OK DLSS 2.0 - Promising technology that ACC could benefit from

To demand in such a tone, you're the one who pays the developers wage and you're the executive producer?

 

Hi guys,

I recently opened a thread about DLSS 2.0 on "ACC General Discussions" forum, asking directly to Kunos if there are any chance to see DLSS 2.0 implemented on ACC. Unfortunately, that thread has been closed so I´m asking to Kunos Dev Team here:

Is there any chance to implement DLSS 2.0 in ACC?

Thank you for your excellent job

Ale.

 

Yep, easy to say, not so easy to do.

Besides, does it work on the vast majority of Nvidia and any AMD cards ?

 

@LATE4APEX no this is Nvidia only and only for RTX cards (2060 and up)

 

So at this point, a waste of resources.

 

Does DLSS even work for simulators where everything is moving ?

 

Would be great if someone of Kunos Dev Team would say something.

Sure they will ;-)

 

https://www.assettocorsa.net/forum/...nos-staff-this-is-for-you.63446/#post-1124564

 

I really appreciate his answer, Aris is THE BOSS in handling and setups, but not on visual development, as he said in that thread by the way.

 

I search this question several times but they give us no answer why not?
I purchased a RTX 2080TI for 1300 euro and i'am only playing racegames
If I had knew i never bought the card

 

They did:

 

Being a full time Pimax VR player on a RTX 2080, DLSS 2.0 is the most wished for feature from me. Anything that can give me more FPS and visual clarity to give me a better experience in the game would be welcomed with open arms. At the moment, even with a 2080, my experience in the game is limited by the FPS I can achieve which currently isn't enough to be optimal.

The people that seem to be against this are obviously either non-RTX users or AMD users. What non-RTX Nvidia users don't seem to realise is that all new NVidia cards from here on out will all very likely support DLSS 2.0, so a future card you buy and use will be able to benefit from this feature being added to ACC. If you are in the AMD camp, with DLSS 2.0 looking to be as successful as it initially does, AMD will likely add similar functionality to their cards at some point. When that occurs, I'm 100% sure you will all want to see something like this implemented for your cards too. No point being narrowminded just because you don't have a card that can take advantage of this right now, as you are likely to benefit from it in the future as well.

As people have stated in posts above (see Drathuu's post), implementation of it doesn't appear to be a massive amount of effort. With NVidia already adding DLSS 2.0 into a custom branch of Unreal engine 4 and with them doing most of the heavy lifting, if cost isn't an issue and Kunos has someone with a bit of bandwidth to push this through, then why not?

 

The only people I’ve heard being “against” it are simply agreeing with the devs who state that it is currently not on their urgent priority list. When you consider that only a fraction of their users have the specific cards to take advantage of DLSS and they have other DLC to code, I can certainly understand that.

 

It was an advertised feature of the game from the beginning. Moreover, isn't it supposed to require very little dev work?

 

It was never advertised by Kunos or promised as a feature during development.

 

The only response we have is from Aris and as much as i like him, he's not a person responsible for that - as simple as that.

RTX cards are becoming the standard, Nvidia tends to progress in a way that makes previous model of said tier perform as good as lower tier model of next generation so 1060 performs as 970. 2060 performs like 1070 etc. That means that in next generation even 3050 will be RTX card. It would be stupid not to implement technology that is fairly easy to implement (as even indie games like deliver us to the moon have it) it gives massive performance gain because you're basically rendering 1/4th of the pixels.
Especially that ACC is very demanding on our rigs and especially that VR users have big problem running the game at all.

By your logic VR wouldn't ever exist. I guarantee you that there's far more RTX card ACC players than ACC VR players. Yet VR is implemented in the game and i guarantee you once again that it took far more effort from Kunos to implement than DLSS will.

 

I don’t have any insider knowledge, but just think about this. Kunos are a small studio which have multiple different priorities to work on. I’m implying that the highest of these priorities for them is meeting release goals for already-planned DLC packs, as they often come with deadlines that they are contractually obligated to meet; eg: race series tie-ins or race team liveries.

Adding content like the above is easy to do, coding is comparatively harder. And right now they have a serious cheating issue which just cropped up and needs immediate focus from the programmers. It’s in their commercial interest to make sure that the game as it stands is bulletproofed for esports use, before adding new code.

I didn’t mean to imply they’d never implement DLSS, but they need downtime to do it in a manner that doesn’t break the game for everyone else, and esp. for AMD users who DLSS wasn’t designed for. FWIW, as a CV1 user I’m very interested in this tech, although my 1080ti won’t run DLSS anyway. But I’d welcome the future scalability being present in ACC. But I understand it’s newness, and am willing to wait a while for it to make it onto the dev’s action-items list. .

 

I understand your point of view but i don't think it works the way you describe it.
DLSS is something that people responsible for graphics do, cheating is something that is a totally different issue and i seriously doubt that it's the same person's job. I know Kunos is a 30 person studio but still.

Different workers have different tasks. Aris isn't going to record audio, audio engineers won't optimise the textures throughout the tracks, and graphics guys won't address the cheating.

DLSS from what i read isn't really hard to implement at all, it's just a small module added to public branch of UE4. I'm not exactly sure what branch of UE4 Kunos are using but if it's something not that custom - then it should be a piece of cake. If it's something a bit custom it might take a bit of work to make it work but not a lot. DLSS is pretty much like a post-processing tool, it does not interact directly with rendering of the game. Moreover what's great with DLSS 2.0 is that process is very simplified because it's no longer required to train the AI on game to game basis. It's a broad solution that works across the board. So literally all it takes is to implement the module to the engine and to decide how much sharpening should be added, or if developers want to they might add a slider for that to control it.

Deliver us to the moon is a game made by KeokeN interactive - it's a studio with 6 employees. And the DLSS works wonders in this game.
I respect Kunos and i'm not going to say some BS that they 'immediately have to implement it' because they don't need to. But i genuinely think that they should implement it within 3 or 4 months because that's a reasonable timeframe to do so.

And just to remind you - we're not talking about some gimmick that adds fancier reflections to the game or shortens load times. We're talking about technology that in somecases more than doubles the framerate. That's a giant technological advancement.

 

If you’re that patient, we are on the same page. I believe Kunos would love to integrate it, given the luxury of time to do so.

 

We are going back and forth debating whether they will or wont add it, but I honestly wouldn't be surprised if it is on their roadmap already. For all we know they might have someone in the team working on adding it as we speak. I doubt it will appear in 1.4 as they have enough on their plate with that and the cheating issues, but I think it will probably come sometime in the next few months. I'm not impatient either and would be more than happy if it was on their 3/4 month roadmap.