Here's an example of 3 edits of a video with Omni:
1. original
2 maker her invisible, put gloves on her
3. while she's talking, two men come and take away the framed picture
4. change her outfit
Seedance 2.0 vs Gemini Omni, tested under the same conditions 👀
I was working on a new storyboard. After generating it with Seedance 2.0, I gave the exact same prompt + storyboard reference + character reference to Gemini Omni as well.
Result: Gemini Omni surprised me with style quality and got closer than I expected in prompt adherence. But Seedance still feels ahead for storyboard execution, motion energy, camera language and environmental interaction. Gemini looks good. Seedance feels directed.
You can see them in the video and the storyboard sheet is shown at the end.
Which one do you think is more useful for this kind of work?
GPT image 2 on ChatGPT
Prompt:
A hyper-detailed miniature diorama of [CITY OR PLACE] seamlessly built directly onto a rustic wooden table surface, as if the entire location is being handcrafted in real time. The scene features the most iconic landmarks, architecture, streets, transportation, cultural elements, and atmosphere of [CITY OR PLACE], surrounded by dense miniature activity with tiny people, vehicles, shops, markets, lights, and environmental storytelling unique to the location. Roads, landscapes, buildings, and structures emerge naturally from the wooden tabletop with no visible platform or base. In the foreground, a realistic human hand carefully uses precision tweezers, brushes, or miniature modeling tools to place and adjust tiny details, emphasizing the handcrafted diorama creation process. Warm cinematic lighting streams through a nearby window, creating soft golden highlights, atmospheric haze, floating dust particles, shallow depth of field, and realistic tilt-shift photography. Ultra intricate miniature craftsmanship, photorealistic textures, cozy workshop ambience, cinematic composition, immersive scale-model realism, ultra realistic 8k detail --ar 2:3 --style raw --v 6
effil tower
GPT Image 2 on @itsPolloAI
Prompt: Use my reference photo to preserve the exact facial features, hairstyle, bangs, facial proportions, skin tone, body posture, and overall silhouette with high identity consistency.
Cinematic portrait of a woman standing alone inside a dark exhibition room. She wears a vintage white shirt layered with a soft pink outer jacket and a black vintage-textured skirt. Minimal makeup. Front-profile pose with her body facing the camera, relaxed posture, one shoulder slightly lowered, hands gently adjusting her hair. Calm, mysterious expression with an intellectual and elegant aura.
She is positioned in the lower center of the frame, surrounded by darkness. A single dramatic diagonal spotlight cuts across the scene from the upper right to the lower left, illuminating only half of her body while the rest fades into shadow. Behind her is a colorful vintage Japanese mural wall filled with retro cartoon illustrations and traditional artistic details.
Foreground shadows naturally frame the composition like torn darkness, creating a moody museum atmosphere with lonely artistic energy and a contemplative cinematic mood. Realistic photography with subtle film grain, analog texture, muted blacks, and vibrant mural accents.
Camera angle: eye-level shot, medium-close portrait framing, subject captured at approximately a 70-degree side profile angle, slight distance from the subject for environmental storytelling, centered cinematic composition.
Camera settings: cinematic low-light photography, shallow depth of field, realistic shadow falloff, Kodak Portra film color grading.
Lighting: harsh focused spotlight from the upper right creating a strong diagonal beam and deep high-contrast shadows, minimal ambient light, rich blacks, soft warm highlights along the face and glasses edges.
Vibe: mysterious, artistic, museum-core aesthetic, intelligent quiet energy, melancholic cinematic atmosphere, indie film still, luxury editorial photography, solitary nighttime mood, understated elegance.
Style: ultra-realistic cinematic photography, analog grain, editorial fashion portrait, natural skin texture, subtle imperfections, realistic color science, masterpiece quality.
Negative prompt: anime, smiling, overexposed lighting, extra limbs, bad anatomy, overly colorful outfit, cyberpunk neon, blurry face, low quality, distorted hands, unrealistic skin, HDR overload, exaggerated makeup, busy composition, duplicate person.
🎬 Short Film Prompt Generator Custom GPT
Create storyboards, character sheets, cinematic shot breakdowns, and access over 50 different production templates — all designed to help you build a full 1-minute video in just 5–6 generations.
📚 Check out the tutorial articles to learn the full workflow and start creating faster.
Pour ceux utilisant la version gratuite de ChatGPT:
1-Dès le début, tapez le mot "menu".
2-Entrez votre image, sélectionnez le numéro de votre choix puis validez.
3-Recommencez.
Vous génèrerez une image à la fois mais vous gagnerez énormément en qualité.