PROMPT RECORD图像记录

AI FILMS Studio 上线了两款新的图生视频模型。Vidu Q3 Turbo 和 Vidu Q3 Drama。不同的工具,不同的用途。 Vidu Q3 Turbo 可以快速将单张参考...

AI 图像 zaesarius Fri Jul 03 21:16:11 +0000 2026

Two new image to video models are live on AI FILMS Studio. Vidu Q3 Turbo and Vidu Q3 Drama. Different tools, different jobs.

Vidu  Q3 Turbo animates a single reference image fast. Built in prompt  enhancement, Movement Amplitude control (Auto / Small / Medium / Large),  synchronized audio, and background music. Both audio options are free,  no extra credits. Resolutions 540pX to 1080pX at 40 to 80 credits per  second.

Vidu Q3 Drama takes a full  script (up to 5,000 characters) and up to 14 named assets for  characters, locations, props. The model tracks each asset ID across  every shot and holds visual identity without drift. It has also three  placement modes: Upload (default), Panorama for wide environments from a  single reference, and Step, which feeds each scene's final frame into  the next generation. The duration is up to 30 seconds. A flat rate of  140 credits per second.

A prompt tip  for Turbo is to always try to describe motion specifically. "Her hand  slowly reaches toward the camera, fingers slightly blurred at full  extension" this outperforms "hand moves" for example.

A  prompt tip for Drama is to write where everyone is, what they do, and  where the camera sits. Summary writing produces inconsistent results.

Both, Vidu Q3 Turbo and Drama also runs in the Nodes Graph Editor.

Full guide with credit tables.

https://t.co/DtXQ1O5o4D

AI FILMS Studio 上线了两款新的图生视频模型。Vidu Q3 Turbo 和 Vidu Q3 Drama。不同的工具,不同的用途。 Vidu Q3 Turbo 可以快速将单张参考图动画化。内置提示词增强、运动幅度控制(Auto / Small / Medium / Large)、同步音频以及背景音乐。两种音频选项均为免费,不额外消耗积分。分辨率从 540pX 到 1080pX,每秒 40 到 80 积分。 Vidu Q3 Drama 可接收完整脚本(最多 5,000 个字符)以及最多 14 个已命名资产,用于角色、场景和道具。该模型会在每个镜头中追踪每个资产 ID,并在无漂移的情况下保持视觉一致性。它还提供三种放置模式:Upload(默认)、Panorama(通过单张参考图生成宽幅环境),以及 S...

查看来源

中文说明

AI FILMS Studio 上线了两款新的图生视频模型。Vidu Q3 Turbo 和 Vidu Q3 Drama。不同的工具,不同的用途。 Vidu Q3 Turbo 可以快速将单张参考图动画化。内置提示词增强、运动幅度控制(Auto / Small / Medium / Large)、同步音频以及背景音乐。两种音频选项均为免费,不额外消耗积分。分辨率从 540pX 到 1080pX,每秒 40 到 80 积分。 Vidu Q3 Drama 可接收完整脚本(最多 5,000 个字符)以及最多 14 个已命名资产,用于角色、场景和道具。该模型会在每个镜头中追踪每个资产 ID,并在无漂移的情况下保持视觉一致性。它还提供三种放置模式:Upload(默认)、Panorama(通过单张参考图生成宽幅环境),以及 S...

原始 Prompt

Two new image to video models are live on AI FILMS Studio. Vidu Q3 Turbo and Vidu Q3 Drama. Different tools, different jobs.

Vidu  Q3 Turbo animates a single reference image fast. Built in prompt  enhancement, Movement Amplitude control (Auto / Small / Medium / Large),  synchronized audio, and background music. Both audio options are free,  no extra credits. Resolutions 540pX to 1080pX at 40 to 80 credits per  second.

Vidu Q3 Drama takes a full  script (up to 5,000 characters) and up to 14 named assets for  characters, locations, props. The model tracks each asset ID across  every shot and holds visual identity without drift. It has also three  placement modes: Upload (default), Panorama for wide environments from a  single reference, and Step, which feeds each scene's final frame into  the next generation. The duration is up to 30 seconds. A flat rate of  140 credits per second.

A prompt tip  for Turbo is to always try to describe motion specifically. "Her hand  slowly reaches toward the camera, fingers slightly blurred at full  extension" this outperforms "hand moves" for example.

A  prompt tip for Drama is to write where everyone is, what they do, and  where the camera sits. Summary writing produces inconsistent results.

Both, Vidu Q3 Turbo and Drama also runs in the Nodes Graph Editor.

Full guide with credit tables.

https://t.co/DtXQ1O5o4D

暗调氛围多分屏广角远景低调光戏剧性画面文字标志或水印多面板布局