In many cases, I find that I will want to take a still image and convert it into a video. Nothing terribly complicated. Most of my shots are landscape and I generally just want to do a very simple push in or some type of dolly move left or right, but it doesn’t seem to matter how I word the prompt. I still get a very shaky handheld camera movement using KLING 3.0. Am I using the wrong model or what? This is truly driving me crazy.
Example prompt:
Slow dolly push forward through a gap in the foliage, revealing a wooden deer blind in the distance. Locked-off tripod camera, static mount, no handheld motion, no camera shake, no jitter. Smooth continuous forward motion at constant speed. Subtle parallax as near foreground leaves pass the frame edges faster than the background trees. Cinematic, natural light, one camera move only.
In the sample video below, it looks like the camera is being dragged across a pile of rocks as opposed to a nice smooth camera motion. This is really driving me nuts that I can’t get this to smooth out. You think that would be an easy point for AI to understand.