從真實影片案例學習 MiniMax H3 提示詞

DAZgen X · @dazgen48

Two linked H3 Turbo tests contrast Japanese prompting and dialogue with English prompting and Japanese dialogue written in romaji.

FoldingEnd A.I. · @EndFolding79421

Two unedited twelve-second H3 experiments use long story prose about an undead awakening and a welcoming restaurant.

FoldingEnd A.I. · @EndFolding79421

Two H3 story-prose experiments explore a distant marriage and a detective’s uneasy arrival in a small town.

伊藤貴將（いとぱん）｜AIエージェント×創作開発 · @takamasa045

A five-page local folktale script becomes an anime through Grok coordination, H3 video and ElevenLabs voices with dialect corrections.

wasabi_vaio · @WasabiVaio

A Krea and H3 pixel-art concept shows a fantasy character exploring an academy in a Metroidvania-style action-game video.

MrJ0SHiE (AI Trump) · @MrJ0SHie

A local H3 fan short adapts hand-drawn designs into ChatGPT character sheets, with Suno background music.

tapehead.lab · @tapehead_Lab

An MV workflow uses faceless H3 Max Turbo previs, converts it to depth and feeds Seedance to rebuild anime characters and backgrounds.

ZhaxRa · @zhaxra

A local H3 I2V test shares action failures at 480p and reports that RTX upscaling smeared the results.

赤染　史郎 · @akazomesirou

A four-model showdown test identifies H3 as the third clip and reports only partial success with requested slow motion.

huangserva · @servasyy_ai

A side-by-side RTX 4090 test compares three-view Ref2VA orbit guidance with a single-image first/last-frame loop.

むぎ@まいにち寧々さん · @EverydayNene

A Colab A100 40 GB H3 test reports 177.77, 67.07 and 24.82 seconds at 1.0, 0.5 and 0.2 megapixels.

Jambomax_a · @jambomax_a

An H3 short-film post presents a Mandalorian and Grogu homage crossed with a Fantastic Voyage-style microscopic adventure.

Jambomax_a · @jambomax_a

An H3 and ComfyUI post shares a locally made Sengoku action short about Sanada Yukimura and Tokugawa Ieyasu.

Kent Dhani · @kentdhani

A local RTX 5090 test replaces three Jurassic World raptors at once rather than swapping one isolated character.

buchi_ · @mix_buchi_

A frog-shaped robot is animated with H3 Ref2V using Singularity and an LLM-written prompt.

Neo · @NeoAIForecast

A Hermes agent uses ComfyUI to assemble ten H3 clips, Qwen images and a 100-second MiniMax Music score on one Spark.

StudioYebisu · @studio_yebisu

An author compares unmodified H3 on Colab against a modified local setup, presenting both outputs at half playback speed.

Glenn Williams · @GlennHasABeard

A locked floor-height shot hides the cat under a sofa while a ball rolls back by itself in a local 15-second clip.

烁皓 · @eternityspring

A lost-cat comic uses H3 ORB360 LoRA imagery inside an interactive browser presentation with movable viewpoints.

Satsuki AI Log · @SatsukiAILog

A background-and-person Reference Edit test improves placement but changes the face and skin texture, according to its author.

PAPADU PAPADU · @papadupapadu

The Filter combines Midjourney imagery, Hailuo H3 video, Eleven audio and manual Premiere editing in a perception-themed film.

Jerrod Lew · @jerrod_lew

One reference image becomes draft images in Magnific One, H3 Max video clips and an edited spec ad with ElevenLabs audio.

Alexey Fateev · @superalesha

A local 3:16 anthem uses 109 H3 video runs; failed reference face swaps lead the creator to masked characters.

FoldingEnd A.I. · @EndFolding79421

A complete prose prompt follows Lucy leaving a diner and sensing a watcher, using eight steps and realism LoRA strength 0.6.

posi_posi · @posi_posi8

A shared Colab notebook starts an H3 WebUI with one button; the author reports a five-second clip in under two minutes.

akatz · @akatz_ai

H3 Ref2VA transfers performances with a character V1 LoRA, prompting and masking in a side-by-side development reel.

Kamimoto(かみもと) · @sep_is_heim

An anime-character test combines start and end images with a Transformation LoRA and a prompt.

Ison Ward · @Isonwardster

Three product views feed a vertical H3 sunscreen ad; a second pass replaces a leaking droplet with water reflections and highlights.

fal · @fal

Fal demonstrates H3 Max extending footage beyond its original frame to change aspect ratio while retaining timing and audio.

パジョカ (Pajoca) · @Pajoca_

A Windows FreeVideo v0.4.0 report generates a four-second 768×1344 image-to-video clip in 210 seconds on a 4070 Ti Super.