Built this in Kling 3.0 using a single hero image as my anchor frame, then treated it almost like directing a mini film rather than just animating a photo.
My workflow:
• Started with a still image of my marionette DJ character
• Used Image-to-Video in Kling 3.0
• Kept it in 9:16, 1080p, 15 sec for music/social content
• Generated single shots first rather than jumping straight into multi-shot
• Focused prompts on micro movements — head nods, shoulder rolls, subtle hand movement — instead of big exaggerated motion
• Added camera language like slow push-in, shallow depth of field, 35mm lens, cinematic shadows to avoid the slideshow look
• After each generation I used Extract Frame to grab the strongest moment, then fed that back in as the next starting frame
• Repeated that process scene by scene to keep character consistency while gradually building more energy
• Exported clips without relying on generated audio so I could sync everything manually to my own track in editing
Biggest lesson: Kling really shines when you think like a director, not just a prompt writer. Small controlled motion + frame extraction + scene progression made the biggest difference.