ULMB VSYNC or RTSS S-sync?

Ask about motion blur reduction in gaming monitors. Includes ULMB (Ultra Low Motion Blur), NVIDIA LightBoost, ASUS ELMB, BenQ/Zowie DyAc, Turbo240, ToastyX Strobelight, etc.
bkragnarok
Posts: 4
Joined: 23 Dec 2018, 02:25

ULMB VSYNC or RTSS S-sync?

Post by bkragnarok » 23 Dec 2018, 03:24

Hi, i'm tryimg to achieve low input lag tearless ULMB, but i have a few questions

I tried in sc2 and lol:
120hz ULMB + VSYNC ON with -0.01HZ capped fps
120hz ULMB + VSYNC OFF + RTSS scanline sync=1080

In the both cases i achieved tearless, but what option has less input lag? vsync on + capped fps or rtss s-sync + vsync off?

Is there a difference in s-sync options? my vertical total are 1144 so i tried 1080, -1 and -50 but i didn't see any difference

In 60hz wich would be better? low lag vsync or rtss s-sync?

Acer XB241H
i7 6700k
GTX 1070
16gb ram

User avatar
RealNC
Site Admin
Posts: 3743
Joined: 24 Dec 2013, 18:32
Contact:

Re: ULMB VSYNC or RTSS S-sync?

Post by RealNC » 23 Dec 2018, 09:24

bkragnarok wrote:but what option has less input lag? vsync on + capped fps or rtss s-sync + vsync off?
s-sync runs vsync OFF, so it has much lower latency (same latency as g-sync.)
Is there a difference in s-sync options? my vertical total are 1144 so i tried 1080, -1 and -50 but i didn't see any difference
Try -400.
In 60hz wich would be better? low lag vsync or rtss s-sync?
s-sync (when it works) is always better. Always.
SteamGitHubStack Overflow
The views and opinions expressed in my posts are my own and do not necessarily reflect the official policy or position of Blur Busters.

bkragnarok
Posts: 4
Joined: 23 Dec 2018, 02:25

Re: ULMB VSYNC or RTSS S-sync?

Post by bkragnarok » 24 Dec 2018, 07:38

Thx for your reply, I tried -400 but has a tear line below the center of the screen

Tested at 120/100/85/60hz and s-sync=1060 is the best for me (1080p)

So, for stable fps, s-sync is always the best, but when vsync is needed for unstable fps fluctuations.. can i set VSYNC ON + S-sync=1060 or in this case VSYNC ON + Capped FPS is better?

User avatar
RealNC
Site Admin
Posts: 3743
Joined: 24 Dec 2013, 18:32
Contact:

Re: ULMB VSYNC or RTSS S-sync?

Post by RealNC » 24 Dec 2018, 10:20

bkragnarok wrote:Thx for your reply, I tried -400 but has a tear line below the center of the screen
Now you know what it does. If you set it -200, the tear line will move down. -100, it will move very near the bottom. The whole point of this value is to find the one that hides the tear line below the bottom of the screen.
So, for stable fps, s-sync is always the best, but when vsync is needed for unstable fps fluctuations.. can i set VSYNC ON + S-sync=1060 or in this case VSYNC ON + Capped FPS is better?
s-sync doesn't work with v-sync. So you need the usual -0.01 normal cap.
SteamGitHubStack Overflow
The views and opinions expressed in my posts are my own and do not necessarily reflect the official policy or position of Blur Busters.

User avatar
Chief Blur Buster
Site Admin
Posts: 11648
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: ULMB VSYNC or RTSS S-sync?

Post by Chief Blur Buster » 25 Dec 2018, 01:18

RealNC is correct.
But there's an easier way than randomly trying negative numbers.
bkragnarok wrote:Thx for your reply, I tried -400 but has a tear line below the center of the screen
Enable the hotkeys and use the hotkeys to move the tearline up/down.

If lag isn't important, VSYNC ON is superior. Use VSYNC ON whenever lag isn't important.

RTSS scanline-sync is superior in lower lag though. But it is simply a complicated way to do something that looks like VSYNC ON but with lots less lag, but is extremely hard to calibrate unless you understand how to calibrate the location of tearlines. It is only for advanced users who understand how negative numbers translate to tearline locations.
bkragnarok wrote:Tested at 120/100/85/60hz and s-sync=1060 is the best for me (1080p)
If your signal is a Vertical Total 1125 (viewed in Custom Resolution) then the corresponding negative number s-sync=1060 is (1060-1125) = -65

You can keep using positive numbers.

However, negative numbers are useful because it's more resistant to resolution-changes (e.g. switching between 1080p and 1440p) and vertical total changes (e.g. large vertical total tweaks) since lag can be lowered from using late-VBI framebuffer flips rather than early-VBI framebuffer flips.
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!

User avatar
Chief Blur Buster
Site Admin
Posts: 11648
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: ULMB VSYNC or RTSS S-sync?

Post by Chief Blur Buster » 25 Dec 2018, 01:33

RealNC wrote:
So, for stable fps, s-sync is always the best, but when vsync is needed for unstable fps fluctuations.. can i set VSYNC ON + S-sync=1060 or in this case VSYNC ON + Capped FPS is better?
s-sync doesn't work with v-sync. So you need the usual -0.01 normal cap.
To be clear (I got confused until I reread and re-parsed)

To be clear, RealNC means:
"RTSS Scanline Sync does not work with normal VSYNC ON in its default refresh-rate-capped configuration. Also, you don't cap below Hz with RTSS Scanline Sync. However, if you use the older Low-Lag VSYNC ON HOWTO instead of the newer RTSS Scanline Sync, you do need the cap-below-Hz, usually 0.01"

You can use RTSS Scanline Sync with any uncapped sync method such as VSYNC OFF, or Fast Sync, or Enhanced Sync if you follow special instructions. This will allow you to permanently avoid tearlines, provided you understand how to correctly calibrate your tearing first before disabling tearing.
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!

bkragnarok
Posts: 4
Joined: 23 Dec 2018, 02:25

Re: ULMB VSYNC or RTSS S-sync?

Post by bkragnarok » 25 Dec 2018, 07:54

Thx for the replies, ok let's see if I understood correctly.

When I use VSYNC OFF the tear line has to be below the edge of the screen in the VBI, right?
But if I want to combine FAST Sync+S-Sync, the tear line has to be above the edge of the screen so that it never touches the VBI, right?
Chief Blur Buster wrote:Enable the hotkeys and use the hotkeys to move the tearline up/down.
I've searched how to use the hotkeys and added the following lines
[Framerate]
SyncFlush=1 (Is it still necessary to add this line in the latest version? since by default it's not included)
SyncHotkeys=1 (To move tear with ctrl+shift+up/down)

In 120hz my VT are 1144 and the tear line disappears in 1056
In 60hz my VT are 1125 and the tear line disappears in 1060
If I want to use a global cfg for 60/100/120hz and I set S-sync=-65, when I switch to 120hz from 60hz, then the tear will be on VBI 1079, so the positive value 1060 will be better than -65?
Chief Blur Buster wrote:However, negative numbers are useful because it's more resistant to resolution-changes (e.g. switching between 1080p and 1440p) and vertical total changes (e.g. large vertical total tweaks)
Now I understand what's the benefits of using negative values, but in my case that I always use 1080p it's not very necessary, right?
Chief Blur Buster wrote:since lag can be lowered from using late-VBI framebuffer flips rather than early-VBI framebuffer flips.
So, will it be better using something like -1 for VSYNC OFF?

User avatar
RealNC
Site Admin
Posts: 3743
Joined: 24 Dec 2013, 18:32
Contact:

Re: ULMB VSYNC or RTSS S-sync?

Post by RealNC » 25 Dec 2018, 10:34

bkragnarok wrote:When I use VSYNC OFF the tear line has to be below the edge of the screen in the VBI, right?
But if I want to combine FAST Sync+S-Sync, the tear line has to be above the edge of the screen so that it never touches the VBI, right?
There is no "above the edge". The first scanline is 1 and the last is 1080 + the VBI area.
SteamGitHubStack Overflow
The views and opinions expressed in my posts are my own and do not necessarily reflect the official policy or position of Blur Busters.

User avatar
RealNC
Site Admin
Posts: 3743
Joined: 24 Dec 2013, 18:32
Contact:

Re: ULMB VSYNC or RTSS S-sync?

Post by RealNC » 25 Dec 2018, 10:42

bkragnarok wrote:SyncFlush=1 (Is it still necessary to add this line in the latest version? since by default it's not included)
It is set to 0 by default. It was never included in the profile by default. You set it to 0 if you want a more stable tear line position, and to 2 is you want an absolutely stable tear line at the cost of increased risk of stutter.
Now I understand what's the benefits of using negative values, but in my case that I always use 1080p it's not very necessary, right?
It doesn't matter. Negative numbers are easier to use. If you don't care, don't use them.
Chief Blur Buster wrote:since lag can be lowered from using late-VBI framebuffer flips rather than early-VBI framebuffer flips.
So, will it be better using something like -1 for VSYNC OFF?[/quote]
1 scanline is not enough headroom. The GPU needs a few scanline to finally sync. If the GPU needs 20 scanlines to sync, at 1080p you should use 1060 so that the sync happens near the last visible scanline.
SteamGitHubStack Overflow
The views and opinions expressed in my posts are my own and do not necessarily reflect the official policy or position of Blur Busters.

User avatar
Chief Blur Buster
Site Admin
Posts: 11648
Joined: 05 Dec 2013, 15:44
Location: Toronto / Hamilton, Ontario, Canada
Contact:

Re: ULMB VSYNC or RTSS S-sync?

Post by Chief Blur Buster » 25 Dec 2018, 13:06

RealNC wrote:It doesn't matter. Negative numbers are easier to use. If you don't care, don't use them.
Yes, they both behave the same.
RealNC wrote:
Chief Blur Buster wrote:since lag can be lowered from using late-VBI framebuffer flips rather than early-VBI framebuffer flips.
So, will it be better using something like -1 for VSYNC OFF?
1 scanline is not enough headroom. The GPU needs a few scanline to finally sync. If the GPU needs 20 scanlines to sync, at 1080p you should use 1060 so that the sync happens near the last visible scanline.[/quote]
Correct, bigger headroom is better. The scanline index in RTSS and the actual tearline location, are usually offset.

It's tighter on faster systems at lower Hz, and can sometimes be exact. However, on slower systems and to allow performance overhead (to allow tearline to jitter downwards away from its prescribed index), you want the tearline to be further upwards (earlier in VBI if VSYNC OFF, or higher above bottom edge if using FastSync/Enhanced Sync) so that timing errors or long frametimes (0.1% frametimes especially) delay a frame a bit. You need to try to squeeze the whole tearline-vibration amplitude into the VBI. A bigger VBI can hide a erratically-vibrating tearline better, so the use of a slightly larger Vertical Total can help Scanline Sync a bit.
RealNC wrote: There is no "above the edge".
<TECHNICAL>
This requires a complex reply.
(The diagrams are useful for other readers to understand RTSS Scanline Sync, so this post adds to discussion)

Actually, Einstein says it is all relative. See photo of old analog VSYNC below. Mathematically, a tearline is simultaneously above and below the refresh cycle when it's in the VBI. Technically, the Vertical Back Porch is above the top edge and the Vertical Front Porch is below the bottom edge, in display engineering!

Image

While the VBI (Vertical Blanking Interval) consists of grand total (Front Porch + Sync + Back Porch), technically, the Front Porch is analog overscan above the top edge and the Back Porch is analog overscan below the bottom edge, while the Sync signal triggers an electron gun movement back to the top edge.

And if you've lived in the 1970s or earlier, a misadjusted VHOLD knob on an old TV allowed you to see the VBI as active scanlines. In fact, if you jacked brightness WAY up on these old televisions then intentionally misadjusted the VHOLD analog knob, the Front/Back porches were often a very dark grey (7.5 IRE, standard analog black level), while Sync was black (0 IRE, below analog black level) so the VBI looked like three-layer ice-cream sandwich if you brightened things to a bleach-out extent. But normally, the whole shebang is completely black.

The refresh cycles and the VBI are just simply endless loop of scanlines continually transmitted out of the GPU output, just as display signals always has been for nearly 100 years (at the image delivery layer).

Image

The above is VSYNC ON. The below is VSYNC OFF.

Image

The name of game of the brand new RTSS Scanline Sync is to aim the tearlines into the VBI -- that metaphorical black "spacer" between refresh cycles. The timing of frame-presentation API is when the tearline occurs, and the game of RTSS Scanline Sync is to time frame presenting in the VBI between refresh cycles.

Although it was possible to make the VBI visible in the analog era -- Now, in the digital era, the VBI is just dummy blank offscreen pixel rows in the signal that are simply delay padding (to give electronincs time to initialize a new visible row of pixels onscreen, or to initialize the start of a new refresh cycle; some newer displays only require literally 1 pixel of sync in nearly nonexistent blanking intervals). What this means is that the topology is identical on a 1930s TV signal versus 2020s DisplayPort signal (at least at the image delivery layer in terms of pixel order, sync order, endless-loop-of-scanlines topology), blanking interval, not the newly digital/micropacketization layers right below it).


(Curious about Horizontal stuff too? Yes. HBI - Horizontal Blanking Interval consisting of Horizontal Front Porch, Sync, and Back Porch also exists -- the horizontal sync -- and sometimes becomes visible to human eyes during ghost signals when 2 analog TV signals are overlapping each other -- and shows up as vertical bars between left/right edges of the ghost signal overlapping the real signal. However, HBI is literally microseconds, while VBI is closer to milliseconds, and so HBI knowledge is not as useful since it literally just literally a quick comma-separator between pixel rows nowadays. However, some people reduce the HBI to gain extra bandwidth headroom to increase the VBI. Reduced Horizontal Totals is good for creating Large Vertical Totals, since trading HBI pixels for more VBI pixels can keep you under the Pixel Clock / Dot Clock maximum bandwidth budget.)

In the digital era, the components of the blanking intervals are just delays. Such as between pixel rows (HBI), or between refresh cycles (VBI), but technically, refresh cycles are like endless series of scanlines (= digital pixel rows) that alternate between active scanlines and VBI scanlines and the topology of a display signal means a VBI is between the previous refresh cycle and the next refresh cycle, so technically, topologically mapped out on a time-basis, the VBI is also simultaneously before and after a refresh cycle, and if we're thinking geometically, a metaphorical tearline in VBI can be above the edge. So, it's legitimate thinking if we're mapping tearlines geometrically.

Whatever the display does internally doesn't matter (there's no concept of above/below, it's just dummy delay loops or wait-for-initialization triggers). But geometrically, when mapping out a display signal.... Above/below both simultaneously exists, full stop.

So, Chief Blur Buster deem it "Yes, there's a such a thing as geometrically thinking that tearlines are above the top edge of the screen" -- that is permitted & legitimate thinking around here, if one wishes to do so for simplicity's sake, since the signal structure is simply an endless orthographical loop of scanlines including VBI scanlines (dummy black scanlines) and active visible scanlines (rows of pixels).

That's why negative indexes exist, it is geometrically thought as "above top edge".
And beyond-bottom-of-screen indexes exist, it is geometrically thought as "below bottom edge". Both answers are definitely correct, because it is simultaneously "below bottom edge of previous refresh cycle, and above top edge of next refresh cycle" -- referring to exactly the same "figurative-tearline-in-VBI" -- and thusly, are both legitimate answers because as Einstein says, it is all relative.

Both large positive indexes and small negative indexes are identical, but they can diverge if your VBI sizes change (Vertical Total tweaks) or resolution changes.
</TECHNICAL>
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!

Post Reply