software Premiere Pro getting AI Editing

sfoster

Staff Member
Moderator
Coolest part IMO is the object removal/replacement... the ability to take away big store signs, replace them with a different name, such a boon for indie filmmakers.

The simple ability to remove light stands and stuff will make it a lot more convenient for indie film too.
Granted the diamond object generation in this looks bad - like really bad - and the sora inserts look awkward and out of place too, but it's just v1.


 
Last edited:
I'm not sure how I feel about it, but that's partially because I consider myself a VFX artist. All of the examples are things that I can already do and I spent a lot of time learning how to do them and make them look good. One thing I will say, this is the truest example of Computer Generated Imagery/CGI. We're approaching a point where soon you could press "D" for Dinosaur, "M" for Monster and the computer will do it without the need for an actual artist. I think Hollywood films are too specific with what they need and that they will require artists for sometime, but low budget kids shows like Power Rangers could use something like that, same with ultra low budget Indies. However, I predict there will be a distinct "look" to this kind of thing and that "look" will instantly devalue the film or show, at least initially.
 
I predict there will be a distinct "look" to this kind of thing

Yeah. For all the slickness in that promo, the few examples given are real out-of-the-box-bog-standard-stock-footage. Now while that might be suitable for churned-out, throwaway video, I don't believe this "AI" algorithm will be any better than the others at not producing homogenised visuals that clash with the rest of the production design on more substantial works - unless the studio/film-maker employs someone (let's call them a digital artist 🧑‍🎨 ) to spend hundreds of hours crafting exactly the right text prompt to bully the software into generating an image that satisfies the director.

For the record: I'm not against this kind of enhance digital assistance - I love my (non Adobe, non-cloud-based) photo editing software's ability to seamlessly stitch texture-rich images together to create an ultra-wide panorama, or focus-stack a series of close-ups into one sharp picture ... - but I wish the purveyors of this technology would stop hyping the beta versions. And stop slapping the "AI" tag onto everything. :grrr: It's not "intelligent" - it's just faster algorithmic processing that we've been using for several decades.
 
Last edited:
I spent a lot of time learning how to do them and make them look good. One thing I will say, this is the truest example of Computer Generated Imagery/CGI. We're approaching a point where soon you could press "D" for Dinosaur, "M" for Monster and the computer will do it without the need for an actual artist.

It's an incredible time.

Using a convoluted piece of software, navigating through sub-menu after sub-menu, watching 8-hour "classes" on how to use an application will all become things of the past.

You'll just be able to speak out loud and ask the computer to do it for you.
They call this "reducing the skill gap" of workers or something like that.
 
Of course I have some comment or other about this implementation, which looks good and useful, but I'm going to save myself the effort here and just reply in a seperate post (when I get a chance) that kind of covers a larger point relevant to all this stuff.

Basically, I don't feel as much need to speak on individual products as I do to speak on the dynamics of oversaturation, skill negation, etc. What happens once everyone can hit a higher mark without effort or training, and what that means for creators.

Anyway, feels like a whole separate topic.
 
Back
Top