[12:51 Tue,26.March 2024 by blip] |
How can video AI be used creatively in film? Not primarily as a camera replacement, because the images are still (?) not really realistic, especially when people are depicted, and you have little or no control over what exactly is generated and how. And yet the synthetically generated images are often fascinating. Not least because they are created from nothing, but also because of their very unique, dreamed-up look. ;) So, if we look at generative video AI as a new tool in the filmmaking toolbox, what can it achieve? When Runway hosted its Gen:48 competition at the end of last year, we were particularly struck by  those clips or sometimes just image sequences that were less interested in realism (a narrow path that, in our opinion, still leads through the Uncanny Valley) than in interesting images of all kinds. Experimental short films, where consistency in time and space is secondary to undesirable, populated by animated creatures - such videos are complex to produce with conventional means using CGI and motion graphics, but much quicker with AI.
OpenAI is now presenting selected feedback from the creative scene on its video AI Sora, which was presented in February - The reviews are very positive, which is hardly surprising in this marketing context. Nevertheless, the statements are not uninteresting to read and the clips are also worth seeing - Sora creates even more realistic, moving images from the Nu (but thanks to a lot of computing power in the background) than previously available AIs, but here too it is clear that it is the strange and surreal images that are more inspiring - images that would be difficult to create without AI. According to Paul Trillo, Sora can be used to implement new and virtually impossible ideas, while the multi-media collective shy kids appreciates the opportunity to create completely surreal things. Quoted creative agencies see the advantage in Sora of being able to visualize concepts in advance and quickly play through different versions, and also that their idea generation is no longer limited by small budgets. In the long term, Sora could give rise to new storytelling that is free of technical hurdles.
So much for a first glimpse of what can be realized with video AIs beyond short memes and fake videos.  more infos at bei openai.comdeutsche Version dieser Seite: Entfesseltes Storytelling mit der Video-KI Sora? |
 |
|
|