Another Experiment in AI Filmmaking
I've been exploring the evolving potential of AI-generated animation and felt that the anime aesthetic is becoming increasingly viable so I decided to give it a try. My process involved generating all images using Leonardo AI, then animating them through Kling 2.0 (which, impressively, seems to release a new version every month, each iteration better than the last).
Here are some of my thoughts:
1. Shooting Ratio: In traditional filmmaking, we often work with a 6:1 or even 10:1 shooting ratio—capturing more than we use. With AI tools, almost every generated frame ends up in the final cut, leaving no room for variation or spontaneity. To achieve a more cinematic ratio, I need a month or two of focused iterations and renderings.
2. Storyboard Integration: If AI tools could implement a "Storyboard-to-Preview" function—allowing us to see what we’re about to generate and adjust focal length, framing, and composition in advance—it would be a game changer.
3. Motion Blur: Current tools struggle with motion blur. Whenever it’s applied, image detail noticeably degrades.
4. Pattern & Symbol Generation: AI is still very limited when it comes to rendering patterns, graphic lines, symbols, or any kind of legible text. That remains a blind spot.
In conclusion, we are at the very beginning of this technology. This is the worst version we’ll ever see. From here, the tools will only get faster, better, and cheaper. Think of how we came from Photoshop v1.0 without layers. Right?
More than ever, I believe you need a strong personal vision, refined taste, artistic direction, and real filmmaking experience. Otherwise, you risk becoming just another voice in the sea of fan-made-content, which trend we're all too familiar with. Open AI’s Ghibli profile sensation, for example.
Entertainment has, and always will have, more value when it’s rooted in an original voice. A unique voice from an individual artist.
