OpenAI recently introduced several videos made by artists who were allowed to test the video AI Sora. One of the videos in particular was impressive: “Air Head”. And what was already suspected when it was released now turns out to be certain: the video has been heavily edited, especially from a human perspective. A whole team worked on the effects and filtering out artifacts.
Advertisement
In “Air Head,” a yellow balloon floats over a man’s body. The balloon was originally obviously not yellow, but red. And even more so, he changed colors several times when trying to generate one or more consistent videos with Sora. That’s what Patrick Cederburg from shy kid tells us, who was responsible for the post-production of “Air Head”, while there was also a producer in Sidney Leeder and a director in Walter Woodman. Now you might wonder why so many people were sitting around watching an AI-generated video? Did they write the prompt together? Hardly likely. Accordingly, numerous questions about the video arose shortly after its publication. For example, it’s longer than OpenAI states how long the AI videos can be, it has sound, which Sora can’t do so far, and much more. OpenAI said about the published videos that they were created with Sora – there was no post-production or further information.
Classic Post-Processing: From Artifacts to Resolution
Cederburg tells the industry magazine FXGuide, it was really interesting to be able to play with Sora: “It’s a very, very powerful tool, and we’re already dreaming about all the ways it can be integrated into our existing process. But I think with any generative AI tool, control is still the most desirable and at the same time the most elusive at the moment.” To generate videos with Sora, you can really only enter one prompt; there is no help for creating multiple videos that build on each other. They therefore proceeded in a hyper-descriptive manner, inserting every detail, such as clothing, into new prompts in order to create new scenes that were then edited together. But: It required a lot of post-processing.
As an example, Cederburg cites a banana, of which the video AI does not have a direct image or any kind of stock photo, but only an idea: yellow, curved, dark ends. However, this creates a new or different looking banana every time. The balloon not only came out in different colors, some of it even had a scary face that had to be removed later. “‘Air Head’ uses footage exclusively generated by SORA, but most of it has been staggered, edited and stabilized, as well as upscaled or high-resolution.” The image sections were also changed, although Sora offers certain options for this; camera settings and angles were difficult to enter. What the filmmaker considers to be essential for filmmakers.
also read
Cederburg and his colleagues are apparently also testing other AI video tools. He says they all have similar problems, for example with camera settings, highlighting Rundway AI, which generated worse and shorter videos than Sora. In the end, there was a lot of material for “Air Head” that had to be edited and put together. There were hundreds of videos generated, each between 10 and 20 seconds long. He assumes a ratio of 300:1. And even the speed had to be adjusted – most of the clips that Sora generated looked like they were shot in slow motion. The voice of the guy with the balloon head, whose name is Sonny, is Cederburg himself. Everything is classically dubbed, without any AI.
Nevertheless, Cederburg still thinks Sora is impressive and obviously the work was so much fun that they are already working on a second part.
(emw)