FPV drone

- Known from
- Ambulance
- Resolution
- 1344×768
- Length
- 124 frames, 5.17s
- Steps
- 20, res_multistep / simple, turbo LoRA off
- Seed
- 7
- Ran on
- RTX 3060, 12GB (local)
What was wired in
<Picture 1>
MarcusCharacter reference plate
Camera clause
subject_definitions: <Subject 1> is the person in <Picture 1>, with a short tapered fade, a neatly trimmed beard, a white linen shirt with the sleeves rolled up and beige chinos. summary: [reference generation] The target video shows <Subject 1> standing still on an alpine lake at dawn, in a high aerial wide shot, as the camera performs a racing-drone dive. retention_analysis: <Subject 1> (appears in [Shot 1]): fully_preserved - his face, identity and clothing (a short tapered fade, a neatly trimmed beard, a white linen shirt with the sleeves rolled up and beige chinos) are held identical to <Picture 1> in every frame with zero drift. detailed_description: The target video is live-action and cinematic. [Shot 1] An aerial wide shot from high above an alpine lake at dawn looks down on <Subject 1>, a tiny figure far below, standing in place at the end of the jetty, breathing softly, hair and clothing stirring in the air, blinking and shifting his weight a little as people naturally do, with a wooden jetty, a red rowing boat, shoreline pines and snow-capped peaks behind him. The camera flies like a racing FPV drone: it dives steeply at fast speed, rolls and banks through the air, skims low and fast over the ground toward him, then pulls up and slows to a stop in a medium shot of <Subject 1>, his face sharp in frame. overall_soundscape: Still morning air, water lapping at the jetty and a bird calling across the lake non_diegetic_music: A single sustained low string note swells partway through the shot, then fades toward the end
Run it yourself
Download everything for this test (.zip, 1.3 MB). It holds the workflow, every input it loads and a README. Copy the inputs into ComfyUI/input, put the models below in place and drag the workflow onto ComfyUI.
| File | Put it in | Size |
|---|---|---|
| minimax_h3_ref2va_pruned_int8_convrot.safetensors | models/diffusion_models | 21.0 GB |
| qwen3vl_32b_minimax_h3_nvfp4_awq.safetensors | models/text_encoders | 15.7 GB |
| minimax_h3_video_vae_fp16.safetensors | models/vae | 5.2 GB |
| minimax_h3_audio_vae_fp32.safetensors | models/vae | 605.3 MB |
No custom nodes. Every node is part of ComfyUI 0.41.0.