New approach with AI video generated content

I did some AI generated animations for my bycicle website.

Some general camera x movement and - the problem with the white background couldnt be solved by AI - so I used the chromakey function in cin.

A weird 16 FPS and a *free *wide ratio for the background which I just dropped in and cin was able to handle that fine :slight_smile:

Here it is: Wie lange ist ein Fahrradhelm sicher?

I know many people dont like AI nowadays but for me a as media designer it is important to follow the trend.

All animations were rendered locally on my laptop and afterwards combined with cin.

Im sure going to edit the clip but it generally is 90% completed. So only some timing and length issues would change.

I meanwhile modified the result with Optical Flow from 16 to 24 FPS which seems to me to be the best option.

An interesting experiment with good results. Indeed, 16 fps made the video too jerky; 24 fps is better.

Thank you for your feedback. It makes a difference between recording at 16 fps and the then reframe to 24 fps vs recording directly with 24 fps.AI generated content looks more flat and synthetic.Leaving off some information brings that kind of bit of AI-charme.In this example I probably should have prepared all content before I work with cin in 24 FPS instead of editing with 16 fps and reframing to 24 fps afterwards.Why? Because the background, which is just one huge wide aspect ratio image is lacking information and would have taken advantage of being created at full resolution (24 FPS).

Hi all, I know this thread is not related with this, But I would like to know if somebody have tried this chat with AI for properties: https://parlantia.com/