Reference Studio on Reeldo
Reference to Video AI Generator
Generate new scenes from reference images, videos, and audio — keep characters, products, and motion consistent. References guide the output; they are not the first frame.
上传参考图像、视频或音频，引导动作、风格与声音。
模型
Sound-On最高画质与动作控制 — 最高 4K
提示：在 Prompt 中用 @image1、@video1、@audio1 引用上传内容（按上方顺序）。
输出设置
分辨率
时长
宽高比
原生音频
与视频一同生成对白、环境音与音效
如何使用 Reference Studio
用多模态参考引导动作、风格与声音。
- 1
上传参考素材
至少添加一张图像、一段视频或一段音频。登录后上传文件，或粘贴公开 URL。
- 2
用 @ 标签编写 Prompt
使用 @image1、@video1、@audio1，与上传顺序对应。说明每个参考控制什么。
- 3
设置输出并生成
选择 Seedance 2 Fast 或 Seedance 2 Pro、宽高比、时长与 Sound-On。可在 Studio 继续或直接生成。
角色外观、服装、场景或艺术风格。
镜头运动、动作、节奏或编舞。
音乐、节奏或环境音 — 配合 Sound-On 使用。
你的参考素材
尚无参考素材 — 请至少上传一张图像、一段视频或一段音频。
示例 Prompt
角色 + 动作
@image1 as the main character. Match the camera movement and pacing from @video1. Cinematic lighting, native ambience.
风格迁移
Use the art style and color palette from @image1. Scene: a rainy street at night. Slow dolly-in, moody atmosphere.
音频驱动
Sync visual beats to @audio1. @image1 defines the performer. Energetic dance cover, vertical 9:16.
限制与提示
- ·至少 1 个参考素材 — 最多 9 张图像、3 段视频、3 段音频。
- ·参考视频输入按输入+输出秒数计费，每秒费率更低。
- ·图像：JPEG, PNG, WebP · 每张最大 10 MB。
- ·视频：MP4, WebM, MOV · 最大 100 MB · 公开 URL 或上传。
- ·音频：MP3, WAV, OGG · 最大 20 MB · 公开 URL 或上传。
- ·Seedance 2 Fast 不支持 4K。开启 Sound-On 可生成原生音频。
- ·生成前会显示积分消耗。
常见错误
- ✕上传了参考素材，但 Prompt 中未提及 @image1 / @video1。
- ✕未上传任何参考素材就点击生成。
- ✕粘贴 Reeldo 无法访问的私有或非公开 URL。
What you can create
Character consistency
Upload portrait or outfit references and generate new shots where the same character speaks, moves, and looks consistent across scenes.
Motion & camera reference
Provide a short reference clip to match camera movement, pacing, or action style while changing the scene or subject in your prompt.
Product & brand
Use product stills as references and describe a new lifestyle or hero scene — the item stays visually accurate while the environment changes.
Audio-guided rhythm
Add reference audio for music, voice, or ambience. Seedance and MiniMax H3 can align motion and sound to your reference track.
Reference to video vs image to video vs text to video
| Feature | Reference to Video | Image to Video | Text to Video |
|---|---|---|---|
| Primary input | Images + video + audio references | One start frame (optional end frame) | Text prompt only |
| Best for | Consistent characters, motion/style transfer, multimodal control | Animating a still photo | Quick concepts from words alone |
| Identity consistency | Strong — references define look, not the first pixel | Limited — tied to the uploaded frame | Varies per generation |
| Reference video / audio | Yes — up to 3 videos and 3 audio clips | No | No |
Supported models & reference limits
Reference to video runs in Reference Studio mode on Reeldo. Seedance 2 Fast, Pro, and 2.5 accept up to 9 images, 3 reference videos, and 3 audio clips per generation. MiniMax H3 supports multimodal references at 768P or 2K with native stereo audio. Kling and Veo on Reeldo use text-to-video and image-to-video only — switch to Seedance or H3 for reference-driven clips. Credits with reference video input are billed on combined input + output duration; see Pricing for details.
How reference to video works
Step 1
Upload references
Add images, video URLs, and audio URLs — characters, motion refs, or sound guides.
Step 2
Write your prompt with @tags
Use @image1, @video1, @audio1 to say what each reference controls in the new scene.
Step 3
Generate & download
Pick Seedance or H3, preview in Studio, then download your reference-driven clip.
Reference to video FAQ
- What is reference to video AI?
- Reference to video generates a new clip guided by reference images, videos, or audio — for example keeping a character's face consistent or matching a camera move. The references steer the output; they are not simply animated as the first frame.
- Reference to video vs image to video — what's the difference?
- Image to video animates a single starting image. Reference to video uses one or more references to control identity, style, motion, or sound while your prompt describes a new scene. Use image to video for photo animation; use reference to video for consistency and multimodal control.
- What is Reference Studio on Reeldo?
- Reference Studio is Reeldo's in-app workflow for reference to video — the same multimodal mode you see in Video Studio with @image1, @video1, and @audio1 tags.
- Which models support reference to video?
- Seedance 2 (Fast, Pro, and 2.5) and MiniMax H3 support reference to video on Reeldo. Kling and Veo use text-to-video and image-to-video modes only.
- How do @image1 and @video1 tags work?
- Tags match the upload order shown in Studio. Mention @image1 in your prompt to tell Reeldo how to use the first reference image, @video1 for the first reference video, and so on.
- How many references can I upload?
- On Seedance, up to 9 images, 3 reference videos, and 3 audio clips per generation. MiniMax H3 supports multimodal references with similar limits. At least one image or video reference is required when using audio references.
- How are credits calculated with reference video?
- When reference video input is used, credits are billed on combined input + output duration (for example 3s reference + 5s output = 8 billable seconds). The estimator in Studio shows the total before you generate.
- Can I use reference to video commercially?
- Paid plans include commercial use. Ensure you have rights to all reference images, videos, and audio you upload.