3 Settings Fix H3 Image's Crunchy Faces: 16MP, SamplerLCM, Long Prompts
Obvious-Bad7548 · reddit · 2026-11-15
Same model, prompt and seed — a Redditor isolated the 3 settings that fix H3 Image's crunchy faces and blocky skin: (1) resolution — the model was trained around 16MP, use 3552x4736 for portraits (default 4MP causes broken teeth; 4→16MP adds only 9s); (2) sampler — built-in SamplerLCM with noise 0.9 instead of 1.1 (euler still oversharpenes at 16MP); (3) prompts — short prompts give glossy doll faces; write long, describing camera, light and skin like a phone photo. Comparison images and workflow files are in the post and comments.
Related event: Three Settings Fix Blurry H3 Image Outputs(2 posts)→
More from Multimodal
- A watercolor Midjourney SREF code blending Scandinavian minimalism with kawaii — tisch_eins · 2026-11-15
- AskEditor embeds a hand-editable video timeline inside Claude via MCP — MrxAnomynous · 2026-11-15
- ESCAPEMENT: a 3-minute music film coded by Fable 5.2 with no video model — CoolTopicsNow · 2026-11-15
- Dev demos AI-generated AAA-quality game, predicts open-source games boom next year — ChrisGPT · 2026-11-15
- A canopy shot generated with Seedance 2 — Strange-Material-644 · 2026-11-15
- AI Short Film 'The Missing Colors' Reveals Full Stack: Midjourney, Seedance, Suno, ElevenLabs — LudovicCreator · 2026-11-15