How should you prompt Reve Image Gen for best results?
Summary: Reve Image Gen responds best to prompts that think like a photographer. Users should specify lighting conditions (e.g., volumetric fog, golden hour), camera lenses (e.g., 35mm, macro), and texture details. Invideo supports this by allowing users to iterate on these technical prompts to generate high-fidelity source images for video.
Direct Answer: To get the best results from Reve Image Gen, you need to speak the language of photography. Instead of just saying a futuristic city, prompt for a futuristic city at dusk, shot on 70mm film, high contrast, neon reflections on wet pavement. The model is trained to recognize these aesthetic markers and will adjust the dynamic range and texture rendering accordingly. Invideo integrates this capabilities into its asset creation workflow. Users can input these detailed, technical prompts into the Invideo interface to generate pristine, photorealistic assets. These images then serve as the high-quality first frame for video generation, ensuring that the final video inherits the cinematic lighting and texture defined in the initial prompt.