Nvidia RTX DLSS/Ray Tracing Discussion

[Exposed]

Well why didn't you post your thoughts up? That's why I stopped no one added anything of substance for me to go through all that again. Besides you don't think I know you're trawling now with that response?

Tell you what, I can do the same for Metro. Only if serious though.

 

[Nunz]

Well you can't run Metro Exodus with RTX at custom 1800p for one.

RTX doesn't work with the resolution scaling available in Metro options? Because HWU seems to have it working fine. Custom resolutions are a thing of the past, almost every modern game engine is offering resolution scaling as part of graphic options.

Two, DLSS isn't about just upscaling. It's taking a 1440p image, taking a 4k image, comparing the two and brute force learning how to approximate the 4k image. That's different than just 1800p upscaling and has the potential to offer better performance and better IQ. You can already see this in Port Royal, DLSS looks better than 1800p upscaled. So this has the potential to transfer over to games. THAT is the point of DLSS and something much more promising than the simple upscaling you've been doing for years. We haven't even seen DLSS 2x yet which takes the native resolution for AI learning.

It has the potential, yes. DLSS may look better in Port Royale but that is a scenario that is exactly the same every time. No game will ever offer that and it's impossible for DLSS to perform the same way it does in a synthetic benchmark as actual real-world gaming. DLSS 2x may be the way of the future, sure.

You keep saying the 1800p videos look better, but that's not what HW concluded with. They said it was on par.

Weren't you just complaining that people are just parroting what the clickbait videos are saying? Now I'm wrong for having a different opinion than those hater videos come on. Looking at the video itself and taking some time to really look at the comparisons, the 1800p slides were flat out better. Added with some sharpening, and now we're really talking. Maybe it has to do with Youtube quality, that could certainly be it. I will try the scaling myself next time I'm home so I can do my own testing on it, and I'll post back.

Like I and others have said before, 4k DLSS is offers a great performance boost at minor detail loss. My specific words were: I thought I'd run into performance issues with RTX Ultra/ Advanced Physics/Hairworks on at 4k but so far I'm surprisingly locked at 60fps. The minor detail lost with DLSS is not noticeable on my 65" screen unless you walk right up to the screen to look at intricate cloth detail you won't see at 3-5 feet anyway.

I mean, at that screen size, and that distance, I have a hard time believing you'd find much issue with the image even with MLAA or some other blur-ridden form of AA. You're kind of the best scenario for DLSS - the detail you're losing with DLSS on @ 4K isn't noticeable because you're too far from the screen to see it even when it's there. For those on the ~30" monitors @ 4K, who are sitting at a normal viewing distance .. I'm not sure it's the same case scenario for them. Would love to see a pic of your setup

Even then it's not even a blur, its as sharp as you would expect 1440p upscaled to 4k to be, with the benefit of virtually zero aliasing present.

I'd agree it's less of a blur, especially post-patch in Metro. It feels more like ... a smudge, I suppose. DLSS seems to be more effective on smaller resolutions (referring to 1440p, not sure about 1080p?) than it does at 4K. 1440p w/ DLSS looks like actual antialiasing, with no blur added; it's impressive, if anything DLSS is adding sharpness, rather than taking it away as it appears to @ 4K.

Have you tested DLSS on a native 4k screen?

No. I've only tried it @ 3440x1440, which it does seem to do a decent job; at least in terms of clarity post-patch, it gives a sharper image than what DLSS seems to be doing @ 4K per videos.

I'm not trying to bash DLSS or make it out to be a useless feature. DLSS 2x certainly will be impressive. My only argument is that DLSS is being pushed out in a form that doesn't really do much that we didn't already have available to us in terms of resolution scaling. I know you have said that custom resolutions break RTX, but using the in-game resolution scaler isn't breaking RTX, at least from the HWU video. Does BF5 react differently? If so, that's more an issue with Frostbite/DICE.

 

[logical]

Nvidia RTX DLSS/Ray Tracing Discussion

Lets talk about the implementation and worth of DLSS and Ray Tracing!

 

[Nunz]

There is a 20+ PFS boost and in certain areas it's badly needed, especially with tessellation turned on.

Strange! I was seeing really good numbers @ 3440x1440 w/ everything maxed, DLSS off. Maybe it was the area I was in? It was still early in the game, but outdoors. On my setup @ 1440p (can't compare w/ you because TITAN X (Pascal)) I'm still pretty early in the game as well and I'm capped at 82 usually with some drops into the 70s at times. DX11 performs better than DX12, as usual though .. lol

 

[acroig]

Strange! I was seeing really good numbers @ 3440x1440 w/ everything maxed, DLSS off. Maybe it was the area I was in? It was still early in the game, but outdoors. On my setup @ 1440p (can't compare w/ you because TITAN X (Pascal)) I'm still pretty early in the game as well and I'm capped at 82 usually with some drops into the 70s at times. DX11 performs better than DX12, as usual though .. lol

There are certain portions in the desert area and others where the FPS tanked to the upper 40's at max without DLSS on.

 

[Exposed]

Oh hi gais!

 

[Seyiji]

Oh hi gais!

Oi wots going on is dis thread...

Well why didn't you post your thoughts up? That's why I stopped no one added anything of substance for me to go through all that again. Besides you don't think I know you're trawling now with that response?

Tell you what, I can do the same for Metro. Only if serious though.

Like I said in my reply I was dealing with family issues at that time.

Also no trolling just 15+ years of being with Nvidia clouding my resolve

 

[Nunz]

There are certain portions in the desert area and others where the FPS tanked to the upper 40's at max without DLSS on.

Gotcha! Can’t wait to see those areas. Not sure why but I’ve been playing small amounts at a time. The game just hasn’t held my attention and really sucked me in the way the other two did.

 

[Exposed]

RTX doesn't work with the resolution scaling available in Metro options? Because HWU seems to have it working fine. Custom resolutions are a thing of the past, almost every modern game engine is offering resolution scaling as part of graphic options.

Metro won't allow RTX/DLSS together outside of their supported resolutions. Thats why their comparisons were DLSS only.

It has the potential, yes. DLSS may look better in Port Royale but that is a scenario that is exactly the same every time. No game will ever offer that and it's impossible for DLSS to perform the same way it does in a synthetic benchmark as actual real-world gaming. DLSS 2x may be the way of the future, sure.

For Port Royale, they got DLSS working up to snuff. With a full game, they are going to need alot more processing time.

Keep in mind regular AI upscaling on static images (nothing to do with Nvidia) takes hours/ DAYS with supercomputing. So for a full game that will be ambitious. That's why we will see further patches as the months go by for DLSS with more and more training time and probably why we saw **** results at first because they wanted to get something out.

Remember the neural network isn't running a single algorithm. Otherwise that would just be regular AA. The reason DLSS runs on tensor cores because it needs to run the AI for whatever area/scene you're in.

Weren't you just complaining that people are just parroting what the clickbait videos are saying? Now I'm wrong for having a different opinion than those hater videos come on. Looking at the video itself and taking some time to really look at the comparisons, the 1800p slides were flat out better. Added with some sharpening, and now we're really talking. Maybe it has to do with Youtube quality, that could certainly be it. I will try the scaling myself next time I'm home so I can do my own testing on it, and I'll post back.

I don't need to look at youtube videos, I have the game running on my display at 4k. 4k DLSS is superior to native 1440p with RTX on. I don't need to run a custom resolution if 4k DLSS is on par. I'm sure the next DLSS patch will make the image quality even better than it is now.

As for sharpening, well that can be done for any game. Including games already running DLSS, as SirPauly pointed out prior. Which incidentally proved DLSS isn't applying a sharpening filter by the way, but actually running a more robust upscaling implementation.

I mean, at that screen size, and that distance, I have a hard time believing you'd find much issue with the image even with MLAA or some other blur-ridden form of AA. You're kind of the best scenario for DLSS - the detail you're losing with DLSS on @ 4K isn't noticeable because you're too far from the screen to see it even when it's there. For those on the ~30" monitors @ 4K, who are sitting at a normal viewing distance .. I'm not sure it's the same case scenario for them. Would love to see a pic of your setup

This is actually the best case for 4k. At 65 inch native 4k I can see intricate 4k details and all AA flaws.

What I'm seeing in motion is that 4k DLSS, even though I know is 1440p upscaled, isn't losing alot of detail compared to 4k native. I had the opportunity to start the game after the post DLSS patch and the first thing I did was check 4k native vs 4k DLSS. Yes 4k native is obviously better, but with advanced physics/RTX/tesselation/hairworks, the 4k performance just slacking a bit. DLSS gave me 60fps locked with much better anti-aliasing and very minor lost of acute details, details not visible unless you put your nose almost right up to the screen, like the cloth detail zoomed in screenshots Seyiji posted earlier.

It's not just me, it's Bill also. And Acroig for his 1440 case. The difference is a molehill, not a mountain.

I'd agree it's less of a blur, especially post-patch in Metro. It feels more like ... a smudge, I suppose. DLSS seems to be more effective on smaller resolutions (referring to 1440p, not sure about 1080p?) than it does at 4K. 1440p w/ DLSS looks like actual antialiasing, with no blur added; it's impressive, if anything DLSS is adding sharpness, rather than taking it away as it appears to @ 4K.

Any blur should be less noticeable at 4k, simply for the fact there's more pixels to work with. This again is something you would need to see in motion, on an actual native 4k screen.

No. I've only tried it @ 3440x1440, which it does seem to do a decent job; at least in terms of clarity post-patch, it gives a sharper image than what DLSS seems to be doing @ 4K per videos.

That's still only 1440p widescreen. Essentially for me to do the same test (which I did), I would need to run 1440p DLSS on my setup. I did that, I could notice a drop off from 4k DLSS. I'm not using a cheap display either, I'm on a high end LG which is better than most computer monitors out there. I don't even think there's a Gsync monitor out yet with the same rich color gamut/HDR.

I'm not trying to bash DLSS or make it out to be a useless feature. DLSS 2x certainly will be impressive. My only argument is that DLSS is being pushed out in a form that doesn't really do much that we didn't already have available to us in terms of resolution scaling. I know you have said that custom resolutions break RTX, but using the in-game resolution scaler isn't breaking RTX, at least from the HWU video. Does BF5 react differently? If so, that's more an issue with Frostbite/DICE.

That would be a question for the developers. I don't know why RTX/DLSS is stringent on resolution, but it kind of makes sense. If traditional image upscaling takes hours/days, then imagine if you "lowered" the resolution on that same image...you'd have to run the upscaling AI network on that image all over again! Presumably similar restrictions would be in place for DLSS.

 

[acroig]

Gotcha! Can’t wait to see those areas. Not sure why but I’ve been playing small amounts at a time. The game just hasn’t held my attention and really sucked me in the way the other two did.

I really liked it. Same Metro formula but enjoyed the different play areas.

 

[Nunz]

Metro won't allow RTX/DLSS together outside of their supported resolutions. Thats why their comparisons were DLSS only.

Busy atm and will expand later, but they have slides showing 0.8 resolution scale (the 1800p comparisons iirc, can't watch to double-check right now) with RTX on. I didn't listen to the guy talking in the video because I couldn't give a **** less what they have to say, I just wanted to see their slides and the image comparisons. If I'm wrong please correct me.

 

[demo]

Honestly, Metro dlss update just looks like they added a sharpening filter with outlines around everything.

 

[Nagorak]

I think DLSS just isn't going to work. It was an inappropriate use of machine learning. There's too much variation in the image data that's generated by a game and it just ends up averaging everything together. It can't come up with some magic filter that can be applied for a better result because it simply doesn't exist.

For some things AI is very effective, for many other things it is absolute crap. In the latter cases there's still no substitute for human intelligence. I think that what DLSS is showing us is that the upscaling and AA techniques that already exist are actually pretty good. There wasn't much hidden room for improvement.

On the other hand, Nvidia has done a great job with its marketing to convince people to simply accept upscaling. So, maybe there is some good in that.

Edit: Also I see the argument that "DLSS isn't just upscaling, it's supposed to do X". It doesn't matter what it's supposed to do. It matters what it actually does. If the technique isn't effective, which is what I suspect, then it really makes no difference if they tried to match an image running in 1,000,000P, it's still not going to work. The truth is sometimes you can crunch a lot of numbers and end up with nothing much at the end...it happens.

 

[Exposed]

Busy atm and will expand later, but they have slides showing 0.8 resolution scale (the 1800p comparisons iirc, can't watch to double-check right now) with RTX on. I didn't listen to the guy talking in the video because I couldn't give a **** less what they have to say, I just wanted to see their slides and the image comparisons. If I'm wrong please correct me.

That's the shading rate, which is the number of pixels used for a shading operation. 1.0 is 1 to 1 pixel ratio. Less looks worse, more looks better. This is different than running in a lower native resolution, and different from other games' in game resolution scalars. It's called a shading rate and not a resolution scalar for a reason...yet another reason to not listen what they have to say.


Nvidia has their own implementation on Turing:


https://developer.nvidia.com/vrworks/graphics/variablerateshading

 

[Exposed]

I think DLSS just isn't going to work. It was an inappropriate use of machine learning. There's too much variation in the image data that's generated by a game and it just ends up averaging everything together. It can't come up with some magic filter that can be applied for a better result because it simply doesn't exist.

For some things AI is very effective, for many other things it is absolute crap. In the latter cases there's still no substitute for human intelligence. I think that what DLSS is showing us is that the upscaling and AA techniques that already exist are actually pretty good. There wasn't much hidden room for improvement.

On the other hand, Nvidia has done a great job with its marketing to convince people to simply accept upscaling. So, maybe there is some good in that.

Edit: Also I see the argument that "DLSS isn't just upscaling, it's supposed to do X". It doesn't matter what it's supposed to do. It matters what it actually does. If the technique isn't effective, which is what I suspect, then it really makes no difference if they tried to match an image running in 1,000,000P, it's still not going to work. The truth is sometimes you can crunch a lot of numbers and end up with nothing much at the end...it happens.

It's pretty darn effective in Port Royale and a good indication of what a properly implemented DLSS solution will do. But yeah, it could be that for now, games are just way too big and varying in terms of graphics and settings for initial DLSS to be effective.

Maybe we'll see better results with a smaller game like Atomic Heart. Though, the current implementation of Metro is pretty darn good for what it gives us. I'm sure we'll see the same for Battlefield 5 in the next patch they promised.

 

[acroig]

Though, the current implementation of Metro is pretty darn good for what it gives us.

Agreed.

 

[SIrPauly]

Don't desire 4k dlss to look like native 4k, but superior while moving. The key is the training image is x64 jittered super-sampled, so the strength, one may imagine is quality with moving with less aa limitations. Needs time and maturity.

 

[NWR_Midnight]

It's fun to watch all the RTX 2080ti owners fall for a placebo effect to substantiate their expensive purchase.

First, Battlefield V and metro do not fully implement Raytracing. They are only using it partially, as neither game would be playable at 4k if it was fully implemented, that is why it is only viable at 1080P if fully utilized. Removing and/or reducing what the game is ray tracing doesn't remove that fact.

Second, DLSS and raytracing at 4k is a not 4k, it is a lower resolution (1440p) upscaled, which means the ray tracing (only partially being implemented) and AA is being applied to the lower image and then being upscaled (I originally thought ray tracing was being done after DLSS upscaling, but realized that to gain frame rate with DLSS, it can only be accomplished by also performing ray tracing at the lower resolution before it is upscaled). This means that it will never look as good as a true native 4k implementation. Basically it's the same as a HD 4k blue ray player that upscales standard Blu ray movies to 4k, they may look a bit better than the original resolution, but it will never be true 4k. (infact, if you ran a game at 1440p on a 4k monitor, doens't the monitor already upscale it to display it at full screen on a the 4k monitor?.. I know my 1440p monitor does it for 1080p... does that mean all games are now 4k.... oh wait)

As for the video's that are showing the comparisons side by side being only at 1080p doesn't change anything. Because going from 4k to 1080p in a video will equally effect both the 4k DLSS example and the Native 4k example, so if the video's where indeed encoded and watchable at 4k resolution, the clarity differences noted would be identical to what we see in the 1080p videos as they are both effected equally as it wouldn't just effect the DLSS examples.

But, hey, keep believing in the placebo effect, if it makes you happy and feel better for spending $1300 on a GPU.

But right now, this is just smoke and mirrors in it's current form. Will it get better, Yes, at least where Ray tracing is concerned. But DLSS, in it's current form, is really just a fancy name for upscaling more so than AA.

 

[kboye]

Scaling up and down, basics still apply. Some material can scale both directions, with unsharp-style sharpening filter to clean it up, and still look very, very good. Film/photography scale well being "analogue" in nature, but a pixel sharp digital render not so much when comparisons are made.

If the scene is so heavily processed there's barely any contrast, or the scene has toon-shader style graphics where texture isn't grainy enough, scaling can look very good when sharpened up slightly. Keep text native and GUI in native res, and it might not be so easy to spot any difference.

But a "noisy" (as in lots of pixel detail, not just diffuse colorwith some shading) game like Metro, ....I think I'd keep the res native and turn down settings instead. I'd do anything to keep the game running at native res.

Edit: just hit me, of course it's a personal preference. It's good to have options!

 

[Exposed]

It's fun to watch all the RTX 2080ti owners fall for a placebo effect to substantiate their expensive purchase.

First, Battlefield V and metro do not fully implement Raytracing. They are only using it partially, as neither game would be playable at 4k if it was fully implemented, that is why it is only viable at 1080P if fully utilized. Removing and/or reducing what the game is ray tracing doesn't remove that fact.

Second, DLSS and raytracing at 4k is a not 4k, it is a lower resolution (1440p) upscaled, which means the ray tracing (only partially being implemented) and AA is being applied to the lower image and then being upscaled (I originally thought ray tracing was being done after DLSS upscaling, but realized that to gain frame rate with DLSS, it can only be accomplished by also performing ray tracing at the lower resolution before it is upscaled). This means that it will never look as good as a true native 4k implementation. Basically it's the same as a HD 4k blue ray player that upscales standard Blu ray movies to 4k, they may look a bit better than the original resolution, but it will never be true 4k. (infact, if you ran a game at 1440p on a 4k monitor, doens't the monitor already upscale it to display it at full screen on a the 4k monitor?.. I know my 1440p monitor does it for 1080p... does that mean all games are now 4k.... oh wait)

As for the video's that are showing the comparisons side by side being only at 1080p doesn't change anything. Because going from 4k to 1080p in a video will equally effect both the 4k DLSS example and the Native 4k example, so if the video's where indeed encoded and watchable at 4k resolution, the clarity differences noted would be identical to what we see in the 1080p videos as they are both effected equally as it wouldn't just effect the DLSS examples.

But, hey, keep believing in the placebo effect, if it makes you happy and feel better for spending $1300 on a GPU.

But right now, this is just smoke and mirrors in it's current form. Will it get better, Yes, at least where Ray tracing is concerned. But DLSS, in it's current form, is really just a fancy name for upscaling more so than AA.

We already know no game is "fully ray traced", that's impossible with current hardware. "Fully ray traced" would be near unlimited rays with near unlimited light bounces.

You must be referring to the performance gained by Battlefield 5. If so, the only thing they removed was the back and forth light bouncing from objects that didn't need to be ray traced, like leaves. That and further bug fixes and optimizations that resulted in much better performance without sacrificing image quality. This is old news and discussed in the Battlefield 5 thread numerous times already.

Also I'm glad you pointed out 4k DLSS is not native 4k. None of us knew that before.

Also keep up that tone, I'm sure a vacation will be in your future if you keep that up!