The great Human Eye fps debate

[NickelPlate]

So there's another debate in a forum I frequent online that the human eye can't detect any difference between 24-30 fps and more. These claims are coming from individuals with "years in the video field" saying the usual things like "30fps for a game like this is fine". Whatever that means.

I personally know this to be untrue as I can clearly see a difference below 60fps and it just gets worse the lower I go. That seems to be the point for me that I don't notice any difference above it though, at least as far as fluidity of motion is concerned.

Back in the days of CRTs though 60Hz refresh rates gave me headaches and I ran my monitor at 75Hz and above. Others claim to be able to detect more than 60fps which may be entirely possible and I'm not going to argue with what they see.

Have there ever been any scientific studies from credible sources that have concrete data suggesting what the human eye is really capable of? I'd like to point out scientific fact to these guys. So far searching just usually comes up with posts in game and video forums from people who supposedly know offering nothing but anecdotal evidence or their own personal observations but that still isn't evidence in my book.

 

[Greenlepricon]

Here's an article from extremetech about it. I'm sure there are actual scientific studies on it, and would check Google scholar if you want an actual medical paper. I can check later on but (weirdly) my job in a lab isn't set up for me to look through them on this computer.

 

[motsm]

If they check out examples like the below, and still deny it, then they have a perception problem, a browser issue, or are just lying to themselves.

www.30vs60fps.com/
www.30vs60.com/mirrorsedge.php
www.testufo.com/#test=framerates-text&pps=360

I find this to be the most convincing test, but unfortunately you can't link presets, and the default is setup more for cameras than games. https://frames-per-second.appspot.com/ If you turn off the motion blur however, and switch it to 30, 60, and 120, I think it's a great demonstration. Personally, I can see differences above 60 in this test rather easily.

 

[CakeMonster]

Yeah, not a debate. Its been settled long ago, and whatever limit there is for perception is way above 60fps.

 

[NickelPlate]

If they check out examples like the below, and still deny it, then they have a perception problem, a browser issue, or are just lying to themselves.

www.30vs60fps.com/
www.30vs60.com/mirrorsedge.php
www.testufo.com/#test=framerates-text&pps=360

I find this to be the most convincing test, but unfortunately you can't link presets, and the default is setup more for cameras than games. https://frames-per-second.appspot.com/ If you turn off the motion blur however, and switch it to 30, 60, and 120, I think it's a great demonstration. Personally, I can see differences above 60 in this test rather easily.

Fantastic links. Thank you for these!

 

[soccerballtux]

CRTs different due to flicker, but yes can detect 60 for certain, probably up to 120, and detect flickers of an image at 1/500th of a second. Navy did testing on that don't recall what for

 

[Despoiler]

There isn't really a debate on this. It's only undereducated vs educated. I'm not claiming to be an expert. A couple of things I've gleaned.

-The first thing is to not confuse analog film with computer rendered frames. They are not the same. 24/30 fps for film is adequate due to motion blur. 18 fps is the lower bound. For computer rendered frames ie no motion blur the minimum bound for all people to perceive smooth motion is 72fps. 60fps covers most people.
-Fighter pilots have been tested and they can perceive a frame of a plane in the 200-300fps area.

 

[BSim500]

Fantastic links. Thank you for these!

Here's another one:-
https://frames-per-second.appspot.com/

It really isn't a "debate" at all when you can literally see it with your own eyes. The only reason this cr*p keeps popping up is because 30fps is all consoles can manage, and some console owners feel to need to "pretend away" what they're being deprived of, out of denial.

As others have said, games are not movies. Motion blur hides a lot but also reduces sharpness. Take the motion blur away and the stutter can look awful. The only reason 24fps was chosen was as a middle ground compromise between the 22-26fps variable frame rates silent movies were commonly shown at by early to mid 1920s's. When "talkies" were introduced in 1926, a fixed fps rate was required (to maintain a static audio pitch). 24fps film wasn't chosen at all based on human perception / "cinematic look" (and neither were 25fps PAL or 30fps NTSC).

 

[Red Hawk]

I don't think there's much of a debate. Any video game player will tell you there's a perceivable difference between 30 FPS and 60 FPS. It doesn't just matter for how smooth animations look, but for how smooth gameplay feels.

 

[skipsneeky2]

Look vs feel is something to consider too,games like CS:GO and twitch shooters feel miles better the higher the framerate is.Doesn't matter if your staying above 60fps and averaging 80,it just feels better hitting 150+.

I am sure if i was a 120/144hz user the difference would be even more dramatic as well.I have tried the fps limit at 60,vsync and none of the cap tricks give even remotely the same feel as the fps doing its own thing.

 

[LoveMachine]

Eye doctor here. I'm working at the moment so I can't provide any specifics, but the long and the short of it is that flicker perception is highly dependent on other visual factors, e.g. the brightness of the screen, contrasting brightness of the surrounding area, peripheral vs. central retinal exposure. Google "critical flicker frequency" to get some examples.

 

[LoveMachine]

Whoops, forgot the take home argument. Unless you isolate and replicate the exact environment, flicker is subjective. Watching a clip will be perceived differently by different people in different environments, so there isn't an absolute "right" answer.

 

[Dankk]

There isn't really a debate on this. It's only undereducated vs educated.

Pretty much this.

This issue has been discussed to death. Anyone who thinks that the human eye can't detect more than 30fps (or 60fps, for that matter), is very misinformed.

And yes, the only reason why 24fps hollywood movies still look good is because of motion blur from the camera. The same idea doesn't really apply to interactive media so well, particularly when it requires fast reflexes and decision making.

 

[gamervivek]

Also take a look at inspection time.

www.en.wikipedia.org/wiki/Inspection_time

 

[know of fence]

Fantastic links. Thank you for these!

Mind you, people see differences only under special conditions (fast movement for one thing), our brain can blend out blur or hitching and we still imagine ourselves in a 3D environment, rather than a slide show. 30 FPS simply was good enough, for certain games paced a certain way played on a slow LCD TV.

The other thing is that "more FPS is better" is a lie that we've been fed for more than a decade. Average FPS is just a stupid GPU benchmark, that says nothing or very little about the actual experience. A game at 60 FPS can hitch or run like butter. Often dual card setups would double FPS without improving image quality at all, in fact additional tearing made it actually worse. Only now word starts getting around that frame time consistency, frame synchronizing, frame latency and pixel transition (response) times are more important for a smooth experience.

This reveal mostly happened because VR headsets provide conditions, where people are much more sensible to these things, we are no longer staring at real-time 2D projection, but rather it's real, head-tracking 3D.

 

[UaVaj]

playback and rendering are apples and oranges.

stop confusing the two.

 

[dacostafilipe]

The other thing is that "more FPS is better" is a lie that we've been fed for more than a decade. Average FPS is just a stupid GPU benchmark, that says nothing or very little about the actual experience. A game at 60 FPS can hitch or run like butter. Often dual card setups would double FPS without improving image quality at all, in fact additional tearing made it actually worse. Only now word starts getting around that frame time consistency, frame synchronizing, frame latency and pixel transition (response) times are more important for a smooth experience.

This.

It's not a problem about frames per seconds but more about the difference between the frames.

You can have 60hz, but if the time between the frames differ to much, it will not "look" fluid.

 

[know of fence]

playback and rendering are apples and oranges.

stop confusing the two.

But that's exactly the problem (one of many actually), unintentional confusion or deliberate equivocating of FPS and FPS.

Is this website rendering or playing back http://www.30vs60fps.com/ ?

That's unclear. Even though when talking games the assumption is rendering, the website is actually using playback of recorded images to demonstrate the effects. Ooops.

 

[flexy]

It is NOT only just because that we perceive, say, 60hz or 120 hz as smoother and fluider than 30hz.

It's also has a technical advantage when a game engine is outputting/updating faster. Obviously, when a game engine renders only at 24hz or 30hz there is some lag implied simply from this fact alone. (Depending on the engine, it might even also ask keyboard/mouse input only at this slow rate as well).

And of course the old 24hz cinema standard was not established "to make things look sufficiently fluid"...the priority there was rather what is the lowest they can go to preserve film. (Even then they need to use tricks in the movies to "double" the 24hz [each frame is actually shown twice], this alone proves that 24hz is no perceived as smooth...and sure not sufficient for computer games).

>>
the human eye can't detect any difference between 24-30 fps and more
>>

Then I recommend an eye doctor for whoever wrote this. Ever seen a true 24hz pan in a movie? It's outright horrible.