Very impressive tutorial. Loved your time travel video at the end. Even though there are obvious limits, the fact it is at the stage an entertaining storyline can be manifested is great. It involves a lot of work, though, at this stage, that is why won't see tons of people doing it.
New stack: - MJ - CapCut or Canva to cut out character - Animate background separately with Pixaverse (better more stable motion brush) or Stable Video - or just camera move with... - Lip Sync with Pika or Sync Labs - Optional cleanup step with Topaz or some of these ComfyUI steps
Thank you, great video! However, Pika doesn`t offer lip sync at the moment and runwayml didn`t work with my animation videos and it also didn`t work with faces in profile
Pika does offer Lip Sync, I just tried it last night. You have to upload an image or video, then you'll see the Lip Sync button. It might not work with animated characters. They have to look human-like for it to recognize the face.
Can you please let me know how to lip sync when there are 2 characters in the same scene? is it possible? Runway keeps detecting the wrong face. (I also have a robot in the scene which isnt getting detected at all)
The only way I know of currently is to cover/paint over one face in the photo and lip synch the first character. Then paint over the second face and lip synch the other character. In your editing program, crop or mask the two clips so you only see the talking faces.
P.S. You need to add 0:00 at the top of your chapters list in the description so that the chapter markers will show up on the videos timeline, otherwise they are just buried in your description area, hard to find and not helpful to viewers
@@aivideoschool I figured that. My pleasure my friend. By the way, in my hunt to find the best quality AI videos, I stumbled across your channel and I saw you reviewing a lot of the platforms that are already somewhat known. But I just stumbled across replikant. It's more of a 3D animation software, but I think it may be relevant to some of your followers as it allows ultra realistic storytelling videos to be created with their platform very easily without having to have years of 3D animation experience. It's a new platform made by epic games I believe.
Glad I showed up in your hunt. Replikant looks so cool but I don't have an Nvidia GPU (I'm on a Mac M2). I wish they had a cloud version like Wonder Dynamics but it looks a lot better than Wonder.
The best way I found prior to Midjourney's update was to uniquely name the character in the prompt and use a detailed physical description that includes age. Then you copy the seed number from an image you like. If you add onto the original prompt with the same seed, it's sort of similar. I don't think Dalle will give you a seed but Leonardo AI will and is also free.
@@aivideoschool thanks my friend. there's a new way to make ai movies you have to make a huge video about it which is creating a full body shot image for a character and his background and make the character move with viggale ai. this is much much better than the slight motion we have with runway, pika or haiper.