Stutter Stutter or Smooth as Butter

[jimortality]

Brands Hatch especially, even in practice screen stutter. I think it's not just me and maybe down to the unreal engine? I've tried all different things so if you're smooth as Butter, please post your settings and how you're eliminating it. Thank you.

Acer Predator X34p
I 7-6700k
Nvidia 1080ti
16gb Ram a lam.
And not forgetting my brand new Master Cooler 212 Evo

 

[slatanek]

BTW an interesting thing...I've been tinkering with ACC the past few days trying to find some optimized settings...and while I was at it, I tried the bonnet view, as mentioned here before.

It's actually quite interesting. I've gained about 10% of fps just by using that view alone. And the interesting thing is it's not GPU-based gain, but very clearly a CPU-based one. For some reason, running in bonnet camera is noticeably easier on the CPU than cockpit cam or even dashboard cam. It's not due to the instruments on the dashboard, since you can turn an overlay with the dashboard on even in the bonnet cam and it doesn't affect framerate at all. So I guess something about the cockpit?

But it's an interesting way to gain some fps basically "for free". And the bonnet cam can even be set in the way so it is very close to the dashboard cam, all the way back and to the left (though I wish the adjustment range was a bit wider still)...

I might be using it more often.

Easiest way to measure actual real world latency would be to use a slow-mo camera with both the in-game as well as your sim racing wheel in frame. My experience with ACC is that the lag and overall smoothness improves a lot when ur framerate stays above 90fps, which is not easy with current optimization. It seems like all the other games running at 60fps r smooth and ACC running at 60fps feels like 40-45fps. I have a 144hz monitor with FreeSync2 so my frames r synced from 40fps-144fps and only when my framerate stays above 90fps does feel really smooth.

 

[Demithio]

Easiest way to measure actual real world latency would be to use a slow-mo camera with both the in-game as well as your sim racing wheel in frame. My experience with ACC is that the lag and overall smoothness improves a lot when ur framerate stays above 90fps, which is not easy with current optimization. It seems like all the other games running at 60fps r smooth and ACC running at 60fps feels like 40-45fps. I have a 144hz monitor with FreeSync2 so my frames r synced from 40fps-144fps and only when my framerate stays above 90fps does feel really smooth.

I have a 144hz Freesync1 monitor that I'm using with a 1070 GTX. The issue with freesync is that its software, not hardware like real Gsync. It seems to be ok smoothness wise as long as its about 72 or so fps or higher. Its not perfect, nor was it perfect when running an AMD 5700 XT. AC2 is a GPU hog for certain, its clearly not optimized.

ACC, like AC, seems to suffer from framerate timing and vsync issues. Maybe some people have thought of this already but I found using RivaTuner Statistic Sever locked to 85 fps with vsync (no freesync) on is the smoothest I've seen (granted that you can maintain 85 fps). You must also set your desktop refresh rate to whatever fps you're trying to obtain in ACC, for instant mine is set to 85hz when I use this method. I usually run with freesync on since I cant maintain 85 fps constantly. Anything less, such as 60, just doesn't look or perform well.

 

[RasmusP]

You must also set your desktop refresh rate to whatever fps you're trying to obtain in ACC, for instant mine is set to 85hz when I use this method.

I do the same for some games. Although I have the Dell 3440x1440 gsync monitor + 1070, sometimes gsync has some microstutters and vsync does a smoother job.

About what I quoted:
You can go into steam - > library - > right click into acc properties - > start options/commands
And then type
-refresh 85

This should put your monitor to 85 Hz as long as acc is in full screen mode.
I only used this for counter strike yet. Dirt rally has an internal hz setting that overwrites the hz setting without the need to change anything on the desktop.
Didn't need to change hz for any other game yet.

 

[RasmusP]

Easiest way to measure actual real world latency would be to use a slow-mo camera with both the in-game as well as your sim racing wheel in frame. My experience with ACC is that the lag and overall smoothness improves a lot when ur framerate stays above 90fps, which is not easy with current optimization. It seems like all the other games running at 60fps r smooth and ACC running at 60fps feels like 40-45fps. I have a 144hz monitor with FreeSync2 so my frames r synced from 40fps-144fps and only when my framerate stays above 90fps does feel really smooth.

IIRC someone did this and the weird stuff about acc is that there are two different input lags:
Steering wheel to in-game tyres and then in-game tyres to virtual wheel.

So quite some people are using the dash view without the virtual wheel visible to make it seem like there's less input lag.

Btw 120 fps aren't enough for a camera to measure it. Sadly I can only do 400 fps for 0.5s as an alternative.

 

[Neilski]

IIRC someone did this and the weird stuff about acc is that there are two different input lags:
Steering wheel to in-game tyres and then in-game tyres to virtual wheel.

Oh dear, that's pretty disturbing... Can't imagine a good reason for that.

Btw 120 fps aren't enough for a camera to measure it. Sadly I can only do 400 fps for 0.5s as an alternative.

Yeah, it's a shame that the cameras in phones (etc.) don't allow you to severely downgrade the resolution to gain duration. My phone's sensor (if I understand correctly) is limited by the on-board memory in the 960 fps mode. Fine, but even with that same memory you could record for waaaay longer and still have enough resolution to see what's going on in many situations. Turns out my phone can only do 0.184 seconds of 960 fps at 720p - even less than I had thought a few weeks back. I'd happily tweak it to record at 240p or lower if it would let me - that would allow for more like 2 seconds... (Would also be nice if it allowed me to vary the speed too!)

 

[RasmusP]

Yeah, it's a shame that the cameras in phones (etc.) don't allow you to severely downgrade the resolution to gain duration. My phone's sensor (if I understand correctly) is limited by the on-board memory in the 960 fps mode. Fine, but even with that same memory you could record for waaaay longer and still have enough resolution to see what's going on in many situations. Turns out my phone can only do 0.184 seconds of 960 fps at 720p - even less than I had thought a few weeks back. I'd happily tweak it to record at 240p or lower if it would let me - that would allow for more like 2 seconds... (Would also be nice if it allowed me to vary the speed too!)

I get that they can't really disable pixels in the sensor so a lower resolution might still be too much output from the sensor?

But I really don't get why they don't implement more fps modes. Like 2 seconds at 300 fps etc.
That would only mean more steps in the shutterspeed, which the camera probably can do anyway for photos.

 

[Neilski]

I get that they can't really disable pixels in the sensor so a lower resolution might still be too much output from the sensor?

But I really don't get why they don't implement more fps modes. Like 2 seconds at 300 fps etc.
That would only mean more steps in the shutterspeed, which the camera probably can do anyway for photos.

Well they already have to heavily sum the pixels to get it down to 720p (I'm almost entirely certain they don't just dump some of the pixels)... And even if the extra work of summing the pixels more heavily was too much at 960fps/240p, you'd still expect something like 600fps to work just fine.
It may be more of a marketing/GUI-design decision to limit the speed/resolution options I guess.

 

[Miguel Batista]

I get that they can't really disable pixels in the sensor so a lower resolution might still be too much output from the sensor?

But I really don't get why they don't implement more fps modes. Like 2 seconds at 300 fps etc.
That would only mean more steps in the shutterspeed, which the camera probably can do anyway for photos.

They don't need to disable pixels. They just need to record less pixels. You can capture with all of them but only 50% or another random value get recorded. This could then be done by different algorithms that ensure you have a proper image capture (keeps resolution, etc).

 

[Neilski]

It's possible that a decent part of why input lag doesn't bother me (as much as it does some people) could be that I invariably use a dash or bonnet cam, rather than a cam which lets me see that the wheel isn't moving perfectly in sync with mine...

Easiest way to measure actual real world latency would be to use a slow-mo camera with both the in-game as well as your sim racing wheel in frame.

So... I finally decided to make the measurement today, on AC (not ACC), using the "both steering wheels in view" fashion as @slatanek proposed, along with my 960 fps phone camera which sadly doesn't possess any kind of motion triggering (the numlock=downshift idea from @gyrtenudre is also pretty cool and I may play with that another time).
My monitor is a bog-standard 60 Hz beast and I use vsync. I was expecting something in the ballpark of 3 frames worth of input lag (thus maybe 50 ms).
I was very bummed to find that it's more like 100 ms
I can't see an option in the INI file for AC to tweak the pre-rendered frames - do you know where I should look for that @RasmusP ?
I do plan to play with RTSS and indeed the Nvidia control panel - am hoping that at least one of them (RTSS I imagine) will take effect without having to close and restart the game each time.

 

[Demithio]

I do the same for some games. Although I have the Dell 3440x1440 gsync monitor + 1070, sometimes gsync has some microstutters and vsync does a smoother job.

About what I quoted:
You can go into steam - > library - > right click into acc properties - > start options/commands
And then type
-refresh 85

This should put your monitor to 85 Hz as long as acc is in full screen mode.
I only used this for counter strike yet. Dirt rally has an internal hz setting that overwrites the hz setting without the need to change anything on the desktop.
Didn't need to change hz for any other game yet.

Good tip. Thanks. You seem to be pushing a lot of pixels for a 1070, as I am at 2560x1440. Think its time to upgrade to a 2070 or the likes.

 

[RasmusP]

Good tip. Thanks. You seem to be pushing a lot of pixels for a 1070, as I am at 2560x1440. Think its time to upgrade to a 2070 or the likes.

I am...
I'm basically waiting for the 3060/70/80 to become pre-orderable. 500€ put aside just like for my 1070 in June 2016.
Anyway I was on fullHD for many years and I really enjoyed slamming everything to the maximum or tweaking beyond with my 1070.
But due to the lack of msaa (and therefore forced sgssaa) with modern games I became very frustrated with the pixel crawling. I also always had micro stuttering due to my old i7 2600k in almost all games I was and am still playing (vsync, 60 Hz).
When I've encountered a lot of PDFs being blurry when I zoomed out a little to make them feel good while reading and also started to edit some photos that were always blurry when not zoomed in quite a bit, I decided I needed a higher pixel density.

Didn't have the budget for all upgrades so decided to keep everything else and splash out for the monitor. Get gsync onboard and get happy.
Paying quite a lot for a fullHD or 16:9 1440p gsync monitor didn't feel right, neither did paying 600€+ for an ultra wide 1440p without gsync...

So I got it all in one. And it's awesome since over a year now.
Best piece of hardware I've ever bought!

But my 1070 is crying for help in a lot of scenarios now

So... I finally decided to make the measurement today, on AC (not ACC), using the "both steering wheels in view" fashion as @slatanek proposed, along with my 960 fps phone camera which sadly doesn't possess any kind of motion triggering (the numlock=downshift idea from @gyrtenudre is also pretty cool and I may play with that another time).
My monitor is a bog-standard 60 Hz beast and I use vsync. I was expecting something in the ballpark of 3 frames worth of input lag (thus maybe 50 ms).
I was very bummed to find that it's more like 100 ms
I can't see an option in the INI file for AC to tweak the pre-rendered frames - do you know where I should look for that @RasmusP ?
I do plan to play with RTSS and indeed the Nvidia control panel - am hoping that at least one of them (RTSS I imagine) will take effect without having to close and restart the game each time.

Rtss has to be pre-configured to be running before you start a game or you'll basically crash the renderer.
But that's only once. After that you can set any limit and things to be displayed in the overlay while the game is running.

Nvidia control panel.. I'd suggest to download the latest Nvidia inspector (I can help you with that) as it gives a much cleaner and lag free interface to work with.


For Assetto corsa:
"Maximum frame latency" setting at
Steam\SteamApps\common\assettocorsa\system\cfg

0 or -1 is the default afaik.. I think 1 should equal to 1 frame... I can't remember as I'm using the Nvidia inspector setting...

But Google for that line and you'll find out

 

[Neilski]

Nvidia control panel.. I'd suggest to download the latest Nvidia inspector (I can help you with that) as it gives a much cleaner and lag free interface to work with.

Yeah I might have to ask for some help on that - I just googled it and to my surprise Google offered no links to any Nvidia websites - is it not made by them?

For Assetto corsa:
"Maximum frame latency" setting at
Steam\SteamApps\common\assettocorsa\system\cfg

Aha, thanks. But darn it, I had searched all of the configs under "Documents\Assetto Corsa\cfg" and found some graphics settings but nothing on this. (Why the heck can't Kunos be a bit more sane/consistent about per-user vs. per-machine stuff on the configs... )

In the meantime, I will be retrying my test with "1" for the max frame latency in that file as soon as I get a moment

 

[RasmusP]

Yeah I might have to ask for some help on that - I just googled it and to my surprise Google offered no links to any Nvidia websites - is it not made by them?
Aha, thanks. But darn it, I had searched all of the configs under "Documents\Assetto Corsa\cfg" and found some graphics settings but nothing on this. (Why the heck can't Kunos be a bit more sane/consistent about per-user vs. per-machine stuff on the configs... )

In the meantime, I will be retrying my test with "1" for the max frame latency in that file as soon as I get a moment

Good question about the inspector.. Afaik it was part of an overclocking tool once but Nvidia stopped development or something.
Some knowledgeable people continued on it as they liked it.

It doesn't do much anyway. Just reads out the Nvidia settings when you start the exe, then you can change a bit more stuff than in the normal Nvidia menu and then you click on apply and it gets stored.

It's basically just notepad++ for Nvidia settings


The reason behind the ini stuff is probably updates and integrity check.
The documents folder isn't updated and isn't checked during the integrity checks.
So all settings that might screw things up too much are in the steam folder.

Don't really know why gyro, damper and frame latency have to be more secure though

 

[RasmusP]

Have to correct myself:
Apparently Nvidia has an official API for people who like to play around.
There's a free one with limited stuff and one to buy where you also sign an NDA.
One German programmer (Orbmu2k) created some tools and also the inspector.

According to him what the profile section of it does, is possible with the free one.
He also says the inspector does nothing more than using the official API to set settings.

And then there's another branch of the inspector with a flashy website with inspector in its name, cluttered with ads, no names listed or anything...

So I do trust that German guy who has put his name onto the Orbmu2k website (Christian Roderfeld) and seems to be trustworthy.
I won't use the other branch though...

OrbLog » Impressum

I have to say I'm on an old driver. Gsync doesn't like some driver versions so when I found a 100% stable one, I kept it. From time to time I test performance on my win 10 boot with the latest driver and as long as I don't see any difference, I don't update.

So I'm on 430.86 and use the inspector from the Orbmu2k site which is version 2.13.

Meaning I don't have the "ultra low latency" stuff but instead just pre-rendered frames to select a number from a list.

I needed to update the inspector at some point when nvidia changed something in their drivers. I'm not sure whether or not you need the other branch' latest version for the latest RTX driver...

 

[Case_]

Oh damn, why is RD not notifying me of new comments in this topic?

But yeah, I'm not surprised by the 100 ms lag in the slightest. I would expect it to be even higher in certain scenarios for me, based on how it feels at least. Sadly I don't have a camera capable of more than 60 fps available, so even though I've done some measurements in the past, I've basically gave up as they're not very accurate. Still, maybe I might try it with ACC again, just to see the rough ballpark.

But anyway, I came here to mention a somewhat different thing which sadly doesn't reduce input lag, but it improves my framerate by roughly 25%, which for me is huge and means the difference between struggling to maintain 60 fps with 15 AI cars to being able to hold it pretty consistently with 25 AI cars.

See, I wasn't aware of it and it's not mentioned a lot, but you can run ACC in DX12. It's actually very simple - you just start the game using the "-dx12" switch.

Now from the little I've seen this discussed, the developers actively discouraged people from using this (and seemed very annoyed by it being even mentioned). Apparently the graphical components and functions are not all DX12 compatible and the performance boost is simply due to "just disabling a select and arbitrary number of incompatible graphical components and effects", to quote exactly, and is apparently "as scientific as closing your eyes and assuming the world has disappeared".

Well, that might very well be. I'm not disputing that at all. But what I can tell you with certainty is that I spent several hours today testing the DX12 version and comparing screenshots between DX11 and DX12 back and forth, and I just can't see any real difference. The only difference I see is that the car dashboards don't seem to work in DX12 - which is easily solvable by turning on the dashboard overlay in HUD.

Other than that, I really don't see any difference, and I also didn't experience any stability issues or any other issues whatsoever. Just smooth sailing with the in-game CPU bottleneck being noticeably lowered for me. I've tried every scenario I could think of, including night and/or rain...everything ran perfectly normal and fine.

So as far as I'm concerned, if this is broken, then I certainly hope it will stay broken like this in the future. An improvement of 25% in exchange of disabled dashboard really seems like a no-brainer to me. And you can even combine this with the performance improvement of the bonnet cam, and get another say 10-15% improvement on top of that (which in my case means I'm finally starting to run into GPU bottlenecks).

But obviously, since this is unofficial - your results might be completely different, should you choose to test it. TRY AT YOUR OWN RISK!

 

[Neilski]

So... I finally decided to make the measurement today, on AC (not ACC), using the "both steering wheels in view" fashion as @slatanek proposed, along with my 960 fps phone camera which sadly doesn't possess any kind of motion triggering (the numlock=downshift idea from @gyrtenudre is also pretty cool and I may play with that another time).
My monitor is a bog-standard 60 Hz beast and I use vsync. I was expecting something in the ballpark of 3 frames worth of input lag (thus maybe 50 ms).
I was very bummed to find that it's more like 100 ms
I can't see an option in the INI file for AC to tweak the pre-rendered frames - do you know where I should look for that @RasmusP ?
I do plan to play with RTSS and indeed the Nvidia control panel - am hoping that at least one of them (RTSS I imagine) will take effect without having to close and restart the game each time.

Update: I played with the settings and made some more measurements. Apologies for the lengthy post below, but...
tl;dr version: I'm now a happy bunny.

It's still a bit of a PITA to measure the input lag, so I wouldn't say that I have extensive and high-quality data, but I think the results are at least fit for my purpose (When your camera only records for 184 ms and an event can last for half of that, it's surprisingly difficult to nail the timing!)

I used @gyrtenudre's numlock LED trick and watched the downshifts in the HUD, and it worked beautifully! I've realised that I normally have three gear indicators on screen (ROFL) with my normal HUD - the circular one, the one in the essentials app, and Sidekick. The first of those was always the most responsive - often the other two updated one or even two monitor frames later (tho occasionally all three changed together). For one set of tests I also had the in-car dashboard gear indicator visible, and it matched the circular one. I also learned along the way something I should probably have realised years ago: LCD panels still update from top to bottom, just like an old CRT...

I only made one successful measurement with my original setup: vsync on, MAXIMUM_FRAME_LATENCY=0. Input lag came out as 92 ms, which is consistent with my original wheel-movement measurement. Given that I have a 60 Hz monitor, I've convinced myself that every input-lag measurement I made must intrinsically have a +/- 8 ms variation (since my button-pressing isn't synced with the display in any way). So a bunch of measurements would be required to map this out but yeah, life's too short and that lag is way too long to ever go back there!

With MAXIMUM_FRAME_LATENCY=1, input lag was 50 - 60 ms (3 meas. made). Big improvement, almost for free...
MAXIMUM_FRAME_LATENCY=1, vsync off and RTSS scanline sync=1000: lag was 34 - 40 ms (just 2 meas.). At this point I was all "yayy!!" and started to drive, to see if I could feel the difference. Placebo or not, I feel I was more consistent, catching slides way more easily, keeping the car much closer to the limit... Awesome.
MAXIMUM_FRAME_LATENCY=1, vsync off and RTSS sync off, so ~200 fps: lag was 17 - 32 ms (4 meas.). Even better, but maybe not worth the noise the GPU fans started to make

I also made some further measurements with other scanline sync positions: -1, 1100, 1. (I now realise that -1 and 1100 are probably about the same on my 1080p screen.) They were all basically indistinguishable with my sparse data, and every measurement fell into the range 34 - 47 ms.

So... mind blown. I still can't believe that my default setup was so horrifically awful. It makes no sense to me that AC would do this out of the box. Perhaps ACC does the same - will try that another time (I rarely use it right now).

Btw, as a sanity check, I also tried driving with vsync on and MAXIMUM_FRAME_LATENCY=HUGE (like 30 or something) and as I suspected, the FFB felt stupidly laggy. I know of no good reason (let me know if you do) for the physics and FFB to be in sync with the graphics, but it seems clear that they are...

Needless to say I'm driving now with RTSS active. I've beaten a few of my PBs already and I can't get over how much easier it is to catch slides.

Thanks, everyone, for the suggestions, questions, and prompting - if I hadn't gone back and shown myself how bad my setup truly was, I'd still be driving it (and cursing as the aliens lapped me again!)

 

[RasmusP]

Update: I played with the settings and made some more measurements. Apologies for the lengthy post below, but...
tl;dr version: I'm now a happy bunny.

It's still a bit of a PITA to measure the input lag, so I wouldn't say that I have extensive and high-quality data, but I think the results are at least fit for my purpose (When your camera only records for 184 ms and an event can last for half of that, it's surprisingly difficult to nail the timing!)

I used @gyrtenudre's numlock LED trick and watched the downshifts in the HUD, and it worked beautifully! I've realised that I normally have three gear indicators on screen (ROFL) with my normal HUD - the circular one, the one in the essentials app, and Sidekick. The first of those was always the most responsive - often the other two updated one or even two monitor frames later (tho occasionally all three changed together). For one set of tests I also had the in-car dashboard gear indicator visible, and it matched the circular one. I also learned along the way something I should probably have realised years ago: LCD panels still update from top to bottom, just like an old CRT...

I only made one successful measurement with my original setup: vsync on, MAXIMUM_FRAME_LATENCY=0. Input lag came out as 92 ms, which is consistent with my original wheel-movement measurement. Given that I have a 60 Hz monitor, I've convinced myself that every input-lag measurement I made must intrinsically have a +/- 8 ms variation (since my button-pressing isn't synced with the display in any way). So a bunch of measurements would be required to map this out but yeah, life's too short and that lag is way too long to ever go back there!

With MAXIMUM_FRAME_LATENCY=1, input lag was 50 - 60 ms (3 meas. made). Big improvement, almost for free...
MAXIMUM_FRAME_LATENCY=1, vsync off and RTSS scanline sync=1000: lag was 34 - 40 ms (just 2 meas.). At this point I was all "yayy!!" and started to drive, to see if I could feel the difference. Placebo or not, I feel I was more consistent, catching slides way more easily, keeping the car much closer to the limit... Awesome.
MAXIMUM_FRAME_LATENCY=1, vsync off and RTSS sync off, so ~200 fps: lag was 17 - 32 ms (4 meas.). Even better, but maybe not worth the noise the GPU fans started to make

I also made some further measurements with other scanline sync positions: -1, 1100, 1. (I now realise that -1 and 1100 are probably about the same on my 1080p screen.) They were all basically indistinguishable with my sparse data, and every measurement fell into the range 34 - 47 ms.

So... mind blown. I still can't believe that my default setup was so horrifically awful. It makes no sense to me that AC would do this out of the box. Perhaps ACC does the same - will try that another time (I rarely use it right now).

Btw, as a sanity check, I also tried driving with vsync on and MAXIMUM_FRAME_LATENCY=HUGE (like 30 or something) and as I suspected, the FFB felt stupidly laggy. I know of no good reason (let me know if you do) for the physics and FFB to be in sync with the graphics, but it seems clear that they are...

Needless to say I'm driving now with RTSS active. I've beaten a few of my PBs already and I can't get over how much easier it is to catch slides.

Thanks, everyone, for the suggestions, questions, and prompting - if I hadn't gone back and shown myself how bad my setup truly was, I'd still be driving it (and cursing as the aliens lapped me again!)

Awesome!
Sadly you didn't test vsync + rtss at 0.01-0.3 fps below your true Hz.
Scanline sync still costs a lot of resources, doesn't it?

That's why I never used it...

I mean you're talking about having 200 fps. My fps drop below 60 at race starts (cpu limit) and my GPU, due to my new monitor is between 60-80%...

 

[Case_]

He's talking about AC, though, not ACC, right? (But yeah, I also missed that in the previous post about 100 ms latency.)

BTW I'm not aware about Scanline sync costing a lot of resources. At least with ScanFlush=1 it shouldn't, but even without it doesn't seem to. I use it pretty much in every sim nowadays.

 

[gyrtenudre]

So... mind blown. I still can't believe that my default setup was so horrifically awful. It makes no sense to me that AC would do this out of the box. Perhaps ACC does the same - will try that another time (I rarely use it right now).

Exactly my feelings with ACC and scanline sync. According to my measurements, which I did 8 per configuration, there's a gain of roughly 30ms in ACC when using scanline, which is pretty close to what you measured too. And I also became more consistent in my driving, I doubt it was placebo. The mind anticipates delay on all the things you expect to happen, but then again most of those slides are not exactly expected, so the 30ms gain is welcome.

Sadly you didn't test vsync + rtss at 0.01-0.3 fps below your true Hz.
Scanline sync still costs a lot of resources, doesn't it?

The -0.01 fps cap I tested in ACC and it produces variable input lag unfortunately (my results should be somewhere in this thread). I would expect the same regardless of the game, but hey, that's just my understanding of it. Scanline shouldn't cost more, if anything is should cost a tiny bit less as it eliminates (at least) some of the buffering and the overhead it comes with (swapping buffers and I don't know what else). Indeed I see the exact same GPU load with either config as long as it has some form of capping (vsync with limiter, scanline, just limiter etc) and it makes sense, the GPU is still rendering 60 frames / sec (unless you're using vsync with triple-buffering and the GPU can render 2xRefreshRate or more).

One other interesting thing to note about scanline sync: when using DSR, it uses the native resolution height, which is ofc lower if DSR > x1, and you can't push the tear line in the non-visible area. (I know there is in-game super sampling to replace DSR, but I prefer DSR visually on ACC). Anyways, luckily enough, a large portion of the lower screen is the dashboard with the cockpit camera, which doesn't really move so I put the tear line on the dashboard and it disappears automagically.

Edit: The only thing I'm not happy about my setup right now is a few micro-stutters I get, around 4-5, and during my outlap only. Then it's smooth as butter . But I suspect this is ACC, scanline shouldn't cause stuttering. If anyone is using ACC+scanline sync I would love to hear if they have the same issue.

 

[RasmusP]

automagically

Beautiful new word for my vocabulary

I've read about scanline sync causing tearing and stuttering when having more than 60-70% gpu load.
Also it's coordinated by the CPU isn't it? So when I'm hitting my CPU limit, would it cause stuttering or tearing?

Guess I'm gonna test it! Probably in acc as with my hardware I can switch from CPU limit to gpu limit with just a few clicks