Codemasters Integrates AMDs Image Sharpening Tech in F1 2019
Click here to post a comment for Codemasters Integrates AMDs Image Sharpening Tech in F1 2019 on our message forum
Undying
In the meantime Remedy is implemeting ray tracing support for their Control and nvidia users. Which one is better?
Denial
Both at the same time
Undying
But if you have to choose one? I know what i want more. 😉
WhiteLightning
Moderator
i dunno, one is for low end, the other for high end, so it depends.
Abc666
i would use image sharpening + higher FPS over raytracing any day with current graphich cards .
Denial
IMO in a single player game like Control I don't necessarily need fast pace framerates and would prefer a visual experience so I'd take the better image quality from RT. As long as I could hit 60fps+ with RT on (1440p) then RT for me. That being said I don't have an RTX card so doesn't matter either way.
Caesar
[youtube=a4xsU9Se0ME]
Denial
DLSS isn't enabled in the game FWIW. They accidentally enabled it before the update was issued to the driver. So it's essentially not functioning at all in that video.
http://blog.codemasters.com/f1-2019-patch-notes/
Also I'm curious why these videos never compare Nvidia's sharpening feature? I don't think its as good as AMD's but for some reason people pretend like it doesn't exist
Fox2232
I want to keep high fps while having extra details.
I definitely do not want TAA/DLSS or any other improperly implemented blur filter.
Denial
Has there even been a real comparison between CAS and DLSS? So far both video's I've seen DLSS wasn't working at all and neither Hardware Unboxed nor WADHD have fixed or even addressed it.
Caesar
So, does AMD FidelityFX Upscaling and FidelityFX Sharpening works with Nvidia Cards? The video below shows the option with RTX 2080ti 😕
UPDATE: GOt it!!!
[youtube=PVuLbL4pFQw]
Denial
This one what? Did you read what I quoted from the developer? DLSS was added as an option before it was implemented - all it does is lower the resolution but it doesn't actually apply the upscaling at all. So all these videos are comparing a lower resolution game with DLSS not running to AMD's sharpening. It's good for showing FidelityFX but it's useless for comparing to DLSS.
Edit: I see you updated your post - yeah FidelityFX works across all video cards.
Fox2232
I have seen some video comparison between CAS and Freestyle. Results looked rather similar, but impact on performance on nVidia's from Freestyle sharpening was few percent while AMD took sub 1% fps loss.
As for DLSS, there is not single implementation worth comparing because none delivers good image quality. One can at best compare it to many TAA implementations where it is not doing badly.
Cave Waverider
No idea. People tend to forget it for some reason. Maybe they don't use GeForce Experience?
Igor's Lab did include it in a comparison test video to augment their written article a while ago, but of course it didn't include F1 2019.
schmidtbag
I would vote raytracing. If a sharper image and better framerate are my priority, I'd rather just turn off or reduce AA. When implemented well (which so far, is rare), raytracing gives a better sense of realism.
I would say it's pretty obvious. DLSS is worse in every regard:
* It's proprietary
* It requires RTX cards
* It only works on a select few games
* It doesn't look as nice
I'm not sure what DLSS' impact is on performance.
Sure, where possible, comparing to DLSS would definitely be nice, but it seems to me it's already a dead technology.
I think it's a little weird how Codemasters acts like they did something to make AMD's image sharpening work, since to my recollection, it works on anything via a simple toggle. But, perhaps they meant that they got it optimized, so it looks better than it otherwise would have.
Denial
I was talking about Nvidia Freestyle - which, while proprietary, works across Nvidia's entire product stack in hundreds of games.
That being said I don't fully agree with your assessment of DLSS. It's proprietary but there is no machine learning framework for gaming.. Microsoft is building one with DirectML but it just started shipping a few months ago - so it's not like Nvidia could have made it non-proprietary even if they wanted. Regardless, while DLSS as a brand might be dead the idea of using AI upscaling is just starting. As the neural nets get better, image quality will increase and as the hardware performance improves it will be faster/easier to train games. The current results with BF5 with DLSS are extremely impressive in my opinion and it annoys me to no end that reviewers are either doing what they did above (demonstrating DLSS in a non working state) or simply just writing the technology off because the image quality was less than ideal at the beginning. Every single DLSS game that's out has received multiple updates to improve image quality and yet I haven't seen a single reviewer go back and test the image quality. The best thing we got was Hardware Unboxed's video of BF5 and the above videos with F1 where everyone concludes "DLSS is bad" but it isn't even enabled lol.
I think the technology of AI upscaling will remain and only continue to improve. I'd like to track the progress of it and I wish reviewers would look at it more frequently or at least make sure it's running before publishing reviews about it. It's honestly embarrassing that those two videos linked above with F1 don't have any disclaimer or update explaining that codemasters accidentally flipped the switch on DLSS without integrating it.
Gomez Addams
The ML used to implement DLSS was built by Nvidia, proprietarily, and must be used by developers for each specific game they want it to be available for. That is because there is a training process involved that must be performed specifically for a game. It is not a general purpose thing users can just enable for any game they want. Also, it is not part of a graphical API like DirectX, OpenGL, or Vulkan.
For more info, read this : https://www.digitaltrends.com/computing/everything-you-need-to-know-about-nvidias-rtx-dlss-technology/
schmidtbag
Ah ok, I guess I misunderstood that part.
You say that as though ML is something games have been in need of, or that Nvidia's approach to DLSS was necessary. It's nice, but the only ML I'm really interested in for games is for computer players; not for graphics.
Nvidia's approach might have required it to be proprietary, but, the gist of how it works doesn't need to be.
Unfortunately, that seems to be the case for most product reviews - the first impression is sometimes all you get, and a lot of the time, reviewers never give an update or 2nd look, even when comparing the product to newer ones.
I do agree that getting more comparisons would be nice, but at the same time, I have a real hard time caring about it. Whether you like the improvements or feel bitter about the quality loss, both the positive and negative the differences are way too subtle. Even with the improvements, the cons outweigh the pros for me to have any real interest in the technology. Of course, you're free to disagree, but I suspect I'm not the only one who feels this way or else it would be tested more often by more reviewers.
All that being said, I find AMD's solution technologically boring and less finely-tuned, so, I'm not especially interested in that either.
BetA
Thank God for reshade...
With it you can use "AMD´s image sharpening - CAS" in allmost every Game, dx9,10,11,12,Vulkan and every GFX Card. 🙂
And it really is a nice sharpener.
And THANKS TO GOD for AMD making it OPEN SOURCE.... Cheers for that...
Denial
DLSS is built on NGX - DirectML is essentially an alternative framework to NGX. Nvidia had to build NGX because DirectML didn't exist. The rest of the stuff you wrote I'm aware of but doesn't have anything to do with NGX vs DirectML or the argument of whether Nvidia could have done DLSS in a way that AMD could benefit.
I don't think anything is necessary until it is - I mean at some point a caveman got tired of being a caveman when everyone else was fine with it for millions of years. We had a technology, machine learning, that was being used for all kinds of image processing - so Nvidia decided to pivot that technology into games. They built an SDK, which does more than just DLSS and said "let's let developers see what they can do with machine learning in video games". It could have lead to better AI, DLSS could have been gods gift to gaming, it could lead to things I can't even think about - it could lead to nothing. Either way I'm a fan of innovation - it drives the industry forward. NGX drove Microsoft to create/finish DirectML - it drove AMD to pursue DLSS alternatives on DirectML. So regardless to whether it was necessary at the start or not it's pushing developers out of caves and potentially starting something great.
This is something that Nvidia has done for a while now regardless to whether gamers like them or not. AMD's GPUOpen alternatives are all in response to Gameworks. AMD wasn't talking about Freesync or using vblank to sync frames to the GPU prior to Gsync. AMD wasn't investing in ray collision hardware until Nvidia pushed RTX. Obviously it goes both ways, with AMD pushing Nvidia on certain things (Async Compute) - I wish Nvidia would be open with their tech and libraries but regardless it's still innovative and drives the industry forward.
