HeyGen Video 1.0 Text-to-Video: short story scenes from a prompt
HeyGen Video 1.0 turns a written scene into a short clip in six frame shapes. How to write a small story, when to use enhancement, and its resolution ceiling.
A courier brings a huge bouquet to the wrong door, realises his mistake, and gives the woman one small flower anyway before riding off. That is one of the example prompts on the page of HeyGen Video 1.0 Text-to-Video, and it tells you more about the model than its one-line description does. It is built for small scenes with a turn in them: a setup, a moment, a reaction. Not a mood board, not a product spin. A story that fits in a few seconds.
The settings, briefly
There is a prompt field with room for a long description, a duration from five to fifteen seconds (five by default), a resolution of 480p or 768p (768p by default), and six frame shapes: 21:9, 16:9, 4:3, 1:1, 3:4 and 9:16, with 16:9 as the default. Prompt enhancement and a seed sit under the advanced settings. That is the whole panel, which is part of the appeal: almost everything that matters is in the words.
Steal the shape of the examples
The examples are written like a short screenplay. One paragraph sets up the place and the people. The next ones are beats, each a single thing that happens. Then a paragraph for the camera, and a last line for the tone. If you write your own scenes the same way, you give the model the order of events, which is the thing short video gets wrong most often when everything is packed into one sentence.
An elderly man sits alone on a park bench feeding pigeons on a cold autumn morning.
A small girl in a yellow raincoat runs up and sits at the other end of the bench. He hesitates, then pours half his bag of seeds into her hands.
The pigeons gather around both of them. She laughs.
The camera starts wide across the empty park, then moves slowly into a two-shot on the bench.
Warm, gentle, quietly funny. Soft overcast light, falling leaves.A chef in a tiny noodle shop flips a wok full of noodles high into the air at closing time.
The noodles land perfectly in a bowl on the counter. The only customer, a tired office worker, slowly starts to clap.
Handheld camera, close on the wok, then a quick pan to the customer's face.
Playful, warm tungsten light, steam in the air.A lighthouse keeper climbs the spiral stairs at night with an oil lamp. At the top, she looks out and sees a small boat struggling in the waves. She turns the great lamp toward it.
The camera follows her up the stairs from below, then holds on her silhouette against the beam sweeping across the sea.
Cinematic, stormy, blue night with warm lamplight.Pick the frame for where it will play
Six aspect ratios is a lot for a model of this kind, and it is worth using them on purpose. 9:16 for a story or a short on a phone, 1:1 or 4:3 for a feed, 16:9 for a page or a player, 21:9 when the landscape is the point. Choose before you write: a tall frame wants the action stacked vertically, close on one person; a wide frame wants two people at either side of it. The aspect ratio guide goes deeper.
Where it stops: resolution
The honest limitation is the ceiling of 768p. That is fine for a phone and for a preview, and soft for a large screen. If you need more, there is a neat route through its sibling: make the opening frame as a still with an image model, then animate it with HeyGen Video 1.0 Image-to-Video, which goes up to 2K and keeps your picture as the first frame. You also get more control that way, because you can approve the look before you pay for motion.
Draft at 480p and five seconds while the story is still changing; the price grows with seconds and resolution and shows on the Generate button. Keep the seed of a take you like, so a longer or sharper rerun stays close to it. The text-to-video prompt guide has more on writing scenes, and the video model comparison on choosing between models.
Open the studio, choose the model on its page and write your first scene.
Questions
What aspect ratios does HeyGen Video 1.0 Text-to-Video support?
Six: 21:9, 16:9, 4:3, 1:1, 3:4 and 9:16. The default is 16:9.
What is the highest resolution?
768p. For a sharper result, make a still first and animate it with HeyGen Video 1.0 Image-to-Video, which goes up to 2K.
Should I turn on prompt enhancement?
Only for a short, one-line idea. When you have written the scene in detail, leave it off so the model follows your words.
