- The Google videos generation model has obtained a major upgrade
- Announced at Google I / S, Veo 3 can combine audio and video in its release
- It is an ultra and American functionality for the moment
AI videos generation tools such as Sora and Pika can create realistic alarming video pieces, and with enough efforts, you can link these clips together to create a short film. One thing they cannot do, however, is to generate audio simultaneously. Google’s new VEO 3 model can, and that could change the situation.
Announced Tuesday at Google I / O 2025, Veo 3 is the third generation of the powerful generation of GEMINI videos generation. With the right prompt, he can produce videos that include sound effects, background noises and, yes, a dialogue.
Google briefly demonstrated this capacity for the video model. The clip was an animation of CGI grade of certain animals speaking in a forest. The sound and the video were in perfect synchronization.
If the demo can be converted into real use, this represents a remarkable shift in the AI content generation space.
“We leave the silent era of the generation of videos,” said Demis Hassabis, CEO of Google Deepmind, during a press call.
Lights, camera, audio
He is not mistaken. Until now, no other AI videos generation model can simultaneously provide a synchronized audio, or audio of any kind whatsoever, to accompany the video output.
It is still not clear if Veo 3, which, like its predecessor, Veo 2, should be able to publish a 4K video, exceeds the leader of the current Videos Openai Sora in the video quality department. Google, in the past, said that Veo 2 is able to produce a realistic and coherent movement.
Anyway, the release of what seems to be fully produced video clips (video And Audio) can instantly make Veo a more attractive platform.
It is not only that Veo 3 can manage dialogue. In the world of cinema and television, substantive noises and sound effects are often the work of Foley artists. Now imagine if everything you have to do is describe to see the sounds you want behind and attached to the action, and he does everything, including video and dialogue. It is a job that takes the animators of weeks or months to do.
In a press release on the new model, Google suggests that you tell the AI ”a news in your prompt, and the model gives you a clip that gives it to life”.
If Veo 3 can follow the output prompts and minutes or, in the end, hours of coherent video and audio, it will not take long before we aimed at the first animated feature generated entirely via Veo.
VEO is live today and available in the United States as part of the new ultra tier ($ 249.99 per month) in the Gemini application and also as part of the new flow tool.
Google has also announced a few updates to its VEO 2 video generation model, including the possibility of generating a video according to the reference objects you provide, camera controls, overpressure to convert portrait into a landscape and add and erase objects.
@Techradar ♬ Sound Original – Techradar
