Physics report card: what each video model gets WRONG, so you can compose around it

by @rileymillercreative in c/cinematic-shorts · · Mood Piece

Spent a month cataloguing the physics failures per model so I can pick the right tool and dodge its weakness. The big ones:

Hair/cloth: Everyone struggles. Kling best, but long hair in motion still occasionally phases through a shoulder. Compose with tied-up hair or minimal cloth movement for safety.

Water: Seedance best (that film look extends to beautiful water), Runway worst (water often moves like gelatin, surface tension wrong). Hailuo's water is okay if it's not the subject.

Fire/smoke: Kling and Hailuo good, Runway's smoke can be weirdly solid/clumpy. Fire is generally safer than smoke across all of them.

Weight/gravity in action: Nobody's great. A jump, a fall, a thrown object — the arc is usually subtly wrong (too floaty, wrong acceleration). AVOID dynamic action if you want believability; this sub is about the HELD moment anyway, which conveniently dodges the whole problem.

Reflections (water/mirror/glass): All weak, same as the still-image models. The reflection is vibes. Keep reflective surfaces non-hero or expect to composite.

The meta-lesson: compose TO the model's strengths. A locked, minimal-motion, atmospheric mood piece dodges 80% of these. The failures all live in dynamic action and complex physics, which the single-shot mood piece naturally avoids.

723 points · 6 comments

Comments

  • @liamwalker_films5 · · 71 points

    water: Seedance best, Runway gelatin — can confirm, Runway water has this horrible viscous quality like everything's in syrup. Seedance water actually has surface tension and caustics. if water is the hero, Seedance, no question.

  • @ethanmartinez_art5569 · · 99 points

    'This sub is about the held moment, which conveniently dodges the whole physics problem' — honestly this is WHY cinematic single-shots are the sweet spot for AI video right now. We accidentally picked the one format that plays to the tech's strengths. Minimal motion = maximal believability.

  • @leilasinghcreative · · 88 points

    It's not even accidental anymore, it's strategic. The montage/action crowd is fighting the physics engine constantly. We just... don't throw anything or make anyone jump, and the uncanny valley closes.

  • @hanakimcreative · · 53 points

    the hair-phasing-through-shoulder thing is SO immersion-breaking and once you've seen it you catch it in everyone's shots. tied-up hair is the coward's fix and I use it constantly. or keep the head still so the hair doesn't swing.

  • @kaismith · · 61 points

    'tied-up hair is the coward's fix' joining 'no text in frame' and 'no reflective surfaces' in the great book of AI-media cowardice. we're all just composing around the holes.