I compare Sora and Runway ML AI video generators in this post. You can readily see the difference between the 480p resolution of the Sora video captured frame in the third image and the 720p of Runway in the featured image. (The second image was my original featured image in a previous post). Sora’s is significantly grainier. Neither images are Photoshopped. I noticed that my Sora videos posted on LinkedIn are very grainy too. I sacrificed quality for quantity with my Sora settings to generate more videos. However, in the future I’ll be selecting 720p resolution for all my videos.
As far as content, the Sora video has more drama, whereas Runway looks smoother and more polished. Sora, in general tends to be more expressive than Runway.
Both video generators flatter the original image input. In some cases, they create completely new designs. Note how the hood reflection of a pole was turned into a design feature in the two videos. Often the highlights, reflections and surfaces on the car are improved and made more appealing. I regularly take frames from the videos and re-render them in Adobe Photoshop. However, in this post the featured image is a frame straight from the Runway video.
I’ve included a tutorial on how to storyboard with OpenAI Sora. In it you can see how to insert specific text prompts at various lengths of the video to provide direction at specific frames throughout.
