As most of You know, one of the main reasons using the VEAI is enhancing old videos with resolution, sharpening etc.
I suggest trying to focus on some of the major Video File Formats (and versions) back in each of the 3 decades, improving the models for 2-3 formats each decade, making it easier to chose the AI model. Most videos doesn’t need a perfect enhancement, but just some improvement.
Making the models for this will have a pretty big area to cover. But it’s better to have a bunch of video files improved by 2 or 3 times what they were, without consuming too much time when enhancing.
Making this a suitable selection for users that have a pc that aren’t that powerful.
This is just a thought, so I will leave like this before I will get deeper into it.