Join our daily and weekly newsletters for the latest updates and exclusive content on our industry-leading AI coverage. He learns more
Google He is going head to head against OpenAISora with the latest version of its video generation model, Veo 2, which says makes More realistic videos.
The company has also updated its Imagen 3 image generation model to produce richer, more detailed images.
Google said Veo 2 has a “better understanding of real-world physics and the nuances of human movement and expression.” It is available on Google Labs Video FX platform – But only on a waiting list basis. Users will need to register through a Google form and wait for access to be temporarily granted by Google at a time of their choosing.
“Veo 2 also understands the language of cinematography: ask it the genre, select the lens, suggest cinematic effects, and Veo 2 will deliver up to 4K resolution,” Google said in a blog post.
While Veo 2 is only available to select users, the original Veo remains around available On Vertex AI. Videos created with Veo 2 will contain Google’s metadata watermark SynthID to identify them as created by AI.
However, Google admits that the Veo 2 may still hallucinate extra fingers and the like, but promises that the new model produces fewer hallucinations.
Veo 2 will compete with OpenAI’s recently released Sora video generation model to attract filmmakers and content creators. Sora was in previews for a while before OpenAI made it happen available To pay subscribers.
Impressively, Google says that in its internal tests measuring “overall favorability” (which videos audiences like best) and “immediate engagement” (how closely videos match instructions provided by the human creator), Veo was favored by evaluators Humans. For Sora and other competing AI models.
Google announced View in maybe This year during the Google I/O developer conference with a video produced in partnership with actor and musician Donald Glover, also known as Childish Gambino.
AI video creation still needs some work
AI video creation has long been an area of generative AI, with large model developers, such as Google and OpenAI, regularly competing with and catching up with relatively smaller companies.
RunwayMLone of the leaders in AI video production, recently launched advanced controls for its videos Gen-3 Alpha Turbo model. Becca Laboratories Pika 2.0 releasedGiving users more control and enabling them to add their own personality to the video. Loma I Announced a partnership with AWS for Bring their models to Bedrock For enterprise use. Loma too She expanded her dream machine generation model.
However, video production using AI still needs to convince both creators and viewers. After the long-awaited release of Sora, People remained skeptical Of her abilities when she continued to generate numbers that defy physics and anatomy. Users felt that it gave inconsistent results.
A trailer from the game’s recent awards as well Show people’s distrust For what they consider “the decline of artificial intelligence.”
But some directors are starting to do so Embrace the possibilities AI video generators can provide. Acclaimed director James Cameron has joined the board of directors Stability AIWhile actor Andy Serkis announced that he is building a production company focused on artificial intelligence.
Google said it is seeing interest from many users. YouTube creators use VideoFX to create backgrounds for short YouTube videos to save time, the company said.
Image updates 3
Google has also updated its image model Imagen 3, which it recently made available through its Gemini chatbot on the web, to be more realistic and offer brighter images.
Imagen 3 can now accurately represent more artistic styles, “from photorealism to impressionism, from abstraction to cartoon.” Google said the model will also follow claims more faithfully.
People can access Imagen 3 through ImageFX.
https://venturebeat.com/wp-content/uploads/2024/12/google-veo-2.png?w=1024?w=1200&strip=all
Source link