Motion blur all the way down (2022)

Efforts to render more realistic motion blur in graphics are pushing beyond simple film-camera shutter simulations toward models that better match human visual perception. Commenters examine how shutter angle, frame rate, display technology (CRT, LCD, OLED, HDR), and eye movements interact to produce or reduce artifacts like strobing, judder, and smearing. They also debate the value and implementation of motion blur in games versus films, arguing that poor or camera-centric blur often harms clarity, while physically based, per-object approaches and future eye-tracked techniques could significantly improve realism.

Article structure & torusphere concept

  • Several readers felt the article naturally splits in two:
    • A clear, useful conceptual intro to motion blur and shutter angle.
    • A dense, “crazy” deep dive into the specific torusphere shader implementation.
  • The second half is seen as valuable mainly to make the accompanying shader/code comprehensible.
  • The torusphere (“motion blur all the way down”) is noted as physically impossible: it’s essentially a loop made entirely of thickened motion blur, not actual “solid” geometry.
  • Some noticed that opacity is artificially boosted; otherwise the torusphere would keep getting fainter.

Perception, frame rate & display technology

  • Strong debate over how high refresh rates must be to appear “continuous” (claims range from ~80 Hz up to multiple kHz), with many citing cursor tests, LED strobing, and human afterimages.
  • Distinction is made between:
    • Display refresh limits (sample-and-hold vs strobing / low persistence).
    • Eye tracking and smooth pursuit vs saccades and afterimages.
  • CRTs and strobed backlights are praised for motion clarity; LCD/OLED sample-and-hold causes “tracking blur.”
  • Some argue even 240 Hz or 400 Hz is visibly discrete; others say beyond a certain point it’s “good enough.”

Motion blur models & shutter angle

  • Discussion of box shutters vs more realistic shutter curves; modern renderers support custom “shutter curves” for cinematic control.
  • Motion blur is framed as either simulating a camera shutter or the human visual system; some see the article as a welcome shift toward the latter.
  • Shutter angle is treated as an expressive tool in cinema (e.g., narrow for harsh/action, long for dreamy).

Games vs films: desirability of motion blur

  • Strong split:
    • Pro-blur: Necessary to avoid strobing at finite FPS; “true 3D” motion blur tied to object/camera motion can look natural and is standard in VFX.
    • Anti-blur: Game implementations often smear the whole screen on camera movement (“ego-motion blur”), mis-blur tracked objects, and feel distracting or nauseating. Many players turn it off.
  • Consensus that per-object, per-pixel blur and eye-tracking–aware methods would be better, but are rarely implemented.

Color space & implementation concerns

  • Some point out that the demos appear to operate in sRGB (nonlinear) instead of linear space, which can make motion blending look wrong and reduce the impact of the examples.
  • The author (in-thread) acknowledges this and plans to revisit the interactive figures.