# Text-to-video, image-to-video or reference-to-video: which should you use for a film?

**Short answer:** For a film with recurring characters, use reference-to-video or image-to-video, not plain text-to-video. Text alone lets the model reinvent faces every shot; a starting frame or reference images give it the same evidence each time.

URL: https://arc.katechon.technology/answers/text-to-video-vs-image-to-video-vs-reference-to-video
Updated: 2026-10-18

## The three modes

Generation modes and what they are good for:

| Mode | Input | Strength | Weakness |
| --- | --- | --- | --- |
| Text-to-video | A prompt | Fast, flexible, good for one-off shots | Faces and places change between clips |
| Image-to-video | A first frame (sometimes a last frame) plus a prompt | The shot starts exactly where you approved it | You need a good frame for every shot |
| Reference-to-video | Reference images of characters, places or props plus a prompt | Identity holds across many shots without a frame per shot | Composition is less controlled than a fixed first frame |

ByteDance says Seedance 2.0 supports text, image, audio and video inputs, and images, audio and video as references. (Source: ByteDance Seed, Seedance 2.0, https://seed.bytedance.com/en/seedance2_0, checked 2026-10-18)

## What we use

- **Reference-video mode** rendered most of our films, including [The Wolf Mother](https://arc.katechon.technology/films/the-wolf-mother), [The Fall of Kalak](https://arc.katechon.technology/films/the-fall-of-kalak) and [Blood Sun: The Fall of Kratos](https://arc.katechon.technology/films/blood-sun): each shot attaches the character and location sheets it needs.
- **Storyboard mode with grid storyboards** rendered [BROODFALL](https://arc.katechon.technology/films/broodfall): each shot starts from an approved board cell, and chained cells make the last frame of one shot the first frame of the next.
- **Plain text-to-video** we use only for establishing shots without recurring characters.

## How to choose

1. Recurring characters across many shots: reference-to-video.
2. Shots that must match an exact composition or continue from the previous shot: image-to-video from a storyboard frame.
3. Product shots where packaging must be exact: image-to-video or references from real product photos.
4. A single atmospheric clip: text-to-video is enough.

## Cost

The mode does not change the per-second video price much; the extra cost is the images. In Arc a Seedream 5.0 Lite image is 13 credits, a few percent of a 10-second Seedance 2.0 Mini clip at 94 credits. Optional reference-image charges on video generation can add a little more.

## Where Katechon Arc fits

Arc projects run in reference-video mode or storyboard mode, so every shot draws on approved sheets or frames rather than text alone.

Arc does not animate an arbitrary photo you upload as a one-off clip in seconds; quick photo-to-video apps are simpler for that.

## People also ask

### Can I make an AI movie from a picture?

A single picture can start a clip with image-to-video, but a movie needs many shots; use the picture as a character or location reference and generate each shot from it.

## Related answers

- [How does text-to-video AI work?](https://arc.katechon.technology/answers/how-does-text-to-video-ai-work): A text-to-video model is trained on huge numbers of video clips paired with descriptions, and learns to turn random noise, step by step,…
- [How do you keep the same character across AI video clips?](https://arc.katechon.technology/answers/how-to-keep-the-same-character-in-ai-video): Give the model the same visual evidence every time: one approved reference image per character, attached to every shot that character is…
- [Can AI turn a storyboard into a video?](https://arc.katechon.technology/answers/can-ai-turn-a-storyboard-into-a-video): Yes. Each storyboard frame can be used as the first frame of an image-to-video generation, so the clip starts from the composition you…
- [How do you make an AI video from photos?](https://arc.katechon.technology/answers/how-to-make-an-ai-video-from-photos): Upload a photo to an image-to-video model and describe the motion you want; the photo becomes the first frame or a reference.
- [What is a grid storyboard in AI video?](https://arc.katechon.technology/answers/what-is-a-grid-storyboard): A grid storyboard is one generated image divided into cells, each cell the frame for one shot, so several consecutive shots are drawn…

## Sources

- ByteDance Seed, Seedance 2.0: [seed.bytedance.com/en/seedance2_0](https://seed.bytedance.com/en/seedance2_0), checked 2026-10-18