Smoothness

Everything about displays and monitors. 120Hz, 144Hz, 240Hz, 4K, 1440p, input lag, display shopping, monitor purchase decisions, compare, versus, debate, and more. Questions? Just ask!
rootsoft
Posts: 42
Joined: 01 Nov 2014, 13:07

Smoothness

Post by rootsoft » 11 Feb 2020, 12:27

Why does 144fps at 144hz feels so much better than at 144hz with higher FPS.

I have experience this with multiple screens.

:oops:
Last edited by rootsoft on 12 Feb 2020, 15:43, edited 1 time in total.

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Smoothness

Post by Chief Blur Buster » 11 Feb 2020, 14:14

Your question is potentially mis-phrased.

Do you mean:

(A) "Why is 144fps at 144Hz smoother than 144ps at a higher refresh rate?"

or do you mean:

(B) "Why is 144fps at 144Hz smoother than higher frame rates at 144Hz?"

Do you mean (A), or do you mean (B)?

Either way, both are true, because framerate=Hz is always smoother.
-- The difference between the two creates a beat frequency (aka stutter).
-- There are 1 stutters per second for 145fps at 144Hz non-VRR.
-- There are 1 stutters per second for 143fps at 144Hz non-VRR.
-- There are 4 stutters per second for 148fps at 144Hz non-VRR.
-- There are 4 stutters per second for 140fps at 144Hz non-VRR.
-- The difference between fps and Hz creates a beat-frequency effect (aka stutters) on fixed-Hz displays.

That's why 240fps at 240Hz is better than 144fps at 144Hz
That's why there are stutters for 144fps at >144Hz (non-VRR).
That's why there are stutters for >144fps at 144Hz (non-VRR).

It's easy to see for yourself. You can witness beat-frequency stutters. Easiest to see if you use a microsecond-precise frame rate capper such as RTSS. Set an RTSS frame rate cap of 142 with VSYNC ON or VSYNC OFF on fixed-Hz 144Hz -- you will see a very regular cadence of 2 stutters per second in CS:GO or any software that easily runs capped-out -- such as older games with virtually zero frame rate slowdowns are excellent games to test beat-frequency stutters. 2 stutters per second is the beat frequency of 142 and 144. That's the difference between 142 (frame rate) and 144 (refresh rate). In-game framerate caps may show this too, but they may have slightly more erratic stutter due to a different capping algorithm that might average things out more. If your frame rate varies a lot, you get erratic stutters.

Even stutters are visible at higher refresh rates. 239fps at 240Hz = 1 stutter per second. If your motion is fast (e.g. 4000 pixels/sec), a 1/240sec stutter is 1/240th of 4000 = a stutter-jump of 16 pixels. It's impressive how even a single framedrop stutter is still visible even at ultra-high Hz. In fact, at CES 2020, I was still able to see single-framedrop stutters for 360Hz VSYNC ON -- 1/360sec stutter is still (just about) human visible during VSYNC ON motion at very fast motion speeds. ASUS has a long-term roadmap to 1000Hz (thanks in part to Blur Busters), the final frontier is not 240Hz nor 360Hz.

Yep, what you call "stutters" are often beat frequency artifacts from the framerate-vs-refreshrate difference.

G-SYNC and FreeSync automates fps=Hz. The refresh rate changes in realtime to match game frame rate. You also want a VRR display with a wide VRR range to capture your entire frame rate range if possible. Thus, I generally recommend 240Hz+ if you want a low-lag VRR display. The great thing with variable refresh rate displays -- is there's no such thing as a single dropped frame on a VRR display -- it just doesn't exist -- it's stutterless framerate changes like www.testufo.com/vrr -- 60fps goes 59fps goes 60fps virtually invisibly -- A tiny framerate change is completely invisible on variable refresh displays!

Image

Image

Even randomized framerates look perfectly smooth because of the above graph.
TestUFO: Animated demo of struggling random framerates looking smooth
Look at how the framerates struggles to equal Hz, yet stutters don't show, thanks to variable refresh rate (G-SYNC or FreeSync). This TestUFO animation uses interpolation to emulate VRR in software. The monitor waits for the software to deliver a frame, before immediately refreshing on the spot (the monitor hardware slaves to the software whenever in VRR range).
TestUFO: Animated demo of frame rate changes from 30fps smoothly to max Hz
Look at how it's done stutterlessly. 30fps, 31fps, 32fps, 33fps....[etc].....142fps, 143fps, 144fps. No beat-frequency stutters, because the Hz is matching fps, and the (erratic locations of) object positions are still in sync with your (analog) eye tracking positions along the axis of the motion vector. It's impressive how variable refresh rate just simply erases single-framedrop stutters, and tiny framerate changes. Sure, large stutters will show (e.g. disk loading stutters that lasts multiple fixed-Hz refresh cycles) and buggy game-engine stutters will still show -- but the typical single frame "missed refresh cycle" stutters are 100% completely gone with G-SYNC and FreeSync. Because there's no such thing as a missed refresh cycle when the computer monitor WAITS for your software to pass a new frame.

TL;DR: If your priority is smoothness, you want fps=Hz, either via VSYNC ON or via VRR.
Use VSYNC ON if smoothness is more important than lag! Or use variable refresh rate (VRR, such as G-SYNC or FreeSync) to automate fps=Hz while avoiding laggy VSYNC ON. On these displays, the display Hz automatically changes to match frame-rate in realtime, dynamically.

Now, if you use blur reduction such as DyAc or PureXP or ULMB, then your beat-frequency stutters are amplified by blur reduction. A single framedrop stutter becomes MUCH more visible with blur reduction. There's no motion blur to mask single-framedrop stutters. They will show up that much more jarringly. You could go with an ASUS TUF series with variable refresh rate ELMB-SYNC (avoid the VA panels, stick to the good ones).

TL;DR: If your priority is smoothness with motion blur reduction, you may have to use VSYNC ON since most monitors won't do VRR combined with blur reduction.
Yes, it adds lag but makes blur reduction infinitely FAR better looking. To compensate, there are lots of low-lag VSYNC ON tweaks, such as RTSS Scanline Sync, NVIDIA Ultra Low Lag, and/or fractional frame rate capping. And yes to smooth strobe-amplified mouse microstuttering/jittering -- with blur reduction you ideally have to jack up your mouse DPI to 1600dpi or higher (even 3200dpi), with an ultra-fine mouse sensor, very clean mouse feet, good esports mousepad, and ultra-low in-game sensitivity to compensate for high mouse DPI. It's best not to use old-fashioned 400dpi settings with LightBoost/ULMB, it stutter-crapfest for strobing during slower-speed and medium-speed eyetracked mouseturns. Use mouse profiles to use lower desktop DPI since your mouse arrow will be a rocket in Windows if you keep same DPI in desktop as your game.
See: HOWTO: Using ULMB Beautifully or Competitively
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

rootsoft
Posts: 42
Joined: 01 Nov 2014, 13:07

Re: Smoothness

Post by rootsoft » 11 Feb 2020, 16:18

I meant B.

I appreciate the great detailed explanation Chief!

camaoyba
Posts: 23
Joined: 24 Oct 2019, 16:18

Re: Smoothness

Post by camaoyba » 12 Feb 2020, 12:20

That said, is it better to cap my fps down to match my 240hz monitor? (CS GO 400+ FPS)

rootsoft
Posts: 42
Joined: 01 Nov 2014, 13:07

Re: Smoothness

Post by rootsoft » 12 Feb 2020, 15:42

camaoyba you will defenetly have a smoother gameplay capping your FPS at 242.

camaoyba
Posts: 23
Joined: 24 Oct 2019, 16:18

Re: Smoothness

Post by camaoyba » 12 Feb 2020, 17:24

rootsoft wrote:
12 Feb 2020, 15:42
camaoyba you will defenetly have a smoother gameplay capping your FPS at 242.
Thanks, will try that! Just out of curiosity, why 242 and not 240?

Stitch7
Posts: 86
Joined: 27 Mar 2019, 08:26

Re: Smoothness

Post by Stitch7 » 12 Feb 2020, 19:10

if vrr 237 or 238 depending on game.

User avatar
Chief Blur Buster
Site Admin
Posts: 11647
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: Smoothness

Post by Chief Blur Buster » 12 Feb 2020, 19:40

I should write a Capping Differentials FAQ.

Understand The Capping Differential (fps vs Hz)!

1. For VRR, G-SYNC, and FreeSync cap 3fps below max Hz. www.blurbusters.com/gsync/gsync101
This can decrease lag because of pick-poison effect fps=Hz. Capping keeps you away from max-Hz (where VSYNC ON lag suddenly starts to occurs, or if using VSYNC OFF, then tearing suddenly occurs at framerates above Hz). Capping latency is lower lag than VSYNC ON lag. It can be a few frames per second below. For higher Hz, use a bigger differential. For lower Hz, can use a smaller differential. We just suggest a 3fps differential to keep things easy for users as it almost always works well.
....FAQ: Why 3fps below? Caps are never perfect -- they aim for an average frame rate -- a 144fps cap means some frames are 1/145sec and some frames are 1/143sec. A 3fps differential gives a good safety margin and breathing room for an imperfect frame rate cap. At lower refresh rates like 75Hz FreeSync, you can use a differential as little as 1fps or 0.5fps. But at 240Hz G-SYNC, sometimes you may want a bigger differential such as 5fps or 7fps when you're using imperfect in-game frame rate caps. A precision frame-rate capper such as RTSS can get by with a smaller capping differential.


2. For VSYNC OFF smoothness-prioritization, cap a few fps away from Hz (either above or below) to prevent stationary or slow tearline
This may sometimes slightly increase VSYNC OFF lag in some games and reduce lag in others ganes (by avoiding GPU-100%-maxout situation starving other parts of the latency chain). Regardless of latency changes, it makes things much smoother. Remember, this is smoothness-prioritized VSYNC OFF. Occasionally the improved smoothness makes aiming easier because you don't have erratic lag. Fluctuating framerates can create fluctuating lag. Sometimes you want consistent latency even if it's higher lag. Sometimes lag consistency wins some esports games.

3. For RTSS Scanline Sync, it is an automatic low-lag fps=Hz automatic cap. No cap differential.
RTSS Scanline Sync (using software found on Guru3D) is what you use if you crave smoothness with low-lag with low-GPU-usage games -- especially strobed -- like older games. Strobing jitter & strobing lag can almost disappear. Use the floor of your framerate range. If your framerate range is 120fps-240fps, then best to even go lower to 100Hz. Recommended with low-Hz while combined with strobing. Only for your games with very high minimum framerate, since it only works well at a refresh rate below your frame rate valley. As a result, RTSS Scanline Sync becomes God Mode of Strobing / VSYNC ON

4. For smoothness-prioritized Low-lag VSYNC ON, cap approximately 0.01fps below what you see at www.testufo.com/refreshrate
Related HOWTO article: HOWTO: Low-Lag VSYNC ON
This is what you want to use if you want to use traditional VSYNC ON but want to reduce framebuffer backpressure latency -- essentially keep framebuffers from piling up. This is similar to Scanline Sync but more forgiving. However, while average lag tends to be much lower than VSYNC ON, you might have a very slow/gradual sawtooth latency effect as the fps & Hz slowly beat-frequencies each other (may potentially be much better than erratic/random latency

It's Best To Understand Your Cap Differential & Why

Note: The above are simply when you prioritize smoothness over other capping reasons. Sometimes you want to use a bigger differential for a reason, e.g. much lower or higher caps. For example 300fps at 144Hz VSYNC OFF for lower-microstutter VSYNC OFF. Or 60fps at 120Hz fixed-Hz for lower-lag emulators (than at 60Hz) on non-VRR displays. Or 100fps or 144fps or 150fps cap on 240Hz to stabilize a very fluctuating frame rate, to keep gamefeel more consistent. Or another capping reason. Understand WHY you cap.
Head of Blur Busters - BlurBusters.com | TestUFO.com | Follow @BlurBusters on Twitter

Image
Forum Rules wrote:  1. Rule #1: Be Nice. This is published forum rule #1. Even To Newbies & People You Disagree With!
  2. Please report rule violations If you see a post that violates forum rules, then report the post.
  3. ALWAYS respect indie testers here. See how indies are bootstrapping Blur Busters research!

camaoyba
Posts: 23
Joined: 24 Oct 2019, 16:18

Re: Smoothness

Post by camaoyba » 13 Feb 2020, 09:37

Thanks Chief!

User avatar
AddictFPS
Posts: 314
Joined: 15 Jan 2020, 14:13

Re: Smoothness

Post by AddictFPS » 14 Feb 2020, 05:58

Chief Blur Buster wrote:
12 Feb 2020, 19:40
1. For VRR, G-SYNC, and FreeSync cap 3fps below max Hz. www.blurbusters.com/gsync/gsync101
This can decrease lag because of pick-poison effect fps=Hz. Capping keeps you away from max-Hz (where VSYNC ON lag suddenly starts to occurs, or if using VSYNC OFF, then tearing suddenly occurs at framerates above Hz). Capping latency is lower lag than VSYNC ON lag. It can be a few frames per second below. For higher Hz, use a bigger differential. For lower Hz, can use a smaller differential. We just suggest a 3fps differential to keep things easy for users as it almost always works well.
If with 240Hz máx VFreq monitor, user set 240Hz VRR-On VSync-On, with game at rock solid sustained 240FPS, and GPU using default Doble Buffer behavior, there are any lag difference Vs overclock monitor to 243Hz and RTSS framerate cap 240 ?

I suppose that this trick is only for:

1) User set VFreq lower than 240, but want maintain the 240Hz fast scanout, or

2) To avoid games that with VSync-On apply Triple Buffer or more

Post Reply