(Long reply)
The number is based on the fastest smooth eye pursuit over the widest FOV over retina-resolution angular. That's found to be about ~10K elements/sec. So you need 1/10,000th that in the temporal timeline, in the sample and hold effect (strobeless). However, it can go beyond, especially if assisted by a head turn (allows faster eye tracking). In order to have motion same as static. Also, need to oversample that a bit too.
Also you need large geometrics (4-8x differences in framerate=Hz), GtG=0.0000 (not just merely less than refreshtime), fast motion, retina resolution, wide FOV, to see human visible differences.
240 vs 360 = useless. Try 120 vs 1000 vs 10,000; that's where the human-benefits visible differences are, much like the blur differences of camera shutter speed (e.g. 1/120sec blur vs 1/1000sec blur).
1000fps 1000Hz sample-hold still has 8 pixels of motion blur at 8000 pixels/sec on an 8K display.
Many armchair researchers (and even a few Ph.D's!) jump to conclusions, so I am just putting this out here...
There are 4 human-visibility thresholds involved.
Very approx magnitudes:
~10 - slideshows become motion
~100 - flicker fusion threshold
~1000 - blurs stop for 1080p 30-FOV
~10000 - stroboscopics stop for 16K 180-FOV
Error Margins: Assumes GtG=0 stutter=0 framepace=perfect, framerate=Hz). This fudges around depending on variables like resolution, motions.
Main Research Portal / cites in papers
blurbusters.com/area51
Stroboscopic Effect of Finite Frame Rates
blurbusters.com/stroboscopic…
1000 Hz Journey article
blurbusters.com/1000hz-journ…
Particularly, the Vicious Cycle Effect section
blurbusters.com/1000hz-journ…
Quest 2-3 is 0.3ms MPRT. That will require 1000/0.3 = 3333fps 3333Hz to say goodbye to strobing. VR strobes out of necessity because we don't have enough framerate & refreshrate necessary.
Refresh rate incrementalism such as 240 vs 360 is useless to humankind. Only barely visible on LCD, mainly of interest to esports.
Retina resolution of small screens would be low (e.g. phones), but big screens would be high (e.g. Las Vegas Sphere). You need a screen big enough that eyetracking that many pixels (e.g. 16000 pixels in 1 second) allows you to see differences between static images and moving images (e.g. blur differences, stroboscopic differences). The more time something is onscreen, the more time you have to detect something isn't equal to analog real life.
For example, several LCD VR headsets (Valve Index, Quest 2, Quest 3) has ~0.3ms pulsewidths, creating 0.3ms MPRT(0%->100%). That will require 1000/0.3 = 3333fps 3333Hz to say goodbye to strobing, getting only 1/3333sec motion blur. VR strobes out of necessity because we don't have enough framerate & refreshrate necessary.
Mainstream benefits is 60 vs 240 vs 1000 *and* GPU keeping up (e.g.
blurbusters.com/framegen lagles framegen) to get the necessary "wow" upgradefeels. That's not happening easily, obviously. But many are unaware of upgrade quality lurking.
Realistically, 1000 is probably the economical limit (for the current human generation), and framegen is capable of getting this far on current GPU tech. Also because Windows 11 has a 1000Hz limit (Windows 10 was limited to 500Hz).
Again: 240 vs 360 (especially in framerate<>Hz situations) is mostly useless for mainstream humankind benefits. Only a 1.5x difference throttled to 1.1x difference due to nonzero GtG and jitters (e.g. use of VSYNC OFF in esports is good for lag but not good for zeroing jitter. Even 70 microstutters/sec at 360Hz = blends to blur like fast vibrating music string, reducing MPRT slightly via high-frequency stutter blending to blur).
Stutter-to-Blur continuum (see 2nd UFO for 20 sec)
testufo.com/eyetracking#spee…
A lot of these perceptuals are in many different papers, but we'd love to see a more cohesive paper combining all the motion thresholds into a combined "What threshold Hz causes all temporal-perceptual weak links to disappear". I'm happy to collab on co-author / being cited / peer review.
Contact me if you want to collab (whether in professional, researcher, open source, or business capacity):
blurbusters.com/contact
Also, there's workarounds to avoid retina Hz, for strobed headset. You can even use eyetrackers to only motionblur (add GPU blur) during eye saccades, to prevent stroboscopic effect, as one band-aid example. Would not work for a multi-viewer screen.
[Bookmark this post if this post is important to your field of work]