Nano Banana, GPT Image, Flux, Seedream, Seedance and more. Try it now for free!

Created 2 years ago · 5 comments· 0 likes
Stable Video Diffusion (SDV 1.1)
Join the conversation
#2 this applies both to the visual impact/reality, as well as to the prompt ... from my short experience, if you want the animation feature to produces something worthwhile, both in start image and start prompt should be something that it can latch onto with regards to creating motion, going from state A to state B ... some dynamic ... honestly speaking I gave up on it for the time being ... due to .. it cannot yet put everything into movement ;) give it time, it'll come around, eventually ...
oh i get it now, you were commention on the animation retake in the first place? I am not sure I have so much worthwhile to say on that? cause I only have used it like? idk more than 5 but less that 10 times? roughly? ... my gut feel from that is: it needs a start image that has some sort of inherent dynamic, movement... if you go through my recents ... you will find "moonfall" ... big green moon and northern lights ... I animated it ... it ended up *animating the northern lights in the sky, but nothing else ... I also animated some portrait scene recently ... with a female in center in front of ... idk what ... the universe, stars? and it did what the animation did to yours ... just move the image from bottom to top ... what I take from this is: the AI for the animation, seeks for, NEEDS, a starter scene that has some remnant of movement dynamic. if ou feed it a stale image, as of NOW, for the time being, it does not always know what part of the image to put into animated movement...
Gary DeVeau has hidden the start image

#3 another example would be the adventurer girl having a drink I animated?