DCS 2.5 Antialiasing comparison

[Repsol2k]

Do u played around with the AISuite? Or all for ur 1080ti on standard?

 

I noticed when i set my AISuite only the Power Target to 120% and the GPU Voltage to 100% and then i close AISuite i got a really good performance.

 

Set my Nvidia Settings to

 

Antialiasing Mode : On

Setting: 8xQ

Transparency Supersampling to 2x

 

Everythings looks good when i turned off MSAA and AF Ingame. But this for non VR with VR i need to test

[StrongHarm]

After much experimentation and points brought out in this thread, These are my grail settings. I attached both pages of the nvidia CP.

 

 

AMD Ryzen 5, ASUS Rog Strix 1080ti 11G, 16G ram, Crucial SSD, windows 10.

 

 

Pixel density is 1.5.

 

Solid 45fps everywhere, even in Vegas, even when I'm throwing the Gazelle around like a go-kart. I can't say how or why it's working for me, but I'm very happy with it.

 

fargo007, be aware that nVidia control panel (or inspector) settings have no effect on VR. If you bring your PD down to 1.2 you can raise your other settings and it will look better and perform better. Full analysis.

[fargo007]

Strongharm,

 

I disagree with the blanket statement. There are some settings in there that do have an effect. Not all, but some. I can tell the difference.

 

I have it set to 1.5, which is a very apparent difference for me from 1.2, and have some settings higher than yours.

 

"Look better" is also subjective. At the end of the day we all have to find what we like. You posted what works for you, this is what works for me. I also fly helos almost exclusively if that matters.

[StrongHarm]

fargo007 you're right, preferences are no doubt subjective. My intention was to save you some time with that analysis. You're the first person to report such a departure from the results. Different hardware builds do create unseen variables. My apologies for insinuating otherwise.

 

Settings in the nVidia control panel and inspector don't effect VR in DCS, and this should not be subjective based upon hardware. As far as I know, and it's been verified by many others; the only setting that will effect DCS (minimal) is 'Power Management Mode'. If you've found a way around it and have enhanced DCS through the NCP or NI, it would be really helpful if you could write up how you did it.

[fargo007]

If the power management mode does have some effect, I don't understand where the absolute conclusion comes from that maintains that is the only setting in the NVCP that has some effect. I apologize if that's been verified by someone with low level knowledge of the DCS rendering code.

 

I do see differences with and without the NVCP settings I posted. I'm going to leave it as-is and try the 1.2 PD with your settings and report back. Who knows? I may find something I like.

[StrongHarm]

My analysis isn't gospel, it's just an attempt at being helpful. I hope you find something useful.

https://forums.eagle.ru/showthread.php?p=3382525#post3382525

[Alf.Snake]

The game need optimization! because the fps dronw HEAVY!!!

[Burning Bridges]

I'm having no luck at all with DCS, I'm lucky that my hardware copes easily as far as FPS go, 8700k and a 1080Ti. Setting MSAA in game to 2x or 8x has no FPS effect.

 

The ingame MSAA is the only thing I can get to control the visual effects.

8x looks good when the aircraft is sitting still, but as soon as I taxi, the lines are shimmering and flickering, the TAD map has flickering lines, runway lines are flickering and shimmering, it looks terrible.

 

The strange this is, I can not get NVCP or Inspector to do anything to my ingame settings. I have tried all combinations, but nothing.

 

Tried new profiles, with DCS launcher as the exe and run as the exe. I tried enhance, override, FXAA on and off, SGSS, but it always looks the same.

 

One trick that I think should work is to leave vsync off in the game, then set it to 1/2 or 1/3 refresh from NVCP or Inspector, then if I check my FPS in game, I can see that the game is taking commands from one of the programs, but in my case, it never does, and FPS is not controlled by external vsync.

 

In the past, any gaming was just a case of override or enhance, supervcaa 64x for AA and get rid of the shimmering with spare grid SS 4x, but now it just wont work.

 

Lastly, if I run NVCP and Inspector profiles, which one overrides which? Do they work together or conflict with each other?

 

 

EDIT : actually, one thing I meant to try, and just pop into my head, was if Inspector profile won't work, then maybe the game is reading the Global settings, so I set some Vsync global settings, and now DCS is responding to them..so some success there, now to try some different AA settings, fingers crossed!

 

It's not only you. I can see the artifacts, the moire patterns, shimmering 3d objects and trees e5c, even in videos of youtubers, wandering over the screen like molasses.

 

People disagree with you because they don't see shit, have terrible monitors or (I suspect) just out of spite because they saw the issues years ago and simply resigned.

 

In fact I spent entire days trying and launching DCS with different settings, DSR, different smooting, AA, AF in the game and in Nvidia panel to get rid of the shimmering, more time than I spent flying. I just can't stand an image that is flimmering and full of artifacts. I kid you not, but I even tried wearing sunglasses while flying :)

 

I am currently at the point where I just give up.

[StrongHarm]

I'm just guessing here fargo, but I think that Power Management Mode determines how much juice the card can get, whereas the graphical preferences in NCP try to force hardware settings over app settings, but DCS hooks the hardware during VR invalidating overrides.

[fargo007]

Except that the drivers are at a lower layer than DCS is.

 

DCS can't make the drivers do, or stop the drivers from doing anything than what the drivers permit it to.

 

Some of the things there, like FXAA are post processing filters, meaning it occurs after DCS is finished.

 

e.g.

 

https://forums.eagle.ru/showthread.php?t=200906&highlight=fxaa+post+processing

 

DCS would have no control over any post processing filters, and the drivers certainly would.

 

Based on that, along with what I am seeing, I don't believe there is a scientific conclusion that there is nothing at all other than power mode in the NVCP that can affect the experience.

 

I did try the 1.2 PD with the settings you posted. It was good, but I definitely like the appearance of 1.5 better, along with 16x Anisotropic, Low terrain texture and no terrain object shadows (as I posted earlier). I never go below 45 this way even on the vegas strip but with the 1.2 settings in the same areas it did.

 

I am using two configs. That one, and another without def. shading and HDR on. The second one is for night missions since def. shading has issues at night (black laser beams, unreadable cockpit, etc). I just use a small batch file to copy the options.lua templates over to options.lua.

 

I had been shutting ASW off but now after reading the info you posted about it, I fully agree it's better to leave it on. That was very good info, thanks.

[StrongHarm]

I don't disagree with that fargo, hardware drivers have access through the hardware abstraction layer in the OS, where DCS has access through the engine. VR is different though. Although I've been in tech for 30yrs I can't give you a reason why NCP settings don't affect VR in DCS, but a little testing shows that it's true. I'm sorry I don't have more information.

[fargo007]

Maybe you are just not seeing/detecting it?

 

It's the blanket statement that I disagree with. There's no data to support it - it's substantiated only by subjective interpretation.

 

My testing does indicate a difference. Exactly consistent with the photos that Snacko posted. At least with FXAA, there's clearly a detectable effect for me. And as pointed out architecturally, with a post processing filter, DCS has no ability to control the application of it. e.g. Which other settings there are post processing?

 

Until we have someone with understanding of DCS's architecture at a low level comment (and this is not you or I, regardless of how many years we've been at a keyboard), we will not know for sure which changes made there have, or do not have effects.

 

Absent that, or any data that can speak for itself, I would still suggest people experiment with these settings, because they may detect or discover differences that others do not.

[StrongHarm]

I most definitely see the effects of nVidia Control Panel settings on my monitor and you're right, the enhancements are visibly obvious. When I change those same settings and view DCS through Rift I see no change to visual quality or FPS. That's conclusive. I've even tried making the changes while connected to a monitor then plugging Rift into the same port (as method does work with some settings in Elite Dangerous), but still no joy. You're asserting however, that based on an unknown variable some people can enjoy NCP settings enhancements in VR while some people can not? I guess I buy that. I have no interest in proving you wrong fargo, but I would love to prove you right. I have a huge interest in enhancing my VR vid through NCP.

 

Can anyone else test and see if you get a response in VR from NCP settings changes?

 

ED Devs, can you comment?

 

Maybe you are just not seeing/detecting it?

 

It's the blanket statement that I disagree with. There's no data to support it - it's substantiated only by subjective interpretation.

 

My testing does indicate a difference. Exactly consistent with the photos that Snacko posted. At least with FXAA, there's clearly a detectable effect for me. And as pointed out architecturally, with a post processing filter, DCS has no ability to control the application of it. e.g. Which other settings there are post processing?

 

Until we have someone with understanding of DCS's architecture at a low level comment (and this is not you or I, regardless of how many years we've been at a keyboard), we will not know for sure which changes made there have, or do not have effects.

 

Absent that, or any data that can speak for itself, I would still suggest people experiment with these settings, because they may detect or discover differences that others do not.

[fargo007]

Conclusive?.... For you.

 

That's the essence of it. You are transposing wide scientific conclusions based entirely on subjective observation.

 

e.g. "because I don't see anything in the dark, it's proven thus that nothing is there."

 

I respectfully question the method of that conclusion, but stand by you in the search for the truth of this.

 

I would also love to have some comment here from ED. If there's fruit on these trees let's pick it. If not, let's please move on.

[StrongHarm]

Yeah dude, as I said; conclusive for me. Based on tests that I thoroughly outlined, decades of tech experience, and years of test experience with this specific product. Open your DCS:A-10 manual and search for my name. I hope you'll throw me a bone. In my last post I surrendered to your tone and basically said "you're right.. lets work together and figure it out." You're the one person I've seen post that you get results from NCP in VR, and there are a lot of people who say they don't. Being the rare bird that you are, why don't you try and figure out how and share it with the community? You could start by outlining your test methods and conclusions (as I requested) and other people can try to reproduce your results.

 

Here's my 20hrs of effort on the subject: https://forums.eagle.ru/showthread.php?p=3382525

 

I try my best to be more helpful than contradictory on these boards, it's more productive for the product and the DCS community. Reddit's a good place to argue, ED boards are a good place for study sims. It's getting harder these days to tell the difference.

[fargo007]

This is my method of observation. It's quite simple:

 

1 - Turn MSAA off in game. Observe staircase effect.

 

2 - Turn FXAA on in NVCP. Observe some mitigation of it.

 

The basis for this goes back to experiments people did with 2.2, e.g.

 

 

We can't tell each other what we see and do not see. So until we have input from ED, or a scientific method to produce data that establishes effect/lack of effect (and I am not aware of any - perhaps the Oculus tools?), this can't be concluded either way.

 

I apologize if I am coming across with a tone, I do not mean to, and am only expressing a disagreement on the method used to substantiate a presented conclusion.

 

I want the community to have the widest available spectrum of configuration settings to tweak for the best possible individual experience, and believe we should leave all lanes open until we have data or an authoritative statement that says otherwise. That is the place I am coming from.

[YoYo]

This is my example with my ReShade profile for DCS 2.5:

 

[Dr_Arrow]

The problem with MFAA on my setup (Nvidia - GTX 1070), that I get a lot of shimmering, while this does not happen with MSAA, although the resulting picture quality is comparable. With MFAA the whole cities are simmering when looking at them at angle, so this is pretty bad and I have to take the massive performance hit with MSAA. Is there any way to get rid of the shimmer? Do you guys also have it?

[ebabil]

i use 2xmsaa in game + mfaa

i think it was better when OB was first released. Now both 2xmsaa and 4xmsaa brings lots of shimmering. Don't know what to do. I can't recognize what plane is in front of me from a few miles

[Deus_010101]

What is SMAA? Where is that setting?

 

I have not seen any affect from FXAA. I might have to mess with that again but using a combination of MSAA and SSAA has good results, no need for MFAA with this.

 

 

I did some comparison you can see here

 

 

https://docs.google.com/spreadsheets/d/1ov9120fk_ej_XhbvPgWrYtgDk7WTG1SzdObXBEw8vjg/edit?usp=sharing