Chat AI

The Generative Video Model Race Continues!

Chat AI
#Veo #Veo 3.1 #Google
image

A week and a half ago, OpenAI released Sora 2 and its Pro version, which immediately spawned a ton of viral videos. We got caught up in it ourselves and generated a capybara — check out the post above.

Today, Google released a beautiful response that’s even more impressive. Meet Veo 3.1 and its Fast version, which generates slightly lower quality content but is a bit faster.

So, what you need to know about Veo 3.1

1️⃣ First, Google has updated light and shadow generation in Veo. Now light falls approximately as it does in the real world, obeying the laws of earthly physics. Shadows look realistic and appear exactly where they should be.

2️⃣ Second, Veo 3.1 supports scenes. This means you can write actual scripts for generation, specifying the content of each scene. It’s easiest to create humorous sketches or backgrounds for advertising. Try a prompt like this:

  • Scene 1: A girl walks along the beach, looks at the sea, romantic atmosphere
  • Scene 2: She holds seashells in her hands that form the shape of a heart
  • Scene 3: She continues walking towards a beach camp where a large group of people is gathered

Add a voiceover advertising island tours and use it in targeted advertising.

3️⃣ Third, they added a “First and Last Frame Control” tool. You can upload a reference for the first and last frame of the video, and Veo will try to generate video so that the image organically transitions from the first to the last frame. For example, you can set a character’s pose at the start and end of the video, and Veo will generate the transition from pose A to pose B. Convenient for marketplace product cards.

4️⃣ Fourth, audio generation has been improved. Speech in Veo is now more or less lively and matches the character (for example, a capybara will speak in a high cartoon voice by default, as befits any well-mannered capybara). Ambient sounds have also been improved — they’re more accurate and better match the scene’s context.

5️⃣ Fifth, videos can be generated with realistic physics. A scarf will flutter in the wind, dandelion seeds will fly according to the laws of earthly physics, and so on.

6️⃣ Sixth, Veo 3.1 and Veo 3.1 Fast are already available in our service!

Video generation takes 2–3 minutes. All the features mentioned above are also available. Hurry up and try it.

You can try it, as usual, in the chatbot @chat_ai_tg_bot

And on the website chataibot.pro

← Previous Article Back to blog Next Article →
Free access to Chat GPT and other neural networks