after 120hz I can't see a difference. Even at 160hz the only way I can tell, I point my DSLR at the screen and see how the refresh rate lines have slowed down dramatically.

I don't think it will be much smoother, but the refresh rate is so high it could potentially negate any frame tearing. Make it worth it imo.

I'm curious, what's the model that does 240hz ? link plz ?

http://www.esreality.com/index.php?a=post&...pid2082653

i wonder what hz you would get if you set it to 320x240 which is quake live's smallest resolution. I know its won't be fullscreen but you could just use your driver to set it to the custom 320x240 res and fit to exact width and height of the unfullscreen quake to get it fullscreen.

I tried 320x240 to gain some fps but it's unplayable.
Quake live's smallest resolution is 1x1. Perfect for pixel rails.

i would use v-sync anyways if i had a CRT still, so tearing wouldn't be a problem.

you would use vsync on a 240hz crt!? :C

i mean it's not necessary to have 240hz to negate frame tearing, as you can use v-sync on a CRT even with lower refreshrates and still don't suffer from inputlag.
if there is any inputlag 125fps@125hz on a CRT, then it's a lesser problem than the tearing with those settings at least in my oppinion.

i disagree, vsync introduces lag i'd want never want in a fps. 240hz/250fps sounds quite reasonable though and shouldn't have much tearing.

what would be the input lag 125hz/125fps on a CRT? 0,5ms?
you can't notice, but you clearly see the tearing if you turn off the vsync.

yea, therefore use 120hz/125fps.

That's interesting, because with 160 Hz, I had frame tearing. : /

try 120hz with 60fps and compare it to 60hz with 60 fps :E

what will happen?

singularity

I had 3 CRT that were running at 200Hz at 800x600, the last one was even running at 255Hz at 640x480. The difference between 200Hz and 255Hz is realy noticable. The frame tearing is substentialy lower.

In fact the real advantage between a 120Hz CRT and a 200+Hz CRT is the frame smoothness/tearing. http://www.esreality.com/post/2378927/quantif...pid2379261

I have to say i myself would have had a hard time understanding it if i didn't tested it myself, so i can understand people's missconception about getting a refresh rate higher than the frames per seconds.

I hope this helps you!

When you say that the frame tearing is substantially lower, what does this look like? Ie. what is the practical appearance to the end user.

I'm a huge fan of high refresh rates, and it frustrates me to no end that refresh rates haven't continued to increase over the years and instead have even lowered going from crt -> lcd. I know they are different technologies and all, but yeah, crank up the refresh rate industry!!

Well i think i am gonna make a post to explain the test methode. By the way, the more a game is fast, the more the refresh rate you need. There is a direct correlation between the speed at which an image move and the refresh rate and screen tearing, people working @television @movies know what i mean and they have a motion tricks to avoid tearing.

Now, for fast FPS the effect is clearly more proiminent than for a RTS or a car game, because the ingame images moves faster. Here again we are back to the same correlation "size of the movement per time" = speed.

I think my english is horid right there but i hope it was understandable.

Well, the higher the refresh rate, the faster new frames are being drawn thus less input lag.

But I noticed myself a difference between 150hz and 160hz on my old trinitron. 150hz being actually smoother with less frame tearing.

screen artifacts are the result of hz/fps being too far apart, not the result of higher hz. tearing (as in seeing a line updating the screen) usually only happens when fps=hz.

I want to see a 1000hz setup. A monitor at a 1000hz, a game at 1000fps, and a mouse at 1000hz. The bandwidth required would be massive between the card and the monitor, but would be crazy impressive to see.

This!!

honestly i don't get why QL is the only serious competitive fps left with such a low fps cap yet it's one of the most dynamic ones(cs had 100fps cap, but csgo has none, most other titles got adjustable caps). The physics are already independent so it shouldn't cause any issues to increase fps to smth like 250, it would bring much better experience for every QL player.
There're 144hz lcd monitors already and people buying those are capped by the game itself :/

I don't know how much work it would take to make the physics independent at any com_maxfps setting. From what I remember of the old qlive forums, it was stated that the physics was only "fixed" up to 125fps. Beyond that it was/is still bugged (eg 333fps).

source? physics are fps independent in ql afaik.

Go ahead and dig through qlive forums if you like. Here's a post from "injx jr" about it:

http://www.quakelive.com/forum/showthread.php...t=uncapped

Trying to find the thread where I believe Sync himself said it. Can't remember if it was on the old beta forums or not. But the question was brought up there, why can't we have uncapped fps if the physics are independent in qlive. The response was it's only independent up to a certain point, that being 125.

"Someonesmart" was injx, and someonealsosmart was sponge.

At the time I wasn't sure if I could quote them since it was on a private forum.

I don't feel any difference between 120hz and 160hz. I have an old CRT (Eizo) which works at 160hz max. I am playing with120hz atm

My Samsung SyncMaster 959NF does 800x600@160hz but I play with 1024x768@120hz because the difference is really small. The smoothest setup I know is com_maxfps 320 and 160hz which you can only do on localhost (at least on q3). The difference with that setup is like going from 60hz to 120hz.

I read on a couple of sites that a human eye cannot really see the difference for anything above 90hz. I'm not sure if that's true. All I know is that going from 60hz to 120hz for me was quite big, and I hate even speccing on my laptop that is only 60hz.

(not to mention input lag from the old LCD was huge, this new 120hz Benq is awesome, less than 4ms.)

the eye is just the lens, it's the brain that counts, and maybe a human can't consciously distinguish the difference between 90 and above but it might well make a difference subconsciously

It's about motion blur quality. Higher the better and motion blur produced by 120 fps is far from perfect at quake's speed.

Yeah, that makes sense. Thanks for the explanation :)

Actually, it could be a little more complicated.
I can see a pattern in the fps/hz mechanism.

With v-sync off, as soon as the game generated a new frame, the screen starts drawing it immediately and 'breaks' the visible motion, provoking tearing.
If no newer frame has to be drawn, it simply draws the previous again until a newer frame comes to break.

By representing time of the frames to be drawn on a diagram, we can see how much it affects visuals.
Please note that this is purely theory as I've never seen a screen working above 160 hz.
However, i can visually confirm the results I saw working.


So, here we go :


Synced rates 125 fps / 125 hz :
perfect scenario (fps on top) :
1.........1.........1.........
*.........*.........*.........

or it could be :
1.........1.........1.........
..*.........*.........*.......

Due to both rates not being perfectly synced at all time, you only see one single line of tearing moving accross the screen very slowly.
It's the setting with the cleanest image (my personal favorite), even though that tearing line can be quite disturbing for some people.


With 125 fps / 150 hz :
<=> 8 ms / 6.66 ms
As it represents a fraction of 6/5, here's the diagram : (6 dots for game frame, 5 dots for screen frame, set font on Courier New to see it properly)

1.....2.....3.....4.....5..... -> 5 frames before it gets synced back in...
*....*....*....*....*....*.... -> 6 * 6.66 = 40 ms.

We saw that we got a tearing line every time the game renders a new frame.
Thus, we actually see 5 tearing lines popping on 5 distinct locations on the screen, every 40 ms.
Although since 40/5 = 8 ms, we do have a tearing line each frame.

I think it looks nice, for those who don't like the single line tearing, it could be the way to go.


125 fps / 160 hz :
<=> 8 ms / 6.25 ms
Fraction of 32/25, so I can't be arsed to show you the diagram :)

The result shows that we have 25 different lines every 200 ms.
Since 200/25 = 8 ms, we still have a line each frame...
What does this prove ? That no matter what the refresh rate is, we'll always have a tearing line somewhere on the screen.

The only thing that vary is on how many locations on the screen the lines will be visible, 25 here because refresh rate is a multiple of 5.

Still, I personnally prefer how 150 hz looks.


For those who can go up to 125 fps / 200 hz :
<=> 8 ms / 5 ms
Fraction is 8/5 (no shit sherlock) :

1.......2.......3.......4.......5....... -> 5 frames before it gets synced back in...
*....*....*....*....*....*....*....*.... -> 8 * 5.00 = 40 ms.

Since the refresh rate is a multiple of 25, same as 150 hz : 5 different lines.

Eventually, as 125 allows multiples 25 and 5 only, the only 2 possibilities are respectively 5 or 25 lines.

Wait, there's another possibility. What if I bought an XL2411T ? =)


125 fps / 144 hz
<=> 8 ms / 6.944 ms
Represents a fraction of 144/125... Shit, these ones are coprime integers.

Basically it means we'll have a tearing line at a different location every frame !

But actually, it looks only slightly different from 160 hz, I couldnt tell you which one is better.

Tbh I've never noticed any tearing on crt. Maybe I just got used to it.

yea.. but what the most common setting: 120hz/125fps? :)

24 tearing lines. If you have perfectly stable fps

edit: Actually few less due to blanking interval.
edit2: Not really (vsync is almost the same as hsync)

It's important to say that these lines never pop at the same place, due to hz and fps not perfectly synced.
that's why 144hz looks pretty similar to 160 or 120 I think.

If they were really perfectly synced in a 125fps/125hz scenario, we should see ONE line splitting the screen, always at the same place and never going away. :P

They pop at the same place every 200 ms if that's what you mean.

That's not what I mean. I'm saying that if both rates didn't fluctuate AT ALL, the lines would always be at the same place.

yes, but why would you want that to happen? qw at 77hz/77fps was horrible to play because of this.

I don't want that to happen.

huh? then i misunderstood half of your posts.

And the best of all has to be 250hz. Double synced : 1 line of tearing, but appearing only half a frame :>

is that bad or good in comparison with the examples above? tldr

Imo, the best are 125hz and 250hz (1 line)
Then multiples of 25 : 150hz, 175hz, 200hz (5 lines)
Then multiples of 5 : 120hz, 130hz, etc... (25 lines)
And finally the rest.

Note that 120hz being lower than 125fps, it induces a small bit of input lag : because the screen refreshes in 8.33 ms, it will always be 0.33 ms later than the game.

Anything below 120hz is crap.
And the best of all has to be 250hz. Double synced : 1 line of tearing, but appearing only half a frame :>

and below that, especifically 60hz and 75hz with 60 or 125fps?

so the achievable optimum is 125fps/125hz? is it possible to overclock/tune/etc. a crt screen to do exactly 125hz when at the same res it was doing 120hz before (800x600)?

Yes it is. Generate cvt or gtf modelines and use your graphics drivers or some tuning soft. I forgot the name.

125hz/125fps technically has it's own disadvantages and imo it's the matter of preference. At 120+ hz even vsync is acceptable.

You can use powerstrip, but you never seem to get _exactly_ any particular refresh rate. Always some heinous decimal. I wouldn't be surprised if the actual refresh rate fluctuates as well, and if the behavior wasn't display dependent.

Use gtf timings instead of cvt

<insta reply @ 7:45 am :>

Edit: Even if refresh rate was fluctuating the tear lines would be slowly crawling around their calculated positions.

Of course.

When i had my Trinitron with XP, I used ReForce to create a modeline in the monitor driver, then created a custom res with Nvidia drivers. (choose GTF timing, gives the best result)

Now with seven and ATI drivers, i have to create a modeline in the registry. The path is in drivers informations, and the key is "DalNonStandardModesBCD1" (something like that).

With seven and Nvidia drivers, I don't know.

And for any refresh rate below 125 hz, the higher the better (due to input lag)

60 hz - 25 tearing lines
75 hz - 5 tearing lines

that's weird, i always found one line updating the screen to be very annoying. no matter if it was hz=fps or hz x2/3 in qw with 77/77 and 77/144 or in q3 with 125/125.

i preferred almost any other setting to experiencing 'the line', but i remember some qw players (def for example) swearing by it. you're right about input lag, although i still play with fps (slightly higher) than hz in every fps.

Actually he isn't right about output lag.
Screen is refreshed in continuous fashion so some pixels will be drawn almost instantly after rendering frame (unless there is vsync). If hz<fps then some pixels just won't be drawn at all. Generally higher hz = lower mean value of output lag.

I disagree.

If Vsync is off, the screen will draws the frame as soon as the game outputs it, even in the middle of a refresh.

What happens is that some frames (parts actually) are sometimes still visible because there's no new frame to be drawn.

125fps A.......................B.......................C...
60hz 1.................................................2.
(both lines have same length)

The frame B will create a tearing line. Half of the visible frame at that point will suffer from 8 ms output lag.
But just before refresh 2 comes in, we will be at 16 ms.
Hence the output lag generated by hz < fps

125fps ............A.......................B.......................
125hz 1.......................2.......................3...........

In that case, the outpug lag will be 8 ms maximum. Increasing refresh rate won't change this fact, it's entirely fps related.
Now, the only benefit of having more hertz is to reduce tearing the way you like (144hz, 150hz, 200hz, whatever floats your boat)
And if you want less that this maximum of 8 ms, you need more fps, period.

I use the definition where pixel's output lag is how long ago displayed pixel was rendered.
I assume that all pixels in frame are rendered simultaneously.

Increasing refresh rate above fps doesn't decrease max output lag but it decreases it's mean value (over time). I agree that it's probably not significant at high refresh rates.

I was referring to this sentence

Note that 120hz being lower than 125fps, it induces a small bit of input lag : because the screen refreshes in 8.33 ms, it will always be 0.33 ms later than the game.

It seems that you used a definition that input lag is max possible input lag. Sounds confusing if you ask me but ofc it's correct.

edit: sry - ofc increasing hz decreases max output lag by 1/old_hz - 1/new_hz according to mine definition.

I understand what you mean.

Increasing refresh rate will of course allow frames to be drawn faster, and the higher refresh rate you have, the quicker frames are 'complete'. (in a time [1 / hz] like you said)
But still you have to wait [1 / fps] anyway until the next frame comes.

You need to wait 1/fps + 1/hz until all pixels are replaced. Notice that they usually aren't displayed for 1/fps+1/hz period of time. I count time from the moment they were rendered.

Why do I use this definition? The moment frame is rendered is the moment when displaying process starts affecting input/output lag.


          1/fps                                                                                                                                               
fps A...............B...................                                                                                                                       
hz  ...|............|............|......                                                                                                                       
    ^               ^    1/hz    ^                                                                                                                             
    |               |            |                                                                                                                             
    | Pixels are ren|dered       |                                                                                                                             
                    |            |                                                                                                                             
                    | New pixels |are rendered                                                                                                                 
                                 |                                                                                                                             
        Last of the old A pixels |                                                                                                                             
                   gets replaced |


Bars on hz line are just to show refresh rate.

Yeah, you count from the moment they start to be rendered until they completely disappear. Makes sense.
Actually we were both right :) But of course, like you said we have to consider 1/fps + 1/hz the ultimate measurement of screen lag.

So what we learn is that 300fps + 100hz (= 3.33 + 10 = 13.33ms) would be better than 125fps + 160hz (= 8 + 6.25 = 14.25ms) for example.

I think we're done. =)

Naah. I've read it again and these are some bogus definitions.

Pixel's input/output delay is time it takes for pixel to update from the moment you move your mouse.

input/output delay = hardware delay + rendering delay + vsync or similar delay + output delay

Output delay is between 0 and 1/hz and mean value for all pixels is 1/2hz.

exactly, it's not really a single line, you can basically see it for quite continuous time on ONE place(or moving slowly) which is a million times more annoying than random tearing EVEN while every frame all over the screen. It would just appear every 2nd/3rd/etc refresh comparing to 125/125 without vsync and not in a frame row, but that's not really making it less annoying

there's also such thing as smoothness where the generated frames are output nearly synced to display refreshes(with fps&hz being 2-5% different to eliminate 1 place tearing) so I'd prefer smth like 128hz/125fps over smth like 160hz eventhough it's a bit less responsive

This thread reached the point where I don't want to learn about this.

: D

CRT > ALL

I use 144hz/125fps/1024*768. Feels a bit better than 120hz at the same resolution.

Isn't this placebo, though? If you are only rendering 125 frames per second you cannot see more than that.

You're probably right. It does genuinely feel smoother to me. I have no arguments to base it upon. I can't remember who; someone said to give it a shot as they used it in qw with 333 fps or some such. I gave it a shot, and it "felt" better.

It didn't decrease the smoothness or feel in game at least, so I've kept it at 144 :)

On a crt a few years ago I was using 180hz and I tested it numerous times agaisnt 120hz, it definitely felt smoother no other way to put it. I agree, nothing to base it on but from personal experience the higher the hz = the better.

input lag is slightly lower on 144

BenQ XL series pwn