This is an AI video generation comparison for
motion-transfer 
prompt:
Image of a woman with a large bird nest on her head in which a pigeon with two eggs is sitting and a driving video of an Asian girl making silly faces and singing to the music....
Log in to see full prompt.
Tested: December 22, 2025
Character orientation matches video mode selected. It is pretty good, except one moment where extra hand comes in :) Love how the nest with bird wobbles realistically on her head. That was the idea behind creating this crazy image :D
Tested: December 22, 2025
The animation is great, except the bird on her head doesn't really wobble, which could add even more realism. Also, Wan zooms in and crops this one each time.
Tested: December 22, 2025
Not that great compared to newer animate models. Distortions, loss of consistency (nails light or dark? is there a bracelet on her hand or not?)
Is the woman’s head decoration clearly a large nest made of straw?
Does the bird remain on her head while the woman moves?
Does the woman perform noticeable, rhythmic body movements to the music?
Are finger gestures clearly synchronized with body rhythm?
Do facial expressions change dynamically with the beat or emotion shifts?
Are lips visibly moving in sync with the imagined or played music?
Is hair motion consistent with her movements (no clipping or unnatural rigidity)?
Are nest materials reacting naturally to her motion (minor sway, not static)?
Are lighting and shadows consistent on both the woman and the nest/bird?
Is the camera framing stable and focused on her upper body and face?
Is there no distortion or warping in her face or hands during motion?
Check out the results from Kling AI (Kling 2.6 Motion Control) vs Wan (Online Platform) (Wan2.2 Animate) vs Pollo AI (Runway Act-Two) for similar or identical prompts side-by-side.
