Vidu Q4 Reference to Video: Consistent Casts, Synced Voices
Last verified: October 29, 2026
Vidu Q4 takes a set of reference images, plus voice samples if you want them, and returns one clip where the characters, styles, scenes and voices all match their references. Picture and sound are generated together, and you can render up to 4K. Put your cast, props and locations in Add Image, add a voice sample in Add Audio if a character needs to sound a certain way, and describe the action in the Prompt box. With Vidu Q4 reference-to-video, each subject stays true to the image you gave it, and the prompt decides what happens.
Vidu Q4 does more than keep a subject steady. In this mode it can also follow a reference sound and switch camera angles by itself while keeping subjects the same from one camera position to the next. It also outputs audio and video together. So one generation can play out as a short scene with more than one angle, not just a single fixed shot.
ShengShu built the model for teams that make a lot of video: short-drama studios, ad agencies, social creators and film crews. In all of that work, continuity is the hard part: the same face, outfit and voice in every clip. You set clip length, frame shape and resolution yourself, so one reference set can give you both a vertical social cut and a widescreen master without rebuilding the cast.
Explore more reference to video
Vidu Q4 settings at a glance
Inputs, clip options and output for each reference-to-video run.
Input
Up to 15 images (JPG, JPEG, PNG or WEBP, up to 10 MB each) + up to 3 audio files (MP3, up to 25 MB each) (optional) + a text prompt of up to 5,000 characters
Video Duration
3 to 16 seconds
Resolution
540P, 720P, 1080P, 2K or 4K
Aspect Ratio
16:9, 9:16, 1:1, 4:3, 3:4
Audio
Switch on or off, on by default
Output
1 video per run
How do you get more control over Vidu Q4 reference scenes?
Name each reference by number in the prompt
Write the camera plan into the prompt
Save the top resolutions for the master you will grade
How do you make a video from reference images with Vidu Q4?
Six steps take you from your reference set to a downloaded clip.
Add reference images in "Add Image"
Upload the characters, products, locations or style frames the clip has to follow. "Add Image" is required, so every run starts from your own pictures. Give each file one clear job: a front-facing portrait for a face, a clean shot of the product, a wide frame of the location. Vidu Q4 builds the video around the subjects it finds in these images. The FAQ lists accepted formats, sizes and file counts.
Add voice samples in "Add Audio"
This upload is optional. Leave it empty for ambient scenes or narration that doesn't need to match a particular voice. Add a sample for each character whose voice should carry into the clip. Use recordings of one speaker with no music underneath, so each file contains only the voice you want.
Write the scene in "Prompt"
Describe who does what, where it happens, how the camera moves and any lines a character speaks. The box takes up to 5,000 characters. Let the images carry how things look and use the words on action. If you only have a rough idea, the "Enhance Prompt" button under the box suggests fuller versions. Your prompt only changes if you pick one of them.
Set "Video Duration", "Resolution" and "Aspect Ratio"
"Video Duration" goes from 3 to 16 seconds in one-second steps and starts at 5. "Resolution" offers 540P, 720P, 1080P, 2K and 4K, with 720P selected to start. "Aspect Ratio" offers 16:9, 9:16, 1:1, 4:3 and 3:4, starting at 16:9. Draft at the starting length and resolution until the action looks right. Pick the ratio for where the clip will be posted, such as 9:16 for vertical feeds. Leave the "Generate Audio" switch on, as it starts, unless you plan to add your own soundtrack.
Start the run with "Generate"
Click "Generate" once your references, prompt and settings are in place. Each run returns one video, so to compare two takes of the same idea, run it again.
Download the finished clip
Watch the result before you download it. Check that each face, outfit and voice matches its reference and that lips follow the dialogue. If one detail drifted, make the prompt more specific about it and generate again. If it holds, download the video and keep the reference set and prompt together for the next clip in the series.
Reference images or a single image: which suits your shot?
How reference-to-video compares with the other way to animate your images in Vidu Q4.
| Criterion | Vidu Q4 | Alternatives | Best For |
|---|---|---|---|
| What you start from | Reference-to-video builds the clip from several image references, with optional audio references, and keeps the subjects in those images consistent. | Vidu Q4's image-to-video mode animates a single image based on your text prompt. | Pick reference-to-video when several characters, props or locations have to appear together in one clip, and image-to-video when one finished picture already has the composition you want to set in motion. |
Common Vidu Q4 reference problems and fixes
Fixes for the problems people hit most with uploads, sound, clip length and prompt length.
My reference image won't upload to Add Image.
Add Audio rejects my voice sample.
The video came back with no sound.
My scene needs more time than one clip allows.
The Prompt box won't take any more text.
Frequently Asked Questions
How many reference images and voice clips can I upload to Vidu Q4?
Add Image takes 1 to 15 images in JPG, JPEG, PNG or WEBP format, each up to 10 MB and at least 300×300 pixels. You need at least one before you can generate. Add Audio is optional and takes up to 3 MP3 files, each up to 25 MB and no longer than 12 seconds. Use the image slots for different roles, one per character, key prop, location or style frame, rather than near-identical shots of the same thing.
Can Vidu Q4 make a reference video from a text prompt alone?
No. Add Image is a required upload in this generator, so every run starts from your own pictures. The prompt then decides what those subjects do.
Do I need to upload audio to get a video with sound?
No. In this mode Vidu Q4 outputs audio and video together, so with the Generate Audio switch on, a run with nothing in Add Audio still comes back with sound. Voice samples are for controlling how a character sounds. Turn Generate Audio off when you want a silent clip to add your own soundtrack to.
Why does one Vidu Q4 generation cost more than another?
The credit cost of each run is calculated from the options you choose and shown above the Generate button once you are signed in. It changes with Video Duration and Resolution, so check the figure again after you adjust either one.
How do I keep the same character across several Vidu Q4 clips?
Use the same reference images in every run, plus the same voice sample if the character speaks. Vidu Q4 builds each video around the subjects in the images you supply, and each run produces one clip. Use the same words to name the character in every prompt, so only the action and camera change between clips.