Useful information
Prime News delivers timely, accurate news and insights on global events, politics, business, and technology
Join our daily and weekly newsletters to get the latest updates and exclusive content on industry-leading AI coverage. More information
Google goes head to head against OpenAI‘s Sora with the newest version of its video generation model, I see 2, which says brands more realistic videos.
The company also updated its Image 3 imaging model to produce richer, more detailed photographs.
Google said Veo 2 has “a better understanding of real-world physics and the nuances of human movement and expression.” It is available in google labs VideoFX platform – but only on the waiting list. Users will need to register through a Google form and wait for Google to provisionally grant them access at a time of their choosing.
“Veo 2 also understands the language of cinematography: ask for a genre, specify a lens, suggest cinematic effects and Veo 2 will deliver them, in resolutions up to 4K,” Google said in a blog post.
While Veo 2 is available only to select users, the original Veo is still available on Vertex AI. Videos created with Veo 2 will contain Google’s SynthID metadata watermark to identify them as AI-generated.
Google admits, however, that the Veo 2 can still cause hallucinations in additional fingers and the like, but promises that the new model produces fewer hallucinations.
Veo 2 will compete against OpenAI’s recently launched Sora video generation model to attract filmmakers and content creators. Sora had been in preview versions for a while before OpenAI made it available to paying subscribers.
Surprisingly, Google says that in its own internal tests that measured “overall preference” (i.e. which videos the audience liked the most) and “adherence” (how well the videos matched the instructions given by the creator human), Veo was preferred by human evaluators. to Sora and other rival AI models.
Google announced Veo in May of this year during its Google I/O developer conference with a video made in collaboration with actor and musician Donald Glover, also known as Childish Gambino.
AI video generation still needs some work
AI video generation has long been an area of generative AI where large model developers such as Google and OpenAI regularly compete and play catch-up with relatively smaller companies.
PistaMLone of the pioneers in AI video generation, recently launched advanced controls for its Gen-3 Alpha Turbo model. Pika Laboratories released Pika 2.0, giving users more control and allowing them to add their own characters to a video. Advanced AI announced a partnership with AWS to bring its models to Bedrock for enterprise use. Luma also expanded its Dream Machine generation model.
However, AI video generation still needs to convince both creators and viewers. After Sora’s long-awaited release, people remained skeptical about his capabilities as he continued to generate figures that defied physics and anatomy. Users felt it gave inconsistent results.
Also a preview of the recent Game Awards. showed people’s distrust of what they perceive as “low AI.”
However, some filmmakers have begun to take advantage of the possibilities that AI video generators can offer. Famed director James Cameron joined the board of directors of AI Stabilitywhile actor Andy Serkis announced that he was building a production company focused on artificial intelligence.
Google said it is seeing interest from many users. The company said YouTube creators have been using VideoFX to create backgrounds for YouTube Shorts to save time.
Image Updates 3
Google also updated its Image 3 image model, which it recently made available through its Gemini chatbot on the web, to be more realistic and offer brighter images.
Image 3 can now more accurately represent more art styles, “from photorealism to impressionism, from abstract to anime.” Google said the model will also follow directions more closely.
People can access Image 3 through ImageFX.
Source link
