The storage problem
Cartridge-based game storage in the early 1990s was constrained by manufacturing economics. A 4-megabyte cartridge for the Super Nintendo cost approximately $10 to manufacture before any other component. An 8-megabyte cartridge cost proportionally more. The largest SNES cartridges — games like Chrono Trigger with its SPC700 sample data, or Star Ocean with its DSP chip — reached 48 megabits (6 megabytes) and required enhancement chips to manage the data efficiently. Every byte of cartridge storage was expensive, which meant every game had to make hard decisions about what to include.
CD-ROM technology, developed by Sony and Philips and commercialised in 1982 as the format for music distribution, stored 650 megabytes — approximately 80 to 150 times the storage of a typical cartridge. The manufacturing cost per disc was under $2. The access time was slow — reading data from a spinning disc was dramatically slower than reading from ROM cartridge chips — but for data that didn't need to be accessed in real time, the capacity advantage was transformative. A full orchestra recording of a game's score required more storage than an entire cartridge. A 30-second video required more than most cartridge games' total data. On CD-ROM, both were trivial.
The Turbografx-CD add-on (1989) was the first consumer game hardware to use CD-ROM storage. The PC Engine CD in Japan enabled games with CD audio soundtracks and voice acting that the cartridge versions of the same games couldn't include. The Sega CD (1992) brought CD-ROM to the Genesis platform. The 3DO (1993) used CD-ROM exclusively. The PlayStation and Saturn (1994–1995) standardised CD-ROM as the primary game distribution format. By 1996, cartridge-based gaming was a Nintendo holdout, and the industry had largely completed the transition.
What the capacity enabled
The most immediately audible change was music. Cartridge games used synthesised or sample-based music generated by dedicated audio chips because storing actual recordings was impossible within cartridge constraints. CD-ROM allowed game developers to include CD-quality audio tracks that played directly from the disc — standard music recordings, indistinguishable in fidelity from a music CD, because they were recorded on the same format.
The practical effect was that games could have orchestral scores without synthesising orchestras from samples. Wing Commander III (1994) — a game that spent $4 million on its budget, unprecedented at the time — included performances by professional actors including Mark Hamill, John Rhys-Davies, and Malcolm McDowell, filmed on professional sets with professional cameras, edited and compressed for disc playback. The game had more in common with a low-budget film production than with any previous video game. The acting was uneven and the compression artefacts visible, but the ambition was new: games as productions with physical sets, union actors, and director's cuts.
Voice acting became possible and then standard. Where cartridge RPGs had given characters text boxes with no vocal counterpart, CD-ROM RPGs could give each character a speaking voice. The first full CD audio soundtrack in a Final Fantasy game was Final Fantasy VII (1997) on PlayStation — not orchestral, because the development team chose synthesised music that would loop cleanly, but including voice acting in its cinematic cutscenes. Players who had read character dialogue in text for a decade heard characters speak for the first time. The transition was significant enough that localisation quality — the quality of English voice acting for Japanese games — became a commercial consideration rather than an afterthought.
The FMV disaster
Full-motion video games — games in which the primary interactive content was filmed video footage — emerged as a genre when CD-ROM made large video files storable and when game developers confused storage capacity with design quality. The reasoning was superficially plausible: if games could include Hollywood-quality video, games could be cinematic experiences. If cinematic experiences were valuable, FMV games would be valuable.
The reasoning was wrong because it misunderstood why cinema is valuable. Cinema is valuable because of what happens within the frame — performance, composition, editing, narrative — not because the frame contains moving pictures. Storing moving pictures on a disc did not give game developers the skills to fill them with content worth watching. The result was a wave of games — Night Trap (1992), Sewer Shark (1992), Phantasmagoria (1995), the Tex Murphy series — that ranged from competently produced to embarrassingly amateurish, and that were universally games in which the interactive element felt subordinate to the video that had prompted the investment.
Night Trap is the FMV game best remembered, largely because it became exhibit A in Senator Joseph Lieberman's 1993 congressional hearings on video game violence. The game's premise — monitoring security cameras to protect sorority girls from "augers" (vampire-like creatures) — was genuinely odd, the acting was poor, and the interactive component consisted of triggering traps at the right moment. It was neither effectively violent nor effectively interactive. Its role in the ESRB's creation was disproportionate to its actual content. The FMV genre collapsed by 1996, not because of congressional attention but because players had discovered that video stored on a disc was not, by itself, interesting.
What changed permanently
The CD-ROM revolution produced changes that outlasted the FMV disaster and the initial period of adjustment. Loading times became a fundamental aspect of game design — something that cartridge games never had to accommodate and that CD-ROM games had to design around. Games that loaded levels from disc had to either provide brief loading screens or preload assets into RAM before players needed them. The engineering challenge of managing disc access while maintaining play continuity shaped how game worlds were structured: the level-based design that predated CD-ROM was partly a consequence of cartridge storage constraints, but it also suited CD-ROM access patterns better than seamless open worlds that required constant background loading.
Nintendo's decision to ship the Nintendo 64 on cartridge in 1996 — explicitly chosen over CD-ROM to eliminate loading times and to maintain cartridge-level data access speeds — was commercially damaging but not without technical merit. The N64's cartridge load times were imperceptible; PlayStation games that used CD-ROM extensively had loading pauses that players experienced as interruptions. The trade-off was storage capacity: an N64 cartridge held 64 megabytes at most, while a PlayStation game could spread across multiple CDs totalling hundreds of megabytes. Square's decision to make Final Fantasy VII on PlayStation rather than N64 was driven partly by this storage difference. The story the game wanted to tell required more data than cartridge could provide.
The CD-ROM also ended the cartridge's role as a physical piracy barrier. Cartridges required specialised manufacturing equipment to reproduce. CDs could be copied by any computer with a CD burner, and CD burners became consumer products by 1997. The software piracy that cartridge manufacturing had made difficult became trivially easy. Sega's Saturn had no effective copy protection. PlayStation's copy protection could be defeated with a small hardware modification sold openly. The era of physical media as piracy protection ended with cartridges. Every subsequent development in game distribution — the shift to online distribution, DRM systems, always-online requirements — was a response to the piracy possibility that CD-ROM had opened.