Hey, pal! I was out of town until today and just missed the 3 day mark on your first ones. I was able to grab your second one. The first thing is your image is a non-standard vertical size, not even true 16:9 (or here not even 9:16 since it is vertical) and your video size is 640x480 which is not 16:9 or vertical either. Your image should always track with your video in terms of aspect ratio. This is why it centered (the default for the node) and crops off the entire top and bottom, so it had no idea what her face looked liked, just that she was blonde which is also primarily the only reason the video pans up, because the model was trying to show the blonde hair you prompted.
I've learned that Wan likes to have descriptive prompts. Something too small like "A young woman with blonde hair reveals her boobs" isn't enough to get any meaningful results you had in mind, and it forces the model to just make something up. Your video did what you asked, to the model, she is revealing as in "showing" or "displaying" her boobs, but because you didn't mention "bare boobs" or "naked boobs" or even just mention her nipples, she wasn't generated to actually show them. Now, this stuff being what it is and how it works, it's possible different passes with different seeds would eventually show them.
Here I've taken your image, cropped off the bottom to create a true-ish 9:16 image to feed Wan 2.2. One trick I've used and used here is to ask Google to create a prompt for Wan 2.2 with replacement language. I asked it:
Take this simple prompt of a woman trying to bring emphasis to her necklace by touching it and holding it for the viewer. This is for a Wan 2.2 I2V prompt so it has to be descriptive: A young woman with blonde hair reveals her necklace
Whatever it spits out I replaced necklace and any other words that need to change to mean her boobs. Here was the original result:
A close-up shot of a young woman with blonde hair gracefully raising her hand to her chest. Her fingers gently pinch and lift a delicate silver necklace away from her collarbone, holding it up toward the camera to present it to the viewer. The camera holds focus tightly on the shimmering pendant of the necklace. Her expression is warm and intentional, with a soft smile, ensuring the viewer's attention is entirely drawn to the jewelry. The motion is smooth, deliberate, and elegant, with subtle, natural micro-movements in her hair and shoulders.
I changed it to:
A close-up shot of a young woman with blonde hair lifts her blouse above her chest exposing her bare breasts and nipples. Her fingers gently touch her breasts, holding them up toward the camera to present them to the viewer. The camera holds focus tightly on her breasts. Her expression is warm and intentional, with a soft smile, ensuring the viewer's attention is entirely drawn to the jewelry. The motion is smooth, deliberate, and elegant, with subtle, natural micro-movements in her hair and shoulders.
Here is the image and the video that came it produced on the first try, which may have been luck, BUT it certainly didn't have her lift her shirt, she destroyed it LOL. But, you get the idea.
View attachment 3856451
