clock menu more-arrow no yes
A woman looks into a man’s eyes while holding his head in her hands Image: Konami

Filed under:

Full-motion video has found the perfect home: horror games

How pre-rendered scenes create fear, and uncertainty, in our favorite games

It would be easy to think that full-motion video is obsolete.

Full-motion video, often shortened to FMV, is a previously rendered or recorded scene that is spliced into a game, rather than rendered by the game’s engine in real-time. Using full-motion video once allowed developers to insert visuals into a game far beyond the scope of their graphics engine. For example, when Final Fantasy 7 needed to depict the massive scale of Midgar, it didn’t use 1997’s crude in-game graphics — it used pre-rendered FMV.

Of course, this explanation paints over huge swathes of the format’s history. Arcade games like Dragon’s Lair built semi-responsive worlds out of hand-animated full-motion video, and an undercurrent of FMV titles like Erica or Sam Barlow’s Telling Lies proved that pre-recorded video content can still be gamified in a compelling way.

But when today’s games can approach levels of graphical quality previously exclusive to pre-rendered scenes played like a movie between interactive segments, is there any storytelling advantage to the dramatic shifts in aesthetic FMVs were once known for?

2019’s Devotion gave us an irresistible answer: yes.

And especially in horror games.

Changing the mood by changing the medium

Devotion’s story largely centers around a young girl, Mei Shin. At the encouragement of her family, Mei Shin enters a televised singing competition, competing against other children for a shot at fame. We return to the competition over and over throughout the game; virtually all of the story’s tragic events can be traced directly back to this situation. And to drive home how narratively important this performance is, Devotion makes it a complete visual aberration.

An older TV shows an image of a person singing Image: Red Candle Games

While everything else in the game is built within a game engine called Unity, the singing contest is, well, real. It’s a blurry, VHS-style recording of real actors on real sets with a real girl singing her heart out. It’s a startling contrast, one that Devotion plays up by placing the FMV on a television built with in-game graphics. It feels like the game is daring you to question its reality. Is what’s on the TV even in the same reality as the sparse furnishings of everyday life in this apartment in Taipei?

This essay on full-motion video’s new home in horror games was written as a companion piece to the video above, also created by Jacob Geller.


The aesthetic disparity takes on a sinister tone in the context of the game’s story. Mei Shin confesses that she doesn’t really want to perform on TV. She just wants to make her father, Feng Yu, happy, and she knows how much this means to him.

Even when Mei Shin falls sick with symptoms clearly resulting from her anxiety, Feng Yu refuses to acknowledge that taking her out of the spotlight might be the right thing to do. Through the use of full-motion video, we begin to understand what’s motivating him. For Feng Yu, the televised performance is more real than his day-to-day life. As such, he’s willing to make any sacrifice to keep his daughter in the limelight.

Devotion’s fragmenting of reality reminds me of another game that used full-motion video to call the reality of the in-game world into question almost two decades ago: Silent Hill 2.

A disheveled man looks into a dark mirror Image: Konami

One of the most indelible images of the PS2 era is the very first shot of Silent Hill 2. The protagonist, James, stares deep into a mirror. His eyes are dark pools, his hair is greasy, the mirror is covered in grime. For a second, the game is a breath away from reality. Of course, the lighting and texturing of this scene are far beyond what the PS2 could render in real-time. Silent Hill 2’s first shot is full-motion video, not anything created in the game’s engine.

The human face is a notoriously tricky subject to render in CGI. There’s a reason Pixar’s first films featured toys, bugs, fish, and monsters; few things make an audience more uncomfortable than an uncanny representation of our own form.

But Silent Hill 2’s character designer, Takayoshi Sato, plunged headlong into that uncanny valley. The game’s videos feature close-up shots of realistically-proportioned characters who do little more than have a quiet conversation with each other. Often due to their technical limitations, they are deeply unsettling. And for the horror of this game, they work.

For all its monsters and darkness, Silent Hill 2 is a story of introspective horror. Unraveling the tangled knot of James’ guilt and mourning takes the entire length of the game. All the while, just like Devotion, radical changes in aesthetic challenge our perception of the character. Is the “true” James the low-resolution model we control around the streets of Silent Hill? Is he the uncanny, shadowy figure present in the game’s pre-rendered sequences?

Or can we only know James from the secrets we learn from a lost VHS tape, a tape that reveals the darkness boiling just below the surface? It’s no coincidence that we begin with James staring into a filthy mirror; the game refuses to give us a clear picture of him.

Silent Hill 2 frequently blindsides the player with these visual shifts as well. You could be deep in the bowels of a hospital, or stumbling through a historical society, open a door, and immediately find yourself in pre-rendered video. More than once, I actually jumped as the game started a pre-rendered scene. Nothing, not even the basic look of the world, can be relied on for long. Reality, and how it’s presented to the player, is proven to be a fluid thing, and the game reminds you of that just as soon as you begin to forget.

While the obvious purpose for full-motion video is to pull players deeper into a game, Devotion and Silent Hill 2 do just the opposite: They force us to confront the artifice of what we’re playing. The protagonists of these games each hold a deep insecurity. Their reality is tearing apart at the seams, and the visual contrast between the in-game graphics and the FMVs make this conflict literal.

Decades ago, game developers might have wished that the entire game could match their FMVs; indeed, Final Fantasy 7’s 2020 remake looks light-years beyond the pre-rendered cutscenes of the original. But ironically, as graphics improve, my appreciation for the inconsistency of FMVs has only grown. In a medium that often falls prey to repetition, they are a truly unpredictable element. They challenge the way we perceive these virtual environments.

Even in the AAA space, FMVs are being used to highlight a space’s instability. In Control, recordings of live actors fill in the foggy backstory of The Oldest House, projected on walls or blaring from old TVs. If you’re lucky, you might even catch an episode of “Threshold Kids,” a completely bizarre children’s program with real(?), filmed puppets. As the characters of the FMV show cry about their dead mother or visit the friendly “uncle mr. bones,” it only drives home the inscrutability of Control’s many mysteries.

In horror, these visual digressions can make the entire fabric of the world feel unstable. Consistency is safe. Radical jumps in look and aesthetic? That’s dangerous.

After all, how can we feel safe if we can’t even trust our own eyes?