IMAGE REFERENCE]
image1 is the exact first frame and identity lock: the old cream-and-brown trailer in side profile on a steep grassy slope, tongue uphill on the left, rear downhill on the right toward the lake, "LAKE LIFE" sign, a piece of firewood wedged under the downhill side of the single tire. Three heavy-set Alabama men in lawn chairs on the roof with a blue cooler.
[TIMELINE]
0-4s: The men clink cans and drink. Neighbor drawls: "Y'all are gonna die up there." Man in the ALABAMA tank top: "Hold my beer!"
4-12s: Clumsy zoom to the tire. A black house cat walks out from under the trailer, bats the firewood, drags it out like a toy. The trailer creaks; the cat flattens its ears and bolts uphill out of frame, never seen again.
12-20s: Zoom out. The trailer starts rolling backward down the slope, speeding up and bouncing. The men grab their chairs yelling "WHOA WHOA WHOA!" Neighbor: "Oh no, OH NO!"
20-23s: Just before the water the men look at each other, shrug, sit back and raise their cans as the trailer plunges rear-first into the lake with a big splash.
23-30s: Zoom in. The trailer sinks slowly. The men stay seated, calm, open fresh cans and toast as the water rises to their waists. The bearded man pulls a phone from his overalls and the three pose for a selfie, cans up, the sinking trailer behind them. Ends on three heads above the water, cans and phone held high.
[AUDIO]
Phone mic only: crickets, cans, laughter, the lines above, cat paw on wood, metal creak, rolling and scraping, splash, bubbling water, cans cracking, a camera shutter click, neighbor laughing hard. No music.
[QUALITY]
Photoreal raw phone footage, identity locked to image1, realistic trailer weight, rolling and water physics, realistic cat. Nobody gets hurt. No cinematic look, no slow motion, no text, no cuts.
Now we just give Seedance 2.5 both references (the video and the first frame) and ask it to use the 3D animation as a reference for the notes our main character is playing.
Add extra details like moving waves... whatever you can think of! Once again, the prompt is in the Space.
🤖 Made with AI
If you make music with Suno, you can export the MIDI and generate the 3D video just like I did using the HTML I shared.
Then, we extract the first frame and transform it with Magnific One! It’s my go-to model for pretty much everything right now.
Prompts in the space.





