Hybrid filmmaking with AI: we filmed Sara Camacho in a living room and took her to an alpine valley

What is hybrid filmmaking?

Hybrid filmmaking is a production method that combines a real, filmed performance with AI-generated locations, light and weather. The actor, the gestures and the camera move come from the shoot; the world around them is built with AI afterwards, using the real footage as the primary reference. It sits between live-action production and fully AI-generated video, and it keeps the one thing audiences can always tell apart: a human performance.

In hybrid filmmaking the performance is real and the world around it is built with AI. We did not generate an actress: we filmed one. Her eyes, her hands, her breathing and the camera move all come from the shoot. The AI changes the place, the light and the weather, and leaves everything else alone.

The text of Pedazos del alma is intimate. It is about growing old with a well-used heart, and about the pieces of your soul that people take with them when they leave. A voice like that needs a face that truly acts, not one that poses. That is how ZerezoFilms AI Studio works: direction and performance first, the tool second.

This is the first of two parts. It covers the shoot: how it was filmed and how that footage became a valley. Part 2 covers what was never filmed: the landscapes, the aging, the forest ending, the score and the subtitles.

The short at a glance

What was filmed and what was generated

Hybrid filmmaking vs. fully AI-generated video vs. virtual production

The three get lumped together, and they are not the same thing. The difference is what is real in front of the camera and when the world gets built:

Pedazos del alma is hybrid filmmaking from start to finish: 23 of its 28 shots start from Sara Camacho's performance. Our other short film, Heartbreakers, was made 100% with AI and no shoot at all. Comparing the two is the clearest way to see what each technique brings to a film.

The shoot: a living room, a grey backdrop and three tricks

We filmed 16 handheld takes in 4K at 24 fps, in a living room. Behind Sara there is only a grey backdrop. There are three practical effects: a floor fan for the wind at the cliff edge, cushions for the fall, and a glass tray of blue kinetic sand for the ashes.

The key light comes from Sara's right in every take. That on-set choice became a rule for everything after it: every reference image and every prompt keeps the light on that side.

How hybrid filmmaking works: from the shoot to the location

• 01Frame-accurate cuts. We split every take to the frame and grouped them by location: cliff, rocks, creek and meadow.
• 02The world and the palette. We designed the valley in Midjourney and locked a cold palette of blues, greys and stone.
• 03References built on real frames. In Seedream we placed Sara in each location starting from real frames of her performance, so the look never changed her face.
• 04Video-to-video. Each block went through Seedance with the real take as the primary reference. The images only add the place, the light and the color.

PRIMARY REFERENCE — @video1 is the main reference for the actress and for every movement. Her face, identity, features, skin, body and proportions, her hair, her performance, her gestures, her timing and the camera move all come from @video1 first. The reference images only add the place, the light and the grade. If an image and @video1 disagree about who she is or how she moves, @video1 wins.

Hybrid filmmaking before and after, shot by shot

Each pair shows the same moment: on the left, the original take in the living room; on the right, the shot as it appears in the film. The pose, the gaze and the framing are Sara's. This is what hybrid filmmaking looks like in practice.

The cliff

The creek and the ashes

The ashes started as blue kinetic sand in a glass tray. In the film they are ash that the water carries downstream, toward the sea. The tray and the sand are removed in every frame; the hands, the rings and the rhythm of the gesture stay.

REMOVE FIRST — @video1 contains two props that must NOT appear in the result: a clear rectangular glass baking tray with thick rounded walls and a rim, and bright pale-blue clumpy play sand inside it. Remove both completely, in every frame. Where the tray stands there is only the creek bed: wet dark river stones under clear shallow flowing water.

The rocks and the meadow

The fall

The wind was a floor fan and the fall landed on cushions. In the film the fan is gone, the wind moves her hair and skirt, and the drop opens on the side away from the rock.

The rules of hybrid AI filmmaking

• 01The real take wins. If a reference image and the video disagree about who Sara is or how she moves, the video wins.
• 02Images set the look, not the framing. When the model copied the photo's framing, the shot became an animated still. The framing always comes from the video.
• 03The camera is the shoot's camera. Asking for a locked camera drained the shot; describing the real move kept it alive.
• 04The landscape has to move. Unless the prompt asks for it, the river and clouds stay frozen, because the grey backdrop never moved.
• 05Remove the props first. The tray and the sand survived the transformation until the prompt opened by removing them.
• 06One light direction. The key light comes from Sara's right in every take, and every reference keeps that side.

Coming in part 2: the five shots that were never filmed, how we aged Sara on her own face, how she turns to ash in the forest without looking burned, and a score built around a silence and a heartbeat.

Hybrid filmmaking at ZerezoFilms AI Studio, Mexico City

Hybrid filmmaking is how ZerezoFilms AI Studio works: the AI production arm of ZerezoFilms, a production company in Mexico City. We use it for commercials —like the Zerezo Bites commercial or the Zerezo Coffee spot, with a real actor at the center— and for auteur cinema, like Pedazos del alma, directed by Rubén Zerecero. Direction and performance come first; AI builds the world around them.

When hybrid filmmaking makes sense for brands and agencies

• 01Your talent, anywhere. The actor, spokesperson or brand ambassador is filmed on a stage or in a room, and the campaign takes them to the mountain, beach or city the idea calls for.
• 02Impossible or remote locations. A cliff edge, a glacier or a street in another country, without moving the crew.
• 03One performance, many worlds. The same take can live in different settings, one per campaign or market.
• 04The emotion stays human. The voice, the eyes and the gestures belong to a person; the AI does not invent them.