demobook

Gemini: Successful multi-character dialogue generation in Google Veo 3

GeminiTry it →Watch full video →MisterQuigleyAI ·

Demo summary

Two successful generations from Google Veo 3 are demonstrated where the model correctly follows the prompt for a man and woman to exchange dialogue at a rainy bus stop with synchronized audio.

All demos from “Google Veo 3 - It's Worse Than You Think - Testing Demo & Review

  1. 10:320:41Testing Google Veo 3 native audio and speech generationA series of generated clips from Google Veo 3 are shown to test if the model can correctly sync speech and audio for two characters at a bus stop. The results show various failures including garbled speech and incorrect character assignments.Google AI StudioText to Video
  2. 11:220:08Evaluating character movement and speech in Google Veo 3A clip generated by Google Veo 3 shows a man in a red shirt asking for the time, demonstrating the model's ability to handle character movement alongside speech, despite some visual glitches.Google AI StudioText to Video
  3. 11:560:15Successful multi-character dialogue generation in Google Veo 3CurrentTwo successful generations from Google Veo 3 are demonstrated where the model correctly follows the prompt for a man and woman to exchange dialogue at a rainy bus stop with synchronized audio.GeminiText to Video
  4. Watch “Google Veo 3 - It's Worse Than You Think - Testing Demo & Review” →

Text to Video

  1. 1:581:20Generate cinematic video with Google Veo 3The video shows how to activate the video tool in Gemini and enter a detailed prompt to generate a slow-motion shot of chocolate chip cookies using the Veo 3 model.@KevinStratvert
  2. 4:081:13Refine video prompts using Gemini AIThe user demonstrates asking Gemini to write a more descriptive, cinematic prompt for a cookie-baking scene and then using that output to generate a better video.@KevinStratvert
  3. 13:380:25Dolly shot camera instructions in Google Veo 3.1A demonstration of using specific camera movement keywords like 'wide angle dolly in' to control the output of Google Veo 3.1.Youri van Hofwegen
  4. 3:180:19Generate video from custom text promptThe user demonstrates Gemini Omni's text-to-video capability by prompting it to show his avatar in a suit eating spaghetti at a restaurant.Paul J Lipsky
  5. 8:041:07Create skydiving video with Google Veo 3.1 via GeminiThe video shows the process of inputting a complex prompt into Gemini to generate an 8-second clip of a grandmother skydiving into a Super Bowl stadium using the Veo 3.1 model.AI Master
  6. 0:580:26Accessing Google Veo 3 in GeminiThe user demonstrates how to navigate to the video generation tab within the Gemini Advanced interface to access the Google Veo 3 model.@SocialtyPro
  7. 0:030:30Generate cinematic video with Veo 3 in GeminiThe video demonstrates high-quality cinematic video generation, including a close-up of a person and a stylized landscape, using the Veo 3 model integrated into the Gemini app.Google
  8. 2:201:02Generate video with audio in Google Veo 3The user demonstrates entering a text prompt into the Gemini interface to generate an 8-second cinematic video of a cookie including a specific narrated voiceover and sound effects using Google's Veo 3 model.@KevinStratvert
  9. 3:340:54Text-to-video generation in Google GeminiThe creator uses Google Veo 3 within the Gemini interface to generate an 8-second video of a scientist speaking specific lines from a text prompt.Tim Harris AI
  10. 1:211:01Generate AI video from text with Google Vids and VeoThe creator demonstrates how to use the Veo 3.1 integration within Google Vids to generate a video of a polar bear from a text prompt and insert it into the timeline.Paul J Lipsky
  11. 0:461:08Generate video from text with Gemini OmniThe user demonstrates how to navigate to the video icon in Gemini and enter a text prompt to generate a cinematic video of a storefront using the Omni model.@KevinStratvert
  12. 11:560:15Successful multi-character dialogue generation in Google Veo 3CurrentTwo successful generations from Google Veo 3 are demonstrated where the model correctly follows the prompt for a man and woman to exchange dialogue at a rainy bus stop with synchronized audio.MisterQuigleyAI