# Sound in the same pass

Period: May – Jun 2025 · Era: 2025 · The Studio Era

Veo 3, Flow and the recipe behind the Veo 3 ad.

Video models learned to speak. Veo 3 generated dialogue, sound effects and ambience in the same pass as the picture, which made short spoken scenes and ads possible in a day. Around it appeared the first director apps for arranging shots into a sequence.

## How it worked

1. Write the script, then have an LLM draft a prompt per shot with the dialogue in quotes.
2. Generate each shot in Veo 3 (eight seconds, native dialogue, effects and ambience), text-to-video or from a still.
3. Assemble and extend in Flow's Scenebuilder with camera controls and asset management, or cut in Premiere.
4. Seedance 1.0 introduced native multi-shot: two or three cuts inside a ten-second clip.
5. For the rest of the stack: Midjourney Video animates any Midjourney image; Higgsfield Soul locks a character; Eleven v3 adds inline audio tags and multi-speaker dialogue.

## Why it mattered

Audio in the same pass killed the separate TTS → lip-sync step for dialogue-light work, and Flow established the director-app category.

## Tools of the era

- Veo 3 (2025): https://deepmind.google/models/veo/
- Flow (2025): https://labs.google/flow
- Seedance 1.0 (2025): https://seed.bytedance.com/en/seedance
- Midjourney Video V1 (2025): https://midjourney.com
- Eleven v3 (2025): https://elevenlabs.io
- Higgsfield (2025): https://higgsfield.ai

## Example videos and posts

- The Kalshi NBA Finals spot: two days, about two thousand dollars · PJ Accetturo: https://x.com/PJaccetturo/status/1932893260399456513
- Veo 3 talking-animal vlogs, explained · Community tutorial: https://www.youtube.com/watch?v=RqQUNnzlsB4
- Veo 3 and Flow: native audio inside a director app · Theoretically Media: https://www.youtube.com/watch?v=oSerkVI-s84
- A stunning AI short film with seven tools · Curious Refuge with Mark Wachholz: https://www.youtube.com/watch?v=Ccxopwerw40

## On the timeline

- [Veo 3](https://aifilmhistory.org/entry/veo-3) (2025-05)
- [Seedance 1.0](https://aifilmhistory.org/entry/seedance-1-0) (2025-06-16)
- [Midjourney V1 Video Model](https://aifilmhistory.org/entry/midjourney-video-v1) (2025-06-18)
- [Hailuo 02](https://aifilmhistory.org/entry/hailuo-02) (2025-06-18)

## Sources

- Google Flow: https://blog.google/technology/ai/google-flow-veo-ai-filmmaking-tool/
- Seedance 1.0: https://seed.bytedance.com/en/seedance
- NPR on the Kalshi ad: https://www.npr.org/2025/06/23/nx-s1-5432712/ai-video-ad-kalshi-advertising-nba-finals
- Ancestra at Tribeca: https://variety.com/2025/digital/news/darren-aronofsky-ai-studio-primordial-soup-google-deepmind-1236403412/

---
Source: https://aifilmhistory.org/workflows/native-audio-director-apps