that's not what diminishing returns means you retard it's about hogging more processing power for less graphical improvement like ray tracing being a resource hog for reflections that devs could do
Yes, and it's actually one of the cleaner real-world illustrations of the concept.
The economic sense of the term applies well
"Diminishing returns" properly describes a situation where each additional unit of input produces a smaller increment of output than the previous unit. Applied to graphics:
- Input: computational power, development budget, artist-hours, engine sophistication
- Output: perceived improvement in "realness" by an average player
The jump from 2D sprites (Pac-Man era) to early 3D polygons (PS1/N64 era) was a categorical leap — characters went from flat abstractions to objects occupying three-dimensional space. Then the jump from PS1 to PS2/Xbox-era graphics was still huge: textures, lighting, and character models became recognizably human rather than blocky approximations.
But compare that to more recent jumps — say, PS4 to PS5, or even mid-2010s to today. The
technical achievement per generation has arguably grown (ray tracing, photogrammetry, physically-based rendering are enormously sophisticated), yet the average player's subjective sense of "wow, that looks real" grows much less per generation than it used to. You need much bigger technical investments now to move the needle on perceived realism by the same amount you could once move it with comparatively modest advances.
Why this happens — the underlying mechanism
A few factors explain why this is a legitimate diminishing-returns case rather than just a vague feeling:
- Perceptual ceiling effects: Human visual perception has limits on discriminating fidelity. Once a game clears a certain threshold of detail, lighting accuracy, and motion smoothness, further improvements become harder to consciously notice, even though they're measurable on a technical level.
- The uncanny valley plateau: As character models approach photorealism, small remaining imperfections (skin texture, eye movement, micro-expressions) become more noticeable and jarring relative to how close the model already is, requiring exponentially more effort to fix for comparatively small gains in acceptance.
- Baseline creep: Each generation resets the "impressive" baseline. Average players calibrate expectations to current-gen output, so a technically massive leap barely registers as a leap subjectively, because the starting point was already fairly convincing.
- Non-graphical factors capping perceived realism: Things like animation quality, physics, AI behavior, and audio increasingly gate how "real" something feels, rather than pixel/polygon fidelity alone. So pouring more resources purely into graphical rendering hits a ceiling imposed by other systems.
A caveat worth noting
The comparison is with respect to
player-perceived realism, not technical capability, artistic quality, or the value of the investment overall. Studios don't get diminishing returns on things like art direction, world density, or performance stability from the same investment — those can still scale well. Also, stylized/non-photorealistic games sidestep this curve entirely since they aren't chasing realism in the first place.
So the term applies specifically and appropriately to the
realism-chasing branch of graphical development, less so to graphics-as-craft in general.
