in

The next milestone for the IAS that generates video was to make them with audio. Google has achieved it with I see 3

Great day for Google. We are in full I/O 2025, the most important software event for the American company. Interestingly, Android is being one of the least sounded names: this year the only thing that matters is AI.

And, related to AI, Google has been working on a model that allows you to generate video through text. That model is I seeand in its new update it is able to generate these videos … with audio.

I see 3. Google has three levels for its generative artificial intelligence of video. I see 1, I see 2 And the new I see 3. Yes, they are much easier names regarding what We are accustomed to us. I see 3 is the most powerful model, capable of generating 4K video with advanced film compression. In this Google I/or gains a key capacity: the generation of video with audio.

Of environmental sounds to dialogues. Google goes with everything with I see 3. This model not only has a higher quality with respect to I see 2: it is the only one of Google capable of generating videos with audio. For example, if in the prompt we detail that we want an urban scene, it will be able to recreate some of the sounds corresponding to it (people walking, traffic, bustle, etc.).

Google goes further, and promises to be able to create even dialogues between characters. This is one of the definitive barriers for text to text to become practically a science fiction function. With I see 3 it will be possible to do everything.

IMPROVEMENTS IN SEE 2. Although I see 3 is the absolute protagonist, I see 2 is updated with new functions. Among them, it premieres new camera controls much more precise for Traveling and Zoom movements, outpainting options to expand the framing (to pass the vertical to horizontal or vice versa video), as well as the possibility of adding or deleting elements of the video.

Flow arrives. Related to VI, Image and Gemini arrives Flow, the new Google tool to create cinematographic videos through AI. It is a new work environment to be able to give free rein to our creations with I see: a video editor with whom we can create both with image and I see.

In addition to functioning as editor, it will have some social function. Through Flow we can access Flow TV, a feed in which we will see content, channels and creators who are generating videos with I see.

Ahead of Open AI. Chatgpt creators surprised the world with Soraits artificial intelligence to generate video from a prompt. The problem? At least, at the time we write these lines, it is not able to generate video.

In December 2024 Google already advanced Sora on the right Showing the capabilities of VER 2, which quadrupled the video output resolution with respect to the Open AI model. It also allowed to create more durable videos, and a “understanding” of spectacular physics, something that makes the difference when creating a natural video.

Your rivals. Rival video generators such as Runway, Luma ai either Pika Labs They allow to add external audio, but in no case generate sound at the time of delivering the final video.

Google has just been punched on the table with I see 3, maintaining the first career position and further complicating things to giants like Open AI.

At the moment, these functions will be available for GEMINI Ultra subscribers in the United States through the Gemini and Flow app, as well as for companies through VERTEX AI.

Image | Google

In Xataka | 14 tools to create free images

What do you think?

Leave a Reply

Your email address will not be published. Required fields are marked *

GIPHY App Key not set. Please check settings

Google already offers synthid to mark content generated by the as such. Now it will help us identify them

Ai Mode, Project Beam, I see 3, the Project Aura, Jules glasses and everything presented in a Google I/or 2025 loaded with ambition