🍿 Creating "Movie-Like" Videos with Runway Gen-3
The Gen-3 AI model allows you to create highly realistic videos using text prompts. Users can control the camera, change the viewing angle, and place light sources — essentially performing all the functions traditionally handled by a director and cinematographer.
Blogger Tao Prompts took it a step further by creating a tutorial on how to craft cinematic scenes using Runway Gen-3. We've reviewed it and compiled the key recommendations for you.
How to Make Videos Cinematic?
1️⃣ To create a video in Gen-3, you need to upload an image from which the model will generate the video. At this stage, you should also choose the first/last frame option, which indicates whether the uploaded frame will start or end the video.
2️⃣ Color Correction. One drawback of Gen-3 animation is that the generated video often has overly bright colors. Therefore, the blogger suggests adding notes for color correction in the prompt. Tao Prompts noted that keywords can sometimes affect the video in unexpected ways: for instance, reducing color saturation might also alter facial animations. Hence, another recommendation is to create videos step by step and experiment.
Keywords:
→ desaturated colors
→ muted colors
→ low contrast
3️⃣ Cinematography. To control the camera, you can use the following keywords:
→ static shot: eliminates background hallucinations and is suitable for scenes with a speaking character.
→ drone shot flying: you can specify how high it should fly.
→ bird's eye view.
→ FPV fly-through: first-person view through a space or building.
→ the camera zoom in/out. Be sure to specify the object of focus, such as a character's eyes or a building (see next point).
→ close up shot on...: focus on a specific object or part of it.
→ full body shot.
➡️ You can find detailed instructions for creating videos in Gen-3 here.
➡️ Try Runway Gen-3 via this link.
More stories:
