Text to video generation for a tracking shot following a frisbee thrown in the woods to a disc golf basket. Run locally using ComfyUI and MiniMax H3.
I’m satisfied with the video generation. It didn’t take that long for a 12 second clip on my 16GB NVIDIA GeForce RTX 4070 Ti Super card. There are some obvious problems with the clip, particularly near the end, where the frisbee seems to shrink and go beyond the basket but then the next clip puts it in there. I asked for it to make an arc and come down and that was it’s interpretation. I could try again with a different approach or a more detailed storyboard for that shot.
The prompt:#
Realistic live-action cinematic look, sports action cinematography: practical film photography style, anamorphic lens, shallow depth of field, natural forest lighting, woodland volumetric mist, vibrant but organic color grading, powerful natural movement.
Scene overview: In a dense, ancient forest, a tanned white person with light brown hair performs a high-precision ultimate frisbee throw. The camera executes a high-speed “chase” trajectory, following the disc’s flight path through heavy foliage. This is a high-stakes display of athleticism and aerodynamics, capturing the fluid motion of the disc slicing through the wilderness.
Storyboard (each shot a separate scene, rapid cuts, all landing on the musical beats):
[0s-2s] Shot 1: Medium side angle: The protagonist winds up and releases the frisbee; muscle tension visible in the arm, the forest depth blurred in the background.
[2s-5s] Shot 2: Extreme close-up “chase cam”: The camera locks inches behind the frisbee as it enters the thickest part of the woods, whipping past trunks and leaves on both sides with high-speed motion blur.
[5s-8s] Shot 3: Wide tracking shot: The frisbee follows a massive arc, climbing toward the canopy on the right, with dappled sunlight breaking through the leaves.
[8s-10s] Shot 4: High-angle tracking: The frisbee dives sharply to the left, slicing through a gap in the branches, maintaining a perfect trajectory.
[10s-12s] Shot 5: Low-angle “impact” shot: The frisbee hits the chains of the metal basket with a satisfying jiggle, the camera following the final settle into the basket.
Camera: Dynamic tracking, rapid cuts between the thrower and the disc, slight frame jitter on the sharp turns, motion blur on the peripheral trees.
Audio: Deep “whoosh” of the frisbee, rustling leaves, birdsong, heavy breathing, rhythmic percussive hits on the turns, a metallic “clink” at the 11s mark, score swelling on the final landing.
No text, subtitles, logos or watermarks of any kind, no animation or cartoon rendering, no overly-CG look, keep the live-action texture.
