What Is an FMV Game

FMV stands for full-motion video. The term settled in the early 1990s, when CD-ROM drives and new consoles first made it practical to store and play back digital video with live actors at scale. Unlike sprites or 3D models, FMV frames are not computed by the processor during play — they are replayed from a recording. That is why FMV of that era looked “like cinema” next to crude triangle-based 3D: image quality was limited not by the console’s real-time horsepower but by disc capacity and codec.

The essence of the format is the role of the video. In an FMV game video is not a reward for clearing a level — it is the game’s material: story, actors, and editing are recorded in advance, and interactivity is layered on top. This is what separates FMV from a cut-scene: a cut-scene illustrates the game, while in FMV the video is the game. On how the cut-scene became a form of its own, see my article on the evolution of cut scenes.

The technical scheme is simple: a video player plays clips, and a logic layer decides which clip comes next. The whole variety of the format reduces to three types of interaction:

• Branching choices — the player picks a path, and the graph jumps to the matching clip; the Bandersnatch scheme, and Late Shift.
• Quick-time events — inputs in rhythm with the video: the reaction decides success, failure, or a rewind.
• Exploration — the player assembles the sequence from a clip archive (“Her Story”, “Immortality”), inspects a frame for hotspots, or navigates a space.

On the technical side the critical parameter is clip-to-clip transition speed: from “player picked a branch” to the first frame of the next clip must pass a fraction of a second, because in FMV transitions never stop. That is why codec choice (H.264, VP8/WebM, HAP) is a design decision, not a technical footnote; how this plays out in my own prototype is described in the devlog on video, hotspots, and branching.

In the 1990s the same scheme answered to other constraints: the Cinepak codec and CD-ROM capacity dictated resolution, frame rate, and runtime. Those constraints became formative — the “digital patina” of heavily compressed video now reads as the era’s aesthetic. I break down that mechanism across twelve 3DO FMV cases in my study of 3DO audiovisual strategies: three archetypal strategies — action reflexivity, interactive drama, spatial immersion — and how Cinepak compromises turned editing into an interface event.

The history of FMV is not a straight line of progress but three separate waves: film-based arcades in the 1970s, the CD-ROM boom of the 1990s, and the revival that has been running since the mid-2010s.

Film arcades: Wild Gunman, 1974

The idea of “film as a game” is older than home consoles. In 1974 Nintendo released the Wild Gunman arcade cabinet: a film projector showed a live actor — a gunslinger — and the player had to draw faster with a light gun. The video was not digital yet, but the FMV principle was: pre-recorded footage responding to the player’s actions. In the 1980s laserdiscs brought animated full-motion video to arcades — Don Bluth’s Dragon’s Lair (1983) and Sega’s Astron Belt — and the industry got used to the idea that a game could consist of video.

CD-ROM and 3DO: FMV goes mainstream

The mass era of FMV began with CD-ROM: 650 megabytes, for the first time, put an hour of digital video inside a game box. Night Trap and Sewer Shark on the Sega CD, “I’m Your Man” on 3DO, and dozens of other titles used live actors and branching footage. FMV became a marketing trump card of the early-1990s console war: “real actors in a game” looked like the industry’s future, and magazines benchmarked consoles by their video playback. What artistic compromises that era accepted — and which of them remain in the format’s language — is the subject of my analysis of twelve 3DO cases.

Why the format declined

By the late 1990s FMV had lost its position for several reasons at once. Real-time 3D (PlayStation, Nintendo 64) freed the image from pre-recorded frames: worlds could now be computed and respond to the player freely, while FMV remained a set of pre-shot branches. Production was expensive — every branch had to be shot, edited, and compressed — and the gameplay was fairly criticized as shallow: “watch and press in time”. The industry drew its conclusions, and for roughly fifteen years FMV survived as a marginal experiment.

Revival: Her Story → Immortality → the 2020s wave

The revival started with design, not technology. Sam Barlow’s Her Story (2015) turned an archive of police interview clips into a database narrative: the player typed search queries and assembled the story alone. Late Shift (2016) brought branching live-action video back in real time. Bandersnatch (2018) introduced a mass audience to the branching-film scheme. Immortality (2022) pushed the “archive” idea to its limit: reconstructing a director’s lost filmography from fragments of shot scenes. Today FMV is a deliberate aesthetic and design choice, not an exploitation of hardware capability — and that is exactly what separates the current wave from the 1990s.

In 2026 the format is demonstrably alive — in the West and in Asia. Tenebris Somnia (Saibot Studios / New Blood Interactive) launches on PS5 on October 16, 2026: a survival horror that combines pixel-art adventure gameplay with live-action FMV scenes filmed by an Argentine crew with practical effects. It is not an isolated case but part of a wave:

• Road to Empress (China): the second installment of the series about Empress Wu Zetian’s path had its global release in June 2026 and passed one million copies in five days, with over 85% positive reviews on Steam.
• Northeast Past (June 2026) — a crime FMV about the gang wars of 1980s–90s northeastern China; Shao Song (August 2026) — a historical FMV about time travel to the Southern Song, already with turn-based battles.
• The wave was paved by The Invisible Guardian (2019), Love Is All Around (2023, 1.9 million copies in three months), and Revenge on Gold Diggers (2025).

The economics remain harsh: an industry survey cited by The World of Chinese estimates that around 90% of FMV projects lose money, and a June 2026 report by analysts East-West Entertainment projects that AI tools could cut live-action production costs by millions of yuan and halve production timelines. The wave is simultaneously expanding on cheaper production and risking saturation — a classic situation for a young genre. One boundary-case answer to that challenge is hybridization: in my FMV prototype FMV-01 shot cutscenes sit next to still frames with hotspots — cheaper to produce and faster to switch between.

The two terms are often confused, but they describe different things. FMV is about material: live-action video used in a game. Interactive cinema is about form: works on the border between film and game, where the cinematic experience comes first. An FMV game can be pure interactive cinema (Late Shift), a game with video inserts (1990s classics), or something in between (my FMV-01). The base definitions are in my pillar article on interactive cinema.

For the full separation framework — three criteria: the recipient’s experience, the type of mechanic, the priority of film language or gameplay — see my essay “Interactive Film or Video Game: Where Is the Line”. The historical comparison of FMV with interactive cinema on DVD is in my DVD-era article: it shows that “video plus menu” existed before game consoles, and the spectrum of hybrid forms predates the term FMV itself.